DISTRIBUTION CHANNEL FOR MOROCCAN ARTISANS By Zyad El Jebbari M.Sc. Science and Executive Engineering Ecole des Mines ParisTech, 2010 SUBMITTED TO THE ENGINEERING SYSTEMS DIVISION IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ENGINEERING IN LOGISTICS ARCHIVES at the 1riT MASSACH3F_ OF EHN(LL JUL 6015 JL 16 201 MASSACHUSETTS INSTITUTE OF TECHNOLOGY June 2015 LIBRARIES 2015 Zyad El Jebbari. All rights reserved. / The author hereby grants to MIT permission to reproduce and to distribute publicly paper and electronic copies of this thesis document in whole or in part or any medium now known or hereafter created1 redacted Signature n u red Signature of Author: ................... .S Master of Engineering in rogram, Eng = gtA Certified by. ....................... ...... Division 102015 Signature redacted............. Dr. Edgar Blanco Research Director, Cent r sportation & Logistics Executive Director, MIT S AoNetwrk Latin America Signature redacted Thesis Supervisor Dr. Tauhid Zaman KDD Career Development Professor in Communications and Technology Assistant Professor of Operations Management, MIT Sloan School of Management Supervisor / zThesis Signature redacted 7... .(ILDr. Ac c e pt e d by : ................ .1-. -------......................D .................-f Yossi Sheffi Elisha Gray II Professor of Engineering Systems Director, MIT Center for Transportation and Logistics Professor, Civil and Environmental Engineering 1 e MITLibraries 77 Massachusetts Avenue Cambridge, MA 02139 http://Iibraries.mit.edu/ask DISCLAIMER NOTICE Due to the condition of the original material, there are unavoidable flaws in this reproduction. We have made every effort possible to provide you with the best copy available. Thank you. Page numbering has been cropped off from the bottom page margins. DISTRIBUTION CHANNEL FOR MOROCCAN ARTISANS By Zyad El Jebbari Submitted to the Engineering Systems Division on May 8th, 2015 in Partial Fulfilment of the Requirements for the Degree of Master of Engineering in Logistics Abstract The handicraft sector in Morocco represents more than 9% of the nation's GDP and employs 2.3 million people. However, despite their talent and expertise, Morocco's artisans are struggling to reach markets other than local regional or national markets. The main distribution model used so far has been the classic retail distribution model, which does not add value to the artisans' work, with most downstream actors playing the role of the middlemen. Other Ecommerce models such as the marketplace model or the crowdfunding model did not solve all the issues faced by artisans and customers. This paper provides the framework for a new distribution channel for Moroccan artisans to improve the efficiency of the value chain linking them to the American consumer by eliminating all unnecessary steps, overhead, and overall inefficiencies of the current distribution models. We first identify the models that are currently used to bring the artisans' crafts to the end consumer in developed markets: the market place (used by the platform Etsy), the crowdfunding model and the classic retail distribution model involving middlemen. We assessed the benefits of all the models by developing a new framework allowing to score each one of them. We then switched our focus to a hybrid model: a crowdfunding platform where the company partners with the artisans to align the different stakeholders' interests. To the extent of our knowledge, this model has not been used yet. It was found that our methodology could help artisans access international markets more efficiently. We finished by determining the optimal strategy using social media to market this platform and engage more efficiently potential customers. Keywords: handicraft, supply chain, crowdfunding, retail, social network marketing Thesis Supervisors: -Dr. Edgar Blanco: Research Director, MIT Center for Transportation & Logistics -Dr. Tauhid Zaman: Assistant Professor of Operations Management, MIT Sloan School of Management 2 Acknowledgments I am appreciative of those individuals who have assisted with my thesis research. I am indebted to the following people: To my family for their support and encouragement. To the Moroccan Minister of Handicraft and Social Solidarity Economy Dr. Fatima Marouan for her support and commitment to the project. To Dr. Bruce Arntzen for encouraging me to pursue an independent thesis project based on my interest in global supply chain. To the management team and staff of the Legatum Center, Dr. Fiona Murray, Dr. Iqbal Qadir, Elizabeth Henry, Julia Turnbull, Anna Omura, Kwadwo Poku and my mentor Ajit Janghiani for their support of and collaboration on my thesis project. To Mrs. Kathleen Campbell, purchasing director of Ten Thousand Villages for generously offering her time and insights on the crafts market in the United States and improving my understanding of the mission of a Fair Trade retailer such as Ten Thousand Villages. To my MIT ORC and EECS colleagues Krishnan Rajagopalan and Sourav Das for their thoughtful input on my research. To Dr. Sinan Aral for generously offering his time, insights and support to improve my understanding and application of his methodology to flexibility in digital marketing. And especially to my thesis supervisors, Dr. Edgar Blanco and Dr. Tauhid Zaman, for their invaluable advice and guidance during many hours of discussion throughout the academic year. It has been a great pleasure to have had the opportunity to work with you both over the course of this thesis project. Thank you. Table of Contents A b stract........................................................................................................................... 2 Acknowledgments..................................................................................................... 3 L ist of Figu res ................................................................................................................. 6 L ist of T ab les .................................................................................................................. 7 1. In trodu ction ................................................................................................................ 8 1.1. Artisans' platforms and current distribution ..................................................... 9 1.2. Motivation & Contribution ................................................................................... 12 1.3. Thesis Outline ..................................................................................................... 13 2. Literature Review ................................................................................................... 15 2.1. General Supply Chain Management Processes and Physical Flows:................ 16 2.2. Funding and financial flows in the supply chain ................................................ 18 2.3. Information flow and relating to the seller/content............................................ 19 3. Meth od s..................................................................................................................... 22 3.1. Artisans' platforms background ........................................................................ 22 3.2. Hybrid platform building and proof-of-concept................................................... 24 3.3. Social network experiments: .............................................................................. 24 4. Evaluating Traditional Supply Chains for Handicrafts........................................ 26 4.1. Data collected from interviews on the Retailer model.................... 26 4.2. Data collected from surveys on the Marketplace Etsy ................... 28 4.3. Data collected from surveys on the crowdfunding model ................. 30 4.4. Initial scoring of the traditional models ............................................................. 32 4.5. The customer survey sent to understand customers' preferences .................... 33 5. The Hybrid platform .............................................................................................. 37 5.1. Description of the model...................................................................................... 37 5.2. Proof of concept on Kickstarter ........................................................................ 38 5.2.1. The Supply chains............................................................................................ 39 5.2.2. Main challenges................................................................................................. 43 6. Evaluating Social Network Impact in the MoroCrafts Hybrid Platform............ 44 6.1. Introduction to Marketing and social networks ................................................ 44 6.2. The Social network experim ents ........................................................................ 45 6.2.1. The m ost efficient strategy............................................................................... 45 6.2.2. Causal relationship between gender and engagement on social media.......... 50 7. D iscussion.................................................................................................................. 52 7.1. Scoring of the different m odels ......................................................................... 52 7.2. Selection of the right supply chain ..................................................................... 53 7.3. Relevance of Facebook Ads ................................................................................ 53 7.4. Best strategy on social networks for social engagement .................................... 54 8. Conclusion................................................................................................................. 57 8.1. Summ ary ................................................................................................................ 57 8.2. Further research .................................................................................................. 59 Appendix....................................................................................................................... 61 References ..................................................................................................................... 71 List of Figures Figure 1 The classic retail distribution model .......................................................... Figure 2 The four types of crowdfunding models ................................................... Figure 3 Flows between suppliers and customers in a supply chain (source clerm iston .com ) ............................................................................................................ Figure 4 The eight essential supply chain management processes........................... Figure 5 Differences between existing artisans' models........................................... Figure 6 How buyers heard about Etsy (source Etsy)............................................. Figure 7 Breakdown of items bought on the Etsy platform (source Etsy) ............. Figure 8 The number of crowdfunding platforms in 2012 (source Corw dsourcing.org).................................................................................................. Figure 9 Breakdown of the different crowdfunding categories in terms of number platform s (source M assolution)................................................................................. Figure 10 Scoring of the first 4 different models...................................................... Figure 11 Some articles offered for sale during the campaign/pilot test ................. Figure 12 Share of revenues and profits per SKU.................................................. Figure 13 Breakdown of costs and margins for each SKU ...................................... Figure 14 Scoring of the different m odels.................................................................. 10 12 15 16 23 29 30 31 of 32 33 39 40 42 52 List of Tables Table 1 Price point and number of items sold for each SKU................................... 39 Table 2 Breakdown of costs and margins for each SKU......................................... 41 Table 3 Description of the accounts and actions performed during the social network 48 experim ent on T w itter .............................................................................................. Table 4 Results for social experiment 1: most efficient strategy............................. 48 Table 5 Logistics regression analysis ....................................................................... 50 Table 6 Results for social experiment 2: influence of artisan gender ...................... 50 Introduction 1. Introduction According to the Ministry of Handicraft and Social Solidarity Economy in Morocco, 20% of the population is represented by the artisans who are living in poverty with no access to the main necessities of life (Ministry of Handicraft, 2012). The Moroccan handicraft industry represents 9% of the total GDP, which is considered a high percentage given that more than 20% of the country's population works in the handicraft industry (Ministry of Handicraft, 2014). There has been an increased demand for authentic handmade products recently due to consumers' awareness of global handicrafts items and their seeking to explore new cultures. Even though the products are widely available in Morocco and there is high demand for them, there is no linkage between the supply and demand. According to a report from the Moroccan Ministry of Handicrafts, the handicraft sector has historically focused its activity on three pairs of products/customers: - Crafts targeting domestic retail consumption, which generates a revenue of $30 million. - Crafts targeting domestic and international tourist consumption, which generates a revenue of $1 billion, mainly from the domestic Moroccan customers. - Crafts with a strong cultural content production to be exported outside the country (Moroccan Ministry of Handicraft and Social Solidarity Economy, 2007). This last category has not being able to exceed $70 million in revenue despite its great potential for both local and international customers and a strong global demand for the items relating to Moroccan civilization, art, and culture. In 2007, the Moroccan government put in place a strategy to multiply the revenue generated from exported crafts by 10 by 2015, but this objective will not be achieved. Because the market is inherently fragmented and Moroccan producers are not organized, structured or developed enough to adapt to new distribution channels and access international markets, the Ministry accepted help from "reference actors" (big companies producing crafts items), which intended to drive international exports. But this strategy was not successful because these companies lacked the necessary technology and operations expertise to access the American market: the revenue from exporting crafts to the US market remained the same, at around 10 million dollars (Moroccan Ministry of Handicraft and Social Solidarity Economy, 2012). 1.1. Artisans' platforms and current distribution In the classic retail distribution model that the reference actors used (Figure 1), the consumer had to pay a very high price compared what the artisans earned because of a lack of supply chain and technology optimization in the process. It is a rather fragmented model with actors that try to transfer (inventory and financing) risk to each other: artisan, Moroccan wholesaler, exporter, US wholesaler/importer, retailer. At best, it was found that in general the final retail prices paid by consumers for artisan crafts represented, on average, 5 to 6 times the initial price paid to the artisan. This is explained by the fact that each actor is taking a risk (e.g. inventory risk, capital risk, assets deployed, demand risk) and a profit margin for that risk We took the example of Ten Thousand Villages to illustrate how a fair trade retailer could improve the typical retail model by cutting some unnecessary middle men in the chain. Today, Ten Thousand Villages is a nonprofit fair trade retailer that markets handcrafted products made by disadvantaged artisans from more than 120 artisan groups in more than 35 countries. It was founded in 1945 by Ms. Ruth Byler, who began a grassroots campaign among her family and friends in the United States by selling handcrafted products to help reduce poverty among artisans. Ten Thousand Villages' model eliminates the Wholesaler and the Wholesale Distributor in Figure 1. Ten Thousand Villages is the importer, the wholesale distributor and the retailer (they also wholesale to other retailers). Some of their product are shipped directly from artisans and they connect to some artisan workshops through in-country fair trade organizations that help them communicate design, organize shipping and transfer funds. These single organizations would cover the functions that have been broken out into two actors of Wholesaler and Exporter. Artisan Wholesaler ( xotr$ Importer WholesalerI Dsrbtr Retailer Figure 1 The classic retail distribution model The other widely used distribution model is the marketplace model illustrated by Etsy. Etsy is an e-Commerce website focused on handmade or vintage items and supplies, as well as unique factory-manufactured items. The items include art, photography, clothing, jewelry, edibles, quilts, and toys. Etsy is modelled after open craft fairs that give sellers personal storefronts where they can display their goods. The company charges artists a flat listing fee (20 cents per item), and takes a commission of 3.5% from all items sold. Etsy charges also payment fees through direct checkout: 3% + $0.25 fee on total sale price (including sales tax and shipping). Since its launch in June 2005, the site has continually added new widgets and features, and has seen corresponding growth. However, as a marketplace, Etsy does not take care of the shipping and logistics, which are left to the seller, who might not have expertise in this area. While it has been extremely successful at providing access to markets for artists, Etsy does not address some main concerns: financing costs should be paid upfront by the artisan, there are no quality controls in place, no "artisan" background checks are done (sometimes wholesalers use it while claiming they are artisans), and it does not "emotionally" connect buyers with the artisans' stories. A final model used is the reward-based crowdfunding model whose early success came from sites like Kickstarter and Indiegogo. Kickstarter is a global crowdfunding platform based in the United States. Kickstarter was launched on April 28, 2009, by Perry Chen, Yancey Strickler, and Charles Adler (Wikipedia). The app is aimed at users who create and back projects and is the first time Kickstarter has had an official mobile presence. Kickstarter is currently open to projects based in the United Kingdom, Canada, Australia and New Zealand, Denmark, Ireland, Norway, and Sweden. Kickstarter has reportedly received more than $1.5 billion in pledges from 7.8 million backers to fund 200,000 creative projects (Wikipedia, 2014). There are 4 main categories of crowdfunding: reward based, lending based, equity based and donation based (Figure 2); one of the most popular is rewards-based crowdfunding (the kind used on Kickstarter and Indiegogo). -1 Figure 2 The four types of crowdfunding models Rewards-based crowdfunding has been a source of funding for small businesses. Businesses and non-profits of all sizes post projects looking for funding on a crowdfunding portal, targeting a certain amount to raise. In return for a donation from backers, a business or non-profit gives some type of incentive for participating (Lambert and Schwienbacher, 2010). However, this model is only present in the US and Western Countries since artisans from the rest of the world (especially Africa) do not have the connection or credibility to leverage the platform to reach users. The crowdfunding market is essentially American (72%) and European (26%), the rest of the world sharing the remaining 2%. The reason is certainly above all confidence in online transactions and possible legal remedies (Crowdsourcing, 2012). 1.2. Motivation & Contribution The motivation for this thesis is to develop a new model of distribution for Moroccan and African artists to help them access international markets. We will use a hybrid crowdfunding platform that we will call "MoroCrafts" to better link artisans to markets by providing customers with authentic handmade products. This platform aims at increasing the artisans' profit margin and improving their quality of life. This project will be used to explore the supply chain and social media dimensions of this business. The results of this research directly enhance the distribution models used to export art and craft items in the context of a global supply chain. When the prepayment of items through a crowdfunding platform is combined with an efficient internal supply chain, both artisans and consumers will be more satisfied. The scoring framework we developed during our research can be applied to assess the efficiency of distribution channels for all sorts of items that countries would like to export to international markets (e.g. traditional cosmetics, accessories, authentic home decor). In addition, social media strategies can be used within this new model to target and engage the right audience. Using the tools developed for MoroCrafts on Twitter to spread the word and build a linked community, the analysis highlights strategies that can be expanded to other social network tools such as Facebook, Instagram or Pinterest. 1.3. Thesis Outline Chapter 2 presents the literature review, which provides an overview of the frameworks used to assess supply chain efficiency and performance. It also provides research done in the field of social media and digital marketing in the field of the handicraft and e-Commerce. Chapter 3 describes the methodology applied in the research, specifically the methodology for evaluating and scoring the current distribution models for art. We also present a new hybrid crowdfunding model and address the issues identified within this scoring framework. We implement a pilot test to assess the score of the platform through the framework and see if this hybrid model performs better than the current models in place. We will finally experiment various digital marketing strategies to engage e-Commerce customers. Chapter 4 examines the results of the data analysis. Chapter 5 presents the recommendations based on the findings in terms of the potential implementation of the model and the use of social network strategies in the e-Commerce space. Chapter 6 concludes with a summary and suggestions for further research. Literature review 2. Literature Review Previous research has tried to solve the problems described in chapter 1 separately (i.e. reducing the financing and inventory risks, eliminating the middlemen, funding) but no study has focused on providing a comprehensive solution for e-Commerce supply chains linking artisans in developing countries to international markets. We will focus on the three supply chain main flows: the materials flow, the information flow, and the cash flow (Figure 3). In the next section, we will review the research done in the area of efficient processes in the supply chain and the physical flows. In section 3, the classic retail distribution model is evaluated in the context offinancialflows. In section 4, we will outline the challenges of the informationflows within the context of the e-Commerce channel, particularly in relation to the emotional linkage between artisans and customers. MATERIALS INFORMATION CASH Figure 3 Flows between suppliers and customers in a supply chain (source clermiston.com) 2.1. General Supply Chain Management Processes and Physical Flows: One of the most prominent articles specific to streamlining the supply chain in the classic business model and cited numerous times by other researchers is by Douglas Lambert. The article outlines the "eight essential supply chain management processes" to efficiently manage a retail distribution channel (Lambert, 2004). AK VI Figure 4 The eight essential supply chain management processes These processes span from the customer relationship management, where the company should focus on the segmentation of customers, to product development and commercialization (Figure 4). Within that framework, the product returns management should ideally focus on product quality to minimize item returns and use promotions to ensure products will be sold and avoid obsolescence. However, in internet supply chains, inaccurate demand forecasting often leads to larger volumes of product returns for eCommerce retailers (Rabinovich & Knemeyer, 2006). Lambert focuses on including customers and suppliers in the same ecosystem to be sure to maximize the value for everyone (make the pie bigger) and the importance of coordination between all the stakeholders of the system. Some important actors in the Internet Supply Chains are the Logistics Service Providers or LSPs (Rabinovich & Knemeyer, 2006): LSPs are intermediaries who help internet sellers add value to their supply chains by consolidating orders and finding lower transportation rates. The last mile delivery can then be outsourced to FedEx or UPS. Rabinovich & Knemeyer studied the operations and economics of the Supply Chain specifically pertaining to the optimization of product delivery to meet customer needs. In situations where the product demand is high, it is crucial for the business to maintain a large enough product inventory as to ensure efficient and time-sensitive product processing and delivery as Rabinovich found that the majority of customers are time sensitive. From the demand side, in order to maximize revenue by optimizing the cost per product, the supplier must maintain high product demand. A lack of demand can induce high costs to pick, pack and ship small orders and impact the client's willingness to buy the product. Rabinovich's research has also shown that a new company can gain a competitive advantage through asset specificity by using outsourcing and LSPs at three levels: company and strategy level (Netflix), transaction level though segmentation (FreshDirect.com), and product level with the distribution of niche products (seasonal products such as flowers). However, outsourcing to LSPs is not recommended for sellers with high asset specificity (LSPs are reluctant to make specialized investments) and high uncertainty (the seller needs to take control of its logistics network). These conclusions suggests that additional research needs to be done for our case. 2.2. Funding and financial flows in the supply chain Another problem that can arise for small artisans and cooperatives is the lack of financing (Ibrahim and Verliyantina, 2012). High transaction costs incurred by the credit procedures makes this bank loan instrument not very well adapted for the small businesses. An alternative is to get rid of these funding risks by using the fair trade model that provides the artisans 50% of the initial funds to start the production. The microfinance model is also available to artisans but interest rates are very high (between 20% and 30% and sometimes even higher). As far as the fair trade model is concerned, it is still a middleman model involving many intermediaries where the artisan only receives a small percentage of the proceeds. An alternative to mitigate the inventory, funding and financial risks is the crowdfunding model: "Crowdfunding involves an open call, essentially through the Internet, for the provision of financial resources either in form of donations or in exchange for some form of reward and/or voting rights in order to support initiatives for specific purposes" (Lambert and Schwienbacher, 2010, p. 6). This model of raising capital can take the form of donations, sponsoring, pre-ordering/pre-selling, and as a result, the complexity of processes varies greatly. The idea of Community Driven Development (CDD) for individual and group-based microfinance can be implemented using the crowd funding mechanism that has already been proven as an established way to fund social and/or not-for-profit projects, particularly in the Third World (Hemer, 2011). Organizations which have a long tradition of fundraising for social and/or not-for-profit projects (e.g. the Red Cross, Oxfam, NGOs and other organizations for development aid) already employ the crowdfunding model extensively. Ibrahim and Verliyantina propose a crowdfunding system that serves as a bridge to link the different stakeholders of the project/business: crowdfunders, field partners, coaches and non-profit organizations, which work together to screen, supervise, and manage the use of funds. His proposed solution/model provides a comprehensive solution to the financing problem for Small and Medium Enterprises (SMEs). Such a system could be leveraged in a web-based ecosystem that exploits the capabilities of social networks, especially the "viral networking and marketing" functionalities, but the study in question did not assess the impact of the physical and financial flows, which will be part of our distribution model. In addition, trade-offs between transportation costs, lead time, product variety and customer service should be assessed as well (Trent, 2004). 2.3. Information flow and relating to the seller/content The process of crowdfunding also relies on social and viral marketing to encourage people to back and fund a project/business. Research has found that single men were more influential than married men/women with regard to spreading the word regarding a topic (Aral, 2011). This remark is challenging in our case, as we will show later that people who usually buy crafts online are mostly female. Aral explores how companies can incentivize users to recommend on their behalf. The more "generous" shared-incentive model (discount/promo offered to both the sender and receiver of the sharing) generates more sends than the "selfish" single-incentive did. In fact, people shy away from spamming their friends unless they can pass a benefit, promo, discount along to the friends with whom they are sharing the idea/concept. These results are very much in line with the sociological notion of a "gift economy," in which generosity confers status. Also, passive broadcast in social media (generating automated notifications to peers) is found more efficient at recruiting new users globally than active personalized techniques of word of mouth (personalized invites) in the case of a Movie App company (Aral, 2012). This finding will also need to be tested and further studied in the case of real assets. One should also the performance of these tools as way to acquire customers or early adopters through social network since they could be overestimated by homophily features (Aral, 2009). Homophily is the tendency of individuals to associate and bond with similar others. When assessing an efficient strategy to engage users in the internet, we should make sure that our strategy is the real driver that grows the audience and is not only due to the "birds of a feather flock together" effect. Lambert talks about the information flow needed between a regular company's suppliers and customers for its supply chain to work smoothly and efficiently (Lambert, 2004). However, although e-Commerce appears to be an emerging sales channel, only limited sales are occurring through it (Stoddard et al, 2012). This finding may be attributed to the hedonic nature of arts and crafts consumption (appeals to emotions). Indeed, an artistic product may not exist to fulfill a market need, a utilitarian concept, but rather a want, a hedonic concept. It makes arts, crafts, and cultural product marketing different compared to other types of products, and thus creates some marketing challenges (Colbert, 2003). Various studies have shown that tourists are important consumers of crafts (e.g., Stoddard, Evans, & Dave, 2008): 60% of craft artisan retail sales are tourism. In 2008, Etsy posted the results from a customer profile survey. Results from the survey of 30,000 customers indicated that Etsy's online buyers tended to be young (32 years old on average), predominately female (96%), and highly educated (63% graduated from college, 17% still in college). The survey also found that online users were mainly from the United States (82%), but their international user base was growing. They had "registered users" from 172 countries by 2008. The profile for these consumers is enlightening since other studies have found traditional arts and crafts consumers to be older, highly educated females with high incomes (e.g., Evans et al., 2006). Perhaps these findings suggest an emerging segment of arts and crafts consumers who feel more comfortable with electronic purchases. Evans' results also indicate there is a strong connection between retail craft galleries and shops in tourism destinations, which suggests that an effective marketing strategy for craft sales should reproduce this experience online so that people can connect emotionally with the artisans and the products. The literature review showed that some processes such as returns management, supplier and customer relationship management used in the regular distribution channels could be used in our particular case, especially when it comes to managing the physical flows from the suppliers to the end customers. However, the classic retail model does not optimize the returns for the artisans, especially if one needs to carry and manage the inventory risk to set up a responsive supply chain (Fischer, 1997). Outsourcing some risk to LSPs in internet supply chains is not recommended for products that exhibit high asset specificity. On the other hand, the crowdfunding model developed by Ibrahim and Verliyantina for a microfinance platform could be used as a starting point to manage more efficiently the financial flows to fund the artisans. This aspect should be further studied in the case of an international supply chain which deals with real assets. Finally, even if the classic information flow within the whole supply chain (as described by Trent and Fischer) plays an important role in internet supply chains, arts products are rather different from regular consumer products in the sense that they appeal to a hedonic concept rather than a utilitarian concept. We may need to study more in depth the dynamics that could help connect the consumer with artisans through the user experience on the platform and see how likely a consumer is to stay engaged in the community (Aral, 2009). Studying the lead time incurred by the latency of the crowdfunding model will be essential in our project. To the best of our knowledge, no holistic research has been done to optimize a crowdfunding distribution channel from artisans to end consumers in the developed markets. Methodology 3. Methods Based on our initial findings and literature review, we argue it is important to develop a new hybrid model that improves the way artisans' crafts are distributed to American clients. In this section, we will focus on three major points. First, we collected data for existing platforms and compared them on the basis of profit margins, lead time, financial risk and trust. We then explored how to design a hybrid platform to optimize these parameters and target the right customer segment (e.g., Evans et al., 2006). Finally, experiments on social networks were conducted to attract these potential customers on the platform. 3.1. Artisans' platforms background The first step in our methodology was to analyze the features of each of the models: the classic retail distribution model (e-Commerce website or fair trade organizations), the marketplace model, and the crowdfunding model. We then evaluated and scored them using 6 criteria taking both the artisan and the customer perspectives: - On the artisan's side: we assessed access to funding, profit margins, and access for artists in developing countries. In December 2014, we conducted interviews with the artisans during a field study in partnership with the Ministry. These 3 issues were identified as the top 3 challenges for artisans to get access to international markets. - On the customer's side: we evaluated supply chain and logistics management, emotional link and trust, and lead time. We also conducted customer surveys with potential buyers and retail stores and these issues were also mentioned several times as the top 3 criteria for a client to buy crafts coming from Morocco. Figure 5 shows an example how of how we will score the different models based on these 6 criteria from a scale of 0 to 2. irSloan At Reasonable Margin Ameesi for Emotnl link and Trust cfe LIly Cn Lead timne Figure 5 Differences between existing artisans' models Depending on the model, data was collected through different means. First, we interviewed the purchasing director of Ten Thousand Villages to get a sense of the different tradeoffs between trust, lead time, supply chain control and profit margins for a fair trade retailer. Second, we relied on the existing literature and print sources from customer surveys, such as the one published on the Etsy's website, to gather additional data on the dynamics of the marketplace and crowdfunding models such as Kickstarter. 3.2. Hybrid platform building and proof-of-concept After we scored the different models, we developed a new model that will solve the issues of the current models to import/export artisans' crafts. This hybrid model will be a crowdfunding platform that will display existing and new features to help artisans access international markets. The platform will reduce the inventory and financing risks, and increase the artisan's profit margins. A pilot test of the model has also been conducted between Moroccan Artisans and American consumers on Kickstarter, and we collected data concerning the market segment (sex, income, degree of study), conversion rate and cost of acquisition to target them more efficiently on social networks. In addition, a market study was conducted in collaboration with the Moroccan Ministry of Handicraft and a consulting company (Jola) to better assess the customer behaviors and tastes. A survey was sent to 12,000 customers across the United States (the survey can be found in the Appendix). 3.3. Social network experiments: Once the right audience is identified in the prototype phase, we will develop the platform and try to build the community around the project. To do so, we will use experiments on social networks and especially Twitter and Instagram. Twitter is an online social media platform where users can post 140-character segments of text to share with friends. These "status updates" can contain both original text passages as well as links to other sites. There are a few ways to engage with fellow users on Twitter: 1) When one "follows" another user, the status updates of the followed user, or the friend, get displayed on the homepage of the following user. 2) When a user wants to endorse another user's the post, he can re-post, or re-tweet, the original status. 3) A user can also reference a fellow user in his post by mentioning the other account name or directly replying to a post. In this experiment we investigated the efficacy of 3 different strategies: following, retweeting and replying, and measured the incremental benefits of combining 2 strategies (following and retweeting, following and replying). Once the successful strategy is identified, we used it to determine whether there is a causal effect between the gender of the artisan and the number of followers acquired. Indeed, from the interviews and initial research we conducted, potential are women and we want to test the hypothesis whether they are more likely to get interested in buying a craft because a woman or man produced it. We will run the same experiment for one more week by a male artisan and female artisan using the strategy that has been identified as being the most efficient. Chapter 4 will present the main results following the same organization as chapter 3. Evaluating Traditional Supply Chains for Handicrafts 4. Evaluating Traditional Supply Chains for Handicrafts The question we try to answer in this thesis is the best distribution channel and supply chain to distribute the Moroccan artisans' crafts in the United States. In this section, we first highlight the main differences between the different models existing to export artisans' crafts in the American market based on the interviews and customer surveys we conducted or found on the internet. Second, building on these insights, we built a framework for a new hybrid platform optimizing the main 6 parameters identified in the previous section: access to funding, profit margins, and access for artists in developing countries on the artisan's side, and supply chain and logistics management, emotional link and trust, and lead time on the customer's side. A proof-of-concept has been carried over and existing crowdfunding platform and we collected some interesting metrics. That means that we used an existing crowdfunding platform called Kickstarter to run a campaign in January and February, 2015 to test the logistics and the customers' preferences. Finally, experiments using social networks (essentially Twitter) were implemented in March and April, 2015 to develop the best strategy to build a community around this hybrid model. 4.1. Data collected from interviews on the Retailer model We interviewed Ms. Kathleen Campbell, the purchasing director of Ten Thousand Villages as part of our research study to learn more about the brick and mortar retail model. We learnt that the organization made around $28M in revenues each year, purchases account for $7.2M and 650 new items/year are added to the portfolio including jewelry scarves, hair accessory). They hold around $9M in inventory in their warehouses, which means that the items remain around 4 months in the warehouse (Inventory Turn=3). Ten Thousand Villages uses 2 channels to reach their customers: - Retail stores: 38 stores and 40 independent franchises under contract. - e-Commerce /website: $2M revenues and 17% growth year on year. Ms. Campbell confesses that she still has some difficulty conveying the story of artisans through this medium: people do not spend enough time to create the connection and experience as is the case in the store. Clients usually spend more time on the store than online and can learn more about the artisan's story and the culture of the country where the craft has been made. The suppliers are not individual artisans but rather coops or organizations (spanning from a few to hundreds of artisans in Bangladesh, for example). Designers usually attend workshops to get a sense of the new trends and colors, artisans provide samples, and designers/buyers (4 buyers who spend 6-8 weeks traveling in their respective region) guide the artisans without constraining their creativity. The advantage of the approach is that the organization gets an immediate feedback from stores. However, the design process needs to begin 14-16 months ahead of the next season. The production begins 6-8 months later and shipping takes around 2 months. Concerning financing, 50% paid in advance to artisans and 50% at the time of delivery. Providing this upfront payment to artisans to help them buy the raw materials they need is a requirement for every Fair Trade Organization. Invoices are sent to artisans' communities separately. The exporter consolidates the items to be shipped to the United States from the supplier and the bill of lading is then sent to him. 90% of the time, products are packaged at the origin. The inventory risk is carried by Ten Thousand Villages ($9M in inventory in warehouse). UPS manages domestic logistics. There are very few product returns (less than 2%). In terms of pricing, some products are price sensitive (unique products fall in this category) and discounts do not affect the sales volume. Other products, such as ceramics for example, are not very price sensitive (offered at WalMart or Target at lower prices). Gross margin is around 55% to 60% and the price in the store represents 4 or 5 times the initial price in order to get a 10% margin bottom line for the organization. This data allows us to quantify 2 criteria in our scoring model: margins and logistics management. 4.2. Data collected from surveys on the Marketplace Etsy In 2008, the website Etsy ran a survey with more than 30,000 participants. A similar survey was realized in 2010 as well. Almost 46% of the participants were buyers, whereas the other 10% represented sellers, and 39% did both. Ninety-six percent of Etsy's users are highly educated females. Both buyers and sellers are a highly educated group, with 63% graduating from college and 17% still in college. Buyers, on average, are 32 years old and sellers are slightly older, averaging 35 years old. The customer base is mainly from the United States: 82% of buyers and 86% of sellers are from the United States even though the website registered users from more than172 countries. Many buyers find out about Etsy through blogs: 26% find out about Etsy on blogs and 17% find out about the handmade products they buy on blogs. More than 50% of the sellers who responded to the survey participate in blogging, which seems to be the best way to get attraction. The most frequent way buyers find out about Etsy is through a friend (33%) (Figure 6), so word of mouth and networking are of paramount importance: 72% of buyers and 78% of sellers have a profile on at least one social networking site. This suggests that Facebook, Instagram, Twitter or Pinterest pages may be a great marketing tool. How buyers hear about Etsy Where buyers learn about handmade products Friend3% 50% Oine C r1% Web seapr 19% Other 8% 14% Frienids 6% Social networkig site 9 Craft Fairs Is% % Mawim Craft Fai or other event 2 .G-q.stoms Mag14nus 6% 4% AdVerisementM Television rA *| l wraN Figure 6 How buyers heard about Etsy (source Etsy) Buyers shop on Etsy primarily to find unique items, because they value handmade goods and they want to support independent artists. Figure 7 shows that the major items bought on the Etsy platform are jewelry accessories. Top ten categories for sellers Top ten categories for buyers JNWBy ipry Topre 17% PqIWG.S ZM Cmb" f st - 4% t caegkowe fo buyes bTgopo te ten ateorie(soreesy 21%ramework P 19% bottlenc fo3h afr.Iti simtdta Figure 7 Breakdown of items bought -4 on the Etsy platform (source Etsy) Esy makes ilo eri esfo Buyers, who represent 80% of all respondents, are also willing to pay more for high quality, well-crafted items. A recent article in Business Insider raised some concerns about the authenticity of crafts on Etsy. While Etsy started as a way for artisans to get access to a larger market, scale became an issue over time since supply of handmade crafts became a bottleneck for the platform. It is estimated that Etsy makes $47 million a year in fees from sellers. It's the fifth most-visited marketplace in the world, after Amazon, eBay, Best Buy, and Wal-Mart. Many items offered for sale are made wholesale imported from other countries, which does not align with the original mission anymore and may hurt their brand image in the future. All this data allow us to score the model according to the 6 criteria pf our framework. 4.3. Data collected from surveys on the crowdfunding model In 2012, the company Crowdsourcing LLC conducted a research study concerning the crowdfunding platform in the world. The global crowdfunding market has reached 6 billion in 2013 (vs. 500 million in 2009), a growth rate of 88%. According to a study by the World Bank, it is estimated that the global crowdfunding market will reach $ 1,000 billion in 2020. Now, the crowdfunding market is essentially American (72%) and Europe (26%), the rest of the world sharing the remaining 2%. The reason for this geographic concentration is certainly the tax exemption of contributions devices (friendly regulations) and above all confidence in online transactions (Figure 8). N S" NUMBER OF CFPs, 2012 6 BELGIUM jS 4 POLANDl S191 UNITED SLE 4 rA -ID 'IPA MN N 1NORWAY SMI 2 MEXICO SWEDEN DENMARK I IRELAND f 3 FICLAEND C ' CZCH REP. A 1 AUSTRIA E I ESTONIA L I LATVIA I 1ROMAMA 1 HUNGARY A HONG 1 HKAMJAPAN PHIWPINES 2 ARGENTINA 6 NEW ZEALAND SOL rcF: Eased on Cr asoourcir.g.org Directory of Sites as of Apr 201.2 Figure 8 The number of crowdfunding platforms in 2012 (source Crowdsourcing.org) The reward-based model that a platform such Kickstarter uses represents 62% of the whole market (Massolution, 2012). Figure 9 shows the split between the different models and we can clearly see that the Kickstarter model is the largest in terms of number of platforms represented. It is also the category that experienced the highest growth in the past few years. 143 41% cAGR DOWAnONBASED 79% CAGR REWARDBASED 3. 21 U50% CAGR LENDING-BASED 114% cAem EQUITY-BASED 2007 2005 2009 2010 2011 Figure 9 Breakdown of the different crowdfunding categories in terms of number of platforms (source Massolution) Project launchers need to have a bank account in the countries where Kickstarter is present. African and Asian artists are excluded. Kickstarter applies a 5% fee on the total amount of the funds raised. Their payments processor applies an additional 3-5% fee. 4.4. Initial scoring of the traditional models Based on the interviews, each model received a score of - 0 (no sign): no access to free loans, no reasonable margin for the artisan, no access for African artists, no emotional link and trust for the customer, no reliable shipping and logistics management by the platform, long lead time to the customer. - 1 (one X): adequate access to working capital, relatively good margin for the artisan, relatively good access for African artists, adequate emotional link and trust for the customer, relatively reliable shipping and logistics management by the platform, relatively short lead time to the customer. - 2 (two X): good access to free loans, reasonable margin for the artisan, good access for African artists, good emotional link and trust for the customer, reliable shipping and logistics management by the platform, short lead time to the customer. The following table sums up the different findings (Figure 10). The retail model has been split between the classic online retail model and the fair trade retail model. Online Retail Fair-Trade Model Reail Model Marketplace CrowdFunding MoroCrafts Mods Hybrid Priterh Reasonable Emotional link and Trust cusler X X XX XX X X XX xx X X X Lead time Xx Figure 10 Scoring of the first 4 different models 4.5. The customer survey sent to understand customers' preferences In collaboration with the Ministry of Handicraft and a consulting firm, a survey was sent to over 12,000 consumers across the United States. The consumers surveyed for this report purchase home furnishings several times a year. They usually buy home accents more often than any other category, but also seek out and purchase more practical items such as furniture and lighting. They buy these items at home decor retail stores more often than anywhere else, but make frequent online purchases as well. - 58.5%purchase home furnishings several times a year - 25% buy home decor when making a purchase - 68% shop in home accents stores - 47% shop online, too They are inspired by visiting retail stores and by shopping online. Nearly 90% have purchased home decor online. Social media also plays a role, but shelter magazines barely register. They love Pinterest. Seventy-one percent go to retail stores for inspiration. Sixtyseven percent go online to get ideas. Once online, nearly 90% make purchases and 29.5% use social media to get information on products 76.5% love Pinterest. There is a growing interest in handcrafted home decor. A majority of these consumer have purchased artisan products, and a significant portion would pay more for such goods because they believe artisan products have more value. Most of these purchases were made in gift stores. - 70.5% have purchased artisan or fair trade products in the last 12 months - 38% would pay more for such goods. Of those, 54.5% would pay up to 20% more - 73% believe that artisan products have more value - 49% of purchases were made in gift stores - 24% were made online The majority of these consumers had not purchased Moroccan products. They are more interested in price and design than in provenance. A little more than half would be hesitant to purchase from a Moroccan company without a US presence, and in general, they were not interested in learning more about Morocco. - 22.5% have purchased Moroccan products - 71.5% have not - 20.5% say it is important that Moroccan design is made in Morocco - 43% would purchase from a site that shipped directly from Morocco But only 22% are interested in learning more about Morocco. In spite of that, when asked for their impressions of Morocco and Moroccan design, the most prevalent words and phrases were as follows: Exotic and unique, Bright and colorful, Rich culture, Beautiful, Warm and inviting Only a small number of those surveyed had visited Morocco. Of consumers who have purchased Moroccan products, most did so at retail stores. A smaller number did so online. For future purchases, they are most interested in home decor and jewelry. A significant number are also interested in rugs. Only 3.5% had visited Morocco 14.5% want to in the future Of those who have purchased Moroccan products, 70% did so at a retail store 23% did so online. From these customers: - 67% are interested in buying home d6cor products - 61% are interested in jewelry - 36.5% are interested in rugs. The initial results suggest that the average consumer surveyed is an active customer of home decor products. They like to shop in retail stores and online. They are responsive to social networks such as Pinterest and Twitter. They are willing up to 20% more for authentic handcrafted, artisan-made and socially responsible products, only if the design and the price are right. The Hybrid Platform Design and Results 5. The Hybrid platform 5.1. Description of the model Based on our preliminary study, we imagined and designed the requirement of a new platform that solves the issues of the current models in place to import/export artisans' crafts as described in the methodology section. The platform reduces the inventory and financing risks, and increase the artisan's profit margins. Given the nature of the crowdfunding model, artisans produce only crafts requested (and paid for) by consumers, which allow us to put in place a low-to-no inventory model and increase the artisans' margins. Artisans cannot afford holding inventory and taking on the risk that is associated with it and find themselves at the bottom of the supply chain, being squeezed by middlemen/retailers who can take on that inventory/financing risk. Also, given the structure of the crowdfunding model, the artisans are not disconnected and far removed from the end-users as they are able to align their production with the customers' tastes, preferences, and behaviors. As a result, the demand and offer will meet at the "right", optimal price. Additionally, the platform gives the artisans a means to share their stories and emotionally connect with their users in a beautifully designed environment such as Pinterest: it offers a visual discovery, collection, and sharing channel emotionally linking and connecting the craftsmen and buyers (which allows users/buyers to travel and discover remote areas of the world through the platform). That proximity will be even more stressed with users having the option to customize the crafts they are requesting. To solve the marketplace model issues, specifically the lack of background checks and quality control, we are partnering with community managers. These managers constitute another line of defense with regard to quality controls and partner with and handhold artisans to improve their processes, design, and administrative duties. This way, the artisans can focus on their core strength: building beautiful products for buyers across the world. The community managers assist artisans optimizing the lead time (from sourcing to end consumer) through a reliable direct Supply Chain. The business model is rather simple: we set up a visual discovery/sharing, crowdfunding platform and a commission is taken on crafts sold by artisans to customers. The artisans are managed by community managers and costs are limited to the investment in the platform, community managers' salaries and marketing to grow user base if needed. The platform that we will call "MoroCrafts" on the rest of the study will rely heavily on social media and viral marketing to grow that part of the business and solve for that side of the equation. 5.2. Proof of concept on Kickstarter We started with a pilot test between Moroccan Artisans and American consumers and sold 18 items or SKUs as if they would be sold in our platform. Figure 11 shows some of the articles we offered during this campaign: teapots, candleholders, mini tagines, traditional leather bags, silver necklaces and bracelets, rugs, lanterns... 35$ Connect with the Artisans!! '350$ Rug-Rug Made by Middle Atlas Tribes 12$ Fez Medina Bag (Small) - 67$ LANTERNE BOIS BLEUE I lue Maeoelle Wood Lantern 14f5 in) with aplionto color:le, ack,Whle I oe Talk direcly to wastan.a Size: 7 ft33 ft. .-'--..--1 65$ Connect with Us! - - - - - -- --- - 1- Lantern 3$ Kilim R AU9 Rmeel by W taen. hde$ade byM 850$ Rug BENI OURAIN RUG C.udle Helde, wvl s sruan a-Y in) Ant en Sk se coy . We will eheduie a Larg sie (17*in) -65$ eratimnoryou to ask umablany. with optlen to engrav thing you would like to askkastopodlng tt. eudeov"c and.fa u t heot yeur Yo iniesO m coat I 95$ Silver Pendant SilPernd :95 Silver Bracelet in &e Tfelt Region. Whit, brown. Purle We wigl vdeotaps, th .ztissusn oath. ground with their sunet. an d it diectlyto yulo connd iltend 60-856 Bronze Conned WithnU! andI liesty. Aska qestionabot itheir caft,story cuture, family, Size: 5 ft *6.5 ft. "30$ Fez Medina Bag I(LAWg) Grey 25$ Marrakech Evening IClutch (saafl I *Silver bracelet with Arabesque -opaving anddeeigo Figure 11 Some articles offered for sale during the campaign/pilot test 5.2.1. The Supply chains After one month, we could raise $10,212 on Kickstarter from 91 different backers. $1,996 were donations without any counterparty while $8,216 were items bought by customers. If we exclude the website fees of 8% and the donations, the average amount purchased for this campaign was around $53. Table 1 shows the price point for each SKU Item Rugs Bracelet Necklace Quantity Retail Price 350.00 110.00 70.00 6 8 4 $ 15.00 25.00 14 12 Large bag Mini Tagine Bougie bois petit $ $ $ 30.00 20.00 34.00 8 36 10 Small bag Medium Bag $ $ $ $ Bougie khmissa metal petit $ 34.00 4 Teapot colore petit $ 45.00 10 Teapot point petit $ 45.00 12 Petite lanteme bronze $ 60.00 2 Bougie bois carre grand Bougie khmissa metal grand $ $ 65.00 65.00 4 4 Teapot colore grand $ 65.00 2 Teapot point grand $ 65.00 6 Lanteme bois bleue $ 67.00 10 Lanterne bois bronze moyenne $ 95.00 2 Table 1 Price point and number of items sold for each SKU as well as the number of items sold while figure 12 shows the percentage of revenues and profits for each item: Share of revenues and profits per SKU 30.0% 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% 0 U Share O '\ of Revenues 06 ' C 0 Share of Profits Figure 12 Share of revenues and profits per SKU We can notice that there are only 3 items that represent more than 10% of profits and revenues: rugs and bracelets which are high value items and the mini tagines which are affordable (they are the ones that were bought the most in terms of volume. We could consider them as A items. In general, revenues and profits are pretty much correlated expect from the small bags and the blue lanterns that were underpriced by comparison to other items. We used two different supply chains for these products: - A direct supply chain for expensive and bulky items such as rugs through FedEx. Lead time was less than 2 weeks and logistics costs were around 15.3% of the final retail price. - An indirect supply chain for customized and smaller items: we sent them by bulk to the fulfillment center to reach a critical mass and then sent them back to customers via the regular US domestic post office. Lead time was significantly higher (more than 6 weeks) and logistics costs were around 26.4% of the final retail price. (Written after items sent to customers next week). Figure 13 and Table 2 summarizes the breakdown of production cost, shipping costs and margins per item: Item Rugs Bracelet Necklace Small bag Medium Bag Large bag Mini Tagine Bougie bois petit Bougie khmissa metal petit Teapot colore petit Teapot point petit Petite lanterne bronze Bougie bois carre grand Bougie khmissa metal grand Teapot colore grand Teapot point grand Lanterne bois bleue Lanterne bois bronze moyenne Average Share Production cost Share shipping cost 15.3% 35.0% 36.5% 5.0% 35.0% 8.6% 47.6% 36.7% 32.7% 22.0% 30.6% 18.3% 12.8% 30.0% 13.5% 29.4% 29.4% 13.5% 27.2% 17.8% 27.2% 17.8% 29.8% 36.7% 12.6% 55.4% 12.6% 55.4% 37.7% 15.4% 37.7% 15.4% 29.9% 38.8% 29.5% 26.3% 28.3% 25.8% Margins 49.7% 58.5% 56.4% 15.7% 45.3% 51.1% 57.2% 57.1% 57.1% 55.0% 55.0% 33.6% 32.1% 32.1% 46.9% 46.9% 31.3% 44.1% 45.8% Table 2 Breakdown of costs and margins for each SKU Breakdown of the price for each SKU 120.0% 100.0% 80.0% 60.0% 40.0% 20.0% 0.0% o quo %Q e 0 Share Production cost N Share shipping cost 0 Margins Figure 13 Breakdown of costs and margins for each SKU Again if we exclude the commission fees from the Kickstarter website, on average, margins were around 45.8% (small bags had a very low margin, suggesting a misprice), production cost were around 28.3% and shipping cost were around 25.8%. We can notice a difference between the first supply chain were shipping cost represent 15.3% of the final retail price and the second supply chain where they represent 26.4%. Inside the second supply chain, we notice differences as well since the jewelry items that are quite light benefit from the economies of scale from the bulk order: shipping cost represent only 5% or 8.6% of the final retail price for these items. Shipping prices are very high for bigger candleholders because they are on average much heavier than the rest of the items (2.6 kg vs 0.8 kg on average). 5.2.2. Main challenges Mainly delays appeared because we faced technical difficulties when customizing products by engraving initials of the customers. We also needed to wait for artisans to send their items from different regions of the country to a consolidation point where the community manager was based. As for rugs which were sent directly to customers, the process was much smoother and clients were generally satisfied. The goal of the campaign was also to better assess the customer's tastes with regards to crafts, artisans' stories and shipping/delivery preferences. People who received their items quickly were extremely happy while answers were mixed with people receiving their items with a higher lead time. This will encourage us to focus on a few products that will be offered in the crowdfunding section (rugs and jewelry), while we will build some inventory for smaller and cheaper items. Evaluating Social Network Impact in the MoroCrafts Hybrid Platform 6. Evaluating Social Network Impact in the MoroCrafts Hybrid Platform 6.1. Introduction to Marketing and social networks The worldwide home furnishing industry is expected to reach $700 billion by 2015, according to research figures from Global Industry Analysts in 2014. Home improvement has always been an area consumers have showed a need, interest, and passion for but new marketing/information channels coupled with the growth of the interior design industry and a move towards sustainability and socially responsible consuming habits have made residential and industrial spaces remodeling more popular than ever. Some estimates from the same Global Industry Analysts Report have measured the home decoration market to be approximately $65 billion each year in the USA. According to the customer research and hypothesis validation we have done via our proof-of-concept on Kickstarter, we are targeting the 60 million travel and tech savvy American women ages between 25 and 65. The Kickstarter campaign confirmed this finding since 76% of our backers were women. The only way we had to track the customers visiting the campaign was thanks to the number of people watching the presentation video on the landing page. Hence we decided to use this action (watching the video) as a proxy to quantify the conversion rate CR with the following formula CR=Number of buyers/Number of people watching the video. The final conversion rate was 12.4% (91 buyers over 730 people who watched the video). Those demographics are in-line with the Pinterest user base and that is why our design and user-experience is set up to mimic the Pinterest user experience to achieve the users' connection and growth we are targeting. We used several methods to reach that audience via: PR (articles on the MIT News Office, Morocco World News), bloggers (Nomadic Decorator, Twitter posts...), and targeted Facebook ads. We spent around $90 on a Facebook ad campaign and acquired 3 customers, which sets up the cost of customer acquisition at 90/3= $30. Each customer spent an average of $120 per purchase, which was more than the average amount spent by a customer on the website ($53). 6.2. The Social network experiments 6.2.1. The most efficient strategy As mentioned earlier, there are 3 ways to engage with fellow users on Twitter apart from posting the original status: "following" another user, "re-tweeting" a post, or "mentioning" the other account name and replying to a post. In this experiment we will investigate the efficacy of these different strategies and measure the incremental benefits of combining 2 strategies together. We will use three items: 1) Six different Twitter accounts with Twitter applications - OmarCobaga: a male artisan producing candleholders and teapots. - AzizCobaga: a male artisan producing candleholders and teapots. - HamidCobaga: a male artisan producing candleholders and teapots. - ChafikCobaga: a male artisan producing candleholders and teapots. - AichaCobaga: a female artisan producing rugs. - JamilaCobaga: a female artisan producing rugs. 2) One text and pictures bank of original content about artisans products 3) One text bank with supportive/directive phrases to us in mentions, like, "nice post, etc." We also have five major controls to replicate a natural human behavior: 1) Robots will tweet at random times with inter-arrival times between robot actions drawn from an exponential distribution with rate 5.333 hours per action, i.e. each bot should act, on average 3 times during the time it is awake. 2) Robots will all have similar names, screen names, and user profile descriptions about artisans' products 3) Robots will sleep from 11 pm EST to 6 am EST 4) Different robots will not interact with the same user or with each other 5) One robot will have no interaction with other users on Twitter, posting original content exclusively: control robot which constitutes our base scenario For each robot, we define a valid tweet as one that has not been used as a base for a retweet or mention by it or one of the other robots. For each robot, we define a valid user as one that has not interacted with any of the other robots. We will use eleven major steps within our experiment: 1) At 6 am EST, each Robot or Artisan will query the Twitter API for search terms such as #morocco #Africa #ryad #social #impact #travel #craft #crafts #art #artisan #artisanal #handmade #jewelry #teapot #silver #rug #rugs #bracelet #necklace #candle #candleholder to target the relevant tweets and people. At the end of each day, the number of followers is recorded and we determine which strategy is the most efficient to advertise the platform to the target audience in terms of number of followers acquired. The experiment will run for 1 week. 2) For each tweet found, the Robot will pull the user description for the user that posted the tweet. If the description contains the strings described previously, it will save the tweet, else, it will discard. 3) The "original-content bot" JamilaCobaga will pull a tweet from its text bank and tweet. 4) The "follower-bot" OmarCobaga will check the first tweet it has saved. If it is from a valid user, it will follow that user. Else, it will move to the next tweet. It will move through the saved tweets until it has followed 10 users. 5) The "retweet-bot" AichaCobaga will check the first tweet it has saved. If it is a valid tweet from a valid user, it will retweet the post. Else, it will move to the next saved tweet and check. It will move through the saved tweets until it has retweeted 10 times. 6) The "retweet-and-follow-bot" ChafikCobaga will check the first tweet it has saved. If it's a valid tweet from a valid user, it will retweet the post. If it is not currently following the posting user, it will follow him. Else, it will move to the next saved tweet and check. It will move through the saved tweets until it has retweeted 10 times. 7) The "mention-bot" AzizCobaga will check the first tweet it has saved. If it is a valid tweet from a valid user, it will reply to this tweet, adding a comment from the text bank of replies. Else, it will move to the next saved tweet and check. It will move through the saved tweets until it has made 10 replies. 8) The "mention-and-follow-bot" HamidCobaga will check the first tweet it has saved. If it is a valid tweet from a valid user, it will reply to this tweet, adding a comment from the text bank of replies. If it is not currently following the posting user, it will follow him. Else, it will move to the next saved tweet and check. It will move through the saved tweets until it has made 10 replies. 9) If it is before 11 pm EST, upon tweeting, each bot will pull random tweet inter-arrival time from its exponential distribution. It will sleep for this many minutes and repeat its tweet process. 10) At each hour, we will pull the timeline for each bot and save followers count, total number of mentions, and total number of retweets. 11) If it is after 11 pm EST, it will clear its saved tweets queried from the API and sleep until 6 am EST. The same approach can be used for Instagram (manual) and Facebook (algorithm). Table 3 shows a description of the different accounts used as well as the different actions performed. The comparison between Omar, Aziz, Jamila and Aicha accounts will help us determine which strategy is more efficient: posting pictures/original content, following, mentioning tweets or retweeting. Second, the comparison between Chafik and Hamid accounts will help us identify if combining 2 strategies will help increase the strategies performance and quantify it as well. It will also give some information about the influence of the gender on the engagement of potential customers. Name Twitter handler URL Password Gender Age Craft produced Jamila @JamilaCobaga @OmarCobaga @AzizCobaga @AichaCobaga @HamidCobaga @ChafikCobaga https://twitter.com/JamilaCobaga https://twitter.com/OmarCobaga https://twitter.com/AzizCobaga https://twitter.com/AichaCobaga https://twitter.com/HamidCobaga https://twitter.com/ChafikCobaga 123Qwaszx/ 123Qwaszx/ 123Qwaszx/ 123Qwaszx/ 123Qwaszx/ 123Qwaszx/ Female Male Male Female Male Male Omar Aziz Aicha Hamid Chafik 31 31 42 27 32 32 Married Children Action on Twitter Rugs Candleholders and lanterns Teapots and lanterns Rugs Teapots and lanterns Teapots and lanterns No Yes Yes No Yes Yes 0 2 3 0 3 3 posts pictures/original content Follows Replies Retweets Replies and retweets Follows and retweets Table 3 Description of the accounts and actions performed during the social network experiment on Twitter After 1 week, we obtained the following results compiled in Table 4: Number of actions Number of Retweets Number of Favorites Number of Friends Number of Followers Number of Mentions 0 2 0 0 0 Original Jamila 11 FollowOmar 19 0 0 147 23 1 Reweet Alcha 16 11 24 21 0 0 0 0 0 0 0 7 2 195 3 22 8 24 168 17 5 Reply Aziz Reply and follow Hamid Retweet and follow Chafik 0 10 0 Table 4 Results for social experiment 1: most efficient strategy - The original strategy was the worst acquiring only 2 followers. - The retweet strategy is better than the reply strategy as it acquired 7 followers and 2 mentions vs 3 followers and 8 mentions. - The simple follow strategy acquired 147 friends, 23 followers and was mentioned once. - The reply and follow strategy acquired 195 friends, 22 followers and was mentioned 24 times whereas the retweet and follow performed actually less than the simple follow in terms of followers, but not in terms of mentions. We do not really see an incremental change if we combine reply and follow in terms of followers but the strategy actually increases the numbers of mentions: this is the most efficient strategy if one wants to achieve a maximum number of followers as well maximize the number of mentions. We also used the data from our experiment to build a logistic regression model that gives the probability that a user follows the Twitter bot as a function of seven features: the number of followers the user has, the number of friends the user has, the number of followers the bot has, the number of friends the bot has, and the way the bot engaged the user (follow, retweet, reply, etc.). We used a logarithmic transformation of the friend and follower counts. The variables that described the type of engagement (follow, retweet, and mention) were all binary indicator variables. We define Xi as the set of features of user i and the bot during the engagement and define the probability that i follows the bot as p(Xi). We assume that p(Xi) has the following form: 7 log - x)-= lg p(X) 7 k= 1 Xik. We performed a logistic regression on our data and show the results in Table 5. Feature Intercept log(User Friend Count) Follow log(User Follower Count) log(Bot Follower Count) Retweet log(Bot Friend Count) Reply 8 -10.9644*** 3.4066*** 1.6365* -1.9575*** 1.4605 -0.7689 p-value 1.31e-08 6.28e-10 0.0211 1.03e-05 0.2883 0.1096 0.2954 -0.9330* 0.7634 0.0238 Table 5 Logistics regression analysis The significance codes are ***: 0.001, * : 0.01, and *: 0.05 The discussion about the regression will be given in the section 7. 6.2.2. Causal relationship between gender and engagement on social media Once the successful strategy is identified, we used it to determine if there was a causal effect between the gender of the artisan and the number of followers acquired. This is to test the hypothesis if the customer base which is interested in crafts is more likely to buy a craft from a female or a male artisans. We ran the same experiment for one more week by a male artisan and female artisan (Hamid and Jamila) using the reply and follow strategy. After another week of experiments, we obtained the following results compiled in Table 6: Number of actions Reply and followJamila Reply and follow Hamid 513 587 Numberof Retweets Number of Favorites 86 81 Iadditional 0 0 Number of Friends 513 550 Number of additional Number of Mentions Followers 72 99 I Table 6 Results for social experiment 2: influence of artisan gender 21 37 Overall, the results in terms of engagement are quite the same and slightly higher for Hamid who obtained more followers and mentions than Jamila. Discussion 7. Discussion 7.1. Scoring of the different models Thanks to the information collected in the previous sections, we gained a better understanding of the different models in place and we could score the different models based on 6 criteria we agreed on in the methodology section. Each model received a score of 0 (no sign), 1 (one X) or 2 (two X). The following table sums up the different findings (Figure 14). x, x Reasonable xx xx xx Margin x * xx xx cfe Emotiona link and Trust x Supply Chain Lnagemt Lead time xx IM xx xx x xx xx xx x x x Figure 14 Scoring of the different models It appears that the Hybrid model we called "MoroCrafts" achieved the highest scores based on the 6 criteria we set up. The main disadvantage of this model is the lead time that can take up to 2 months based on the proof-of-concept we implemented on Kickstarter. The eCommerce and retail fair Trade model developed by Ten Thousand Villages arrives second with only a weakness in margins given to artisans (due to additional inventory held in warehouses) and emotional link conveyed through the online e-Commerce platform (less time spent on the website to know the culture and artisans story). The crowdfunding model arrives third with a lack of access for African artists, lack of control in the supply chain and longer lead times. The marketplace model is ranked number 4 at the same place as regular eCommerce platforms such as Amazon. While e-Commerce mainly focuses on the satisfaction of the customer, the Marketplace focuses more on the supplier: they do not focus on both sides of the supply chain. 7.2. Selection of the right supply chain As seen in the previous section, the direct supply chain with rugs is very efficient and transportation cost represent only 15% of the total price. High value items such as rugs or silver jewelry could benefit from such pull system using the crowdfunding feature. Reducing the lead time could enable the platform to achieve a perfect score. On the other hand, lower value and higher weight item are not very suitable for such a supply chain and not even for a pull system as customers could get unsatisfied because of longer lead times. What we suggest is to segment products and use the direct airfreight supply chain and a pull system for high value items such as rugs and silver jewelry, while we could use a push system building inventory and using ocean freight for lower value items such as candleholders or lanterns. 7.3. Relevance of Facebook Ads We have seen in the previous section that by spending $90 on a Facebook ad campaign we acquired 3 customers, which sets up the cost of customer acquisition at $30. Each customer spent an average of $120 per purchase, which was more than the average amount spent by a customer on the website (around $53). With a 45.8% margin on average per purchase, the net profit for the strategy with the Facebook campaign was $25 while the profit for a regular purchase is $24.3. The Facebook ad campaign does not bring a high added value in terms of profits, but increases revenues. We might need to calculate the retention rate for such customers as well how much they spend over the long term during the next campaigns or when the platform is launched and generalize it to a larger number of customers by allocating more budget to get significant results. 7.4. Best strategy on social networks for social engagement Based on the number of followers we attracted, the best single strategy is clearly the follow strategy. It outperforms the reply strategy by 20 more followers and the retweet strategy by 16 more followers (almost the same number of actions performed: 20% more or less actions performed). The benefit of adding a retweet or reply action does not add more followers: this may be attributed to some noise since the sample is not big enough. This might also be attributed to the lead time not taken into account in the compilation of results since people react later to a friend request. However, what is interesting to notice is that the reply strategy engages people much more than the retweet one since the messages posted by the artisan have been retweeted 10 times (it did not happen for any of the other strategies). In addition, the artisan was mentioned 24 times, which is 3 times higher than we had with the single reply strategy. This was achieved by performing only twice the number of actions. The retweet and follow strategy is in line with the other follow strategies in terms of followers attracted , but it does not necessarily engage them more in the conversation (5 mentions). The results of the logistic regression from Table 5 indicate the user's friend count, the user's follower count, and the actions of replying and following are significant, with the user's friend count and the user's follower count being the most significant. The follow probability is increasing in the number of friends of the user and decreasing in the number of followers of the user. This makes sense intuitively, as we would suspect less "popular" users on Twitter to follow more readily than "popular" users. The coefficient associated with the action of following illustrates the idea that a user is more likely to follow back if followed. The coefficient for the action of replying also indicates what we described earlier in that replying is not the strongest tactic to gain followers and thus decreases the likelihood that a user will follow back. It only allows people to engage you in a different manner through mentions. Although the follower and the friend counts of the bot are deemed insignificant, we suspect that this is due to an experimental shortcoming of not being able to significantly change the bots' follower and friend counts during our period of experimentation. We plan to see how the significance of these features might change by more drastically changing friend and follower counts of the bots in subsequent experiments. There is definitely more value combining the reply and follow strategy as we can attract both more followers and engage more people in the conversation. Of course, this strategy is riskier than the simple follow and retweet one as we need to upload a predefined set of sentences that should match the content we retweet to other people, but the payoff is bigger in terms of engagement. We might have to test the marginal/incremental benefit of performing 3 actions at the same time in a future experiment: follow, reply and retweet The second gender experiment suggests that followers are equally likely to follow back a female artisan using the follow and reply strategy. The slightly higher results for the male artisan Hamid in terms of engagement (followers acquired, mentions and retweets) could be explained by the fact that this person already had a more followers than Jamila and people trusted him more. In that sense, our experiment was somehow biased, but overall we did not see a real causality effect between the artisan gender and the customer engagement on social networks. This does not confirm the hypothesis that women who represent the main part of our customer base are more likely to engage with similar female artisans. Conclusion 8. Conclusion This final chapter provides an overview of the work covered in this thesis paper. A summary is provided, followed by suggestions for further research. 8.1. Summary The handicraft sector in Morocco represents more than 9.0% of the total nation's GDP and employs 2.3 million people. However, despite their talent and expertise, Morocco's these artisans are struggling to reach markets other than the local regional or national markets. The main distribution models used so far did not optimize the value added for both the artisans and the consumers. The main objective of this project is to assist the Moroccan Ministry of Handicraft and Social Solidarity Economy in answering the following question: can we develop a scoring framework that will help us assess the efficiency of each model from both the artisan's and consumer's perspectives? If such a methodology is possible, will it be possible to implement a new model that can optimize these Key Performance Indicators, and how can we leverage the latest social media techniques to make it more engaging for the customers? The motivation for this thesis was to 1) study and refine the three main existing distribution models according to a framework we developed, 2) evaluate a potential hybrid model allowing to better serve artisans and consumers , and 3) improve the engagement of potential customers thanks to social media strategies. Following a summary of past research, we described the 3 distribution models used to export crafts: retail model, marketplace model and reward-based crowdfunding model. We developed our own based framework to score these models according 6 criteria: 3 criteria on the artisan's side (access to funding, profit margins, and access for artists in developing countries), and 3 criteria on the customer's side (supply chain and logistics management, emotional link and trust, and lead time). None of the models was totally satisfying, so we developed and tested a hybrid crowdfunding model. Based on the proof of concept results we obtained from the campaign we ran on Kickstarter, this model achieved the maximal score and allowed us to select the right supply chain and the right products to reach our preferred customer base. Hence, we designed social media experiments on Twitter to reach potential customers and engage them on a more efficient way. It was found that the best strategy to use on Twitter was to follow and reply to social media users interested in arts and crafts. We could not draw a causal relationship between the artisan gender performing this strategy and the engagement of potential crafts buyers. Based on the research, existing distribution models were successfully modified to come up with a new reward-based crowdfunding platform, directly linking the artist and the customer through an adapted supply chain and a unique user experience allowing the customer to learn more about the country, culture and story of the artisan (Pinterest like platform). It does solve for the funding and margin problems faced by the artisans, as well as the trust and emotion conveyed to the end user/buyer. The finding that Facebook advertising is not necessary to increase profits for the platform is quite interesting. Engaging potential customers by following them and replying to their comments is also insightful as it changes the paradigm that the more a company spends on their marketing budget, the more it attracts potential customers. Finally, based on the successful application of the scoring framework put forward in this thesis, we could apply it to many other areas in the e-Commerce field. A successful, social media strategy can be designed and tailor made to engage-more customers on this new and high growth retail channel. Moreover, recommended refinements to the existing methodologies were proposed. Ultimately, the research in this thesis paper meets the goals set out in the introduction of this paper and achieves the objective: to determine the best platform for artisans to get access to international markets, specifically the American market. 8.2. Further research This thesis applies, for the first time, a scoring framework to assess the efficiency of artisan platforms from the perspective of both the artisan and the consumer. Further opportunities exist to test, extend and improve the methodology applied in this thesis, as follows: - Further evaluate and refine the methodology put forward in this thesis through its application to several other platforms such as furniture e-Commerce stores such as Ikea or Wayfair.... - Extend the scoring measurement methodology by including other criteria such as access to illiterate artisan using forms and shapes, reduced prices for customers ... - Test the incorporation of inventory holding cost in the second supply chain strategy used for lower value items in the context of the hybrid platform - Run the pilot for a longer period of time and with a higher budget for advertising to determine if advertising increases the profits for such hybrid platforms. In addition to that, we also added the social network aspect that is extremely important nowadays to build a strong community of followers around the platform. Further opportunities exist to test, extend and improve the methodology applied in this thesis, as follows: - Extend the experiment to other social network such as Instagram or Pinterest. Assess the incremental value added by combining more than 2 strategies. Run the experiment for a longer period of time and assess how we can mitigate the effects of lead time when potential targets follow you back after a certain period. Use the initial results on customer engagement to model a probabilistic approach on how likely a potential user will follow or mention another one based on inputs such as number of friends already acquired, number of followers, number of common friends, type of action (follow, retweet, reply or favorite) ... Extend and tailor the experiment to other e-Commerce retail channels in order to engage potential customers on social media and increase revenues without having to spend extra money on marketing campaigns. Appendix Customer survey from Ministry study in collaboration with Jola Consumer Survey 01 How frequently do you make home dMcor purchases? 100% 80% 20% D% once ser Ista Om am: a par taos sp Once a teak Several r Once a week gurs a VW MAunwer C30am. tiOn an* a year Le Once Ores S.AM 5C*a Yom 15.32% a anwth 11MSS Onew a mael Q2 What will your next home dMcor purchase be? 40% 0Cos4 UgeFg who Ddear Mlor wvwt Mes Covring taoes Oter Access IAlswertMak IRA^ Decoreti ALcOmOdSf Liglieg Vail Sn 0 FbrCaoirt tes buls Omare DOM. VOWe 43% 2"*0% U - -ii PfNUFr kno 03 Where do you go to find Information about home d6cor products? (Feel free to choose more than one.) 43 W% sUder magn Aneer Clae SCeaI e niS etaraS media Oen IeMl FSVcriM 11e1 Im Or y wtwlviale Naspeaus- e 3.32% Reail On flue -*o Favonte 13A7% 97.W% OeWa eta 't2e,% ings Ralar ftin 10404% Q4 Where do you typically shop for furniture and home dicor? (Feel free to choose more than one.) 100% 90% 40% 20 eme OiS inv FUMf ft Sra USd9? i stare kawe FuftsrgS Aoe nedintt Sre MCaNis lass Do Uhf, Ilme et rea r St 8PAe- ore Gift tore GE 2&A&% aus.. Ins SW Cipaitned Sim tamonle 140% COtt.lo OW Q5 Which social media sites do you use to got inspiration for your home? (Feel free to choose more than one.) '00% 0% FaeSo Pflrs nEbb Bbgam 1elee Gegk fibber Pi Anwer Ch|2e Facen lisper Ptas GergI. Ftbr Q6 What percent of home d6cor products do you buy *n-41no? 100% 40% 20% 0% mmon m --= LaseOfn 10fn25% 29% to -- S0% M% to 75% Mane sa 7n% 1% Asawer Chaie Lasshs '% 33A7% 10% 33,57% 265% 4444% 60% t 751 Mkrs t arl 7&4 Q7 Have you purchased any handmade products In the past 12 months? 10% 40% 0% No 7I 05% Q8 If yes, what type of store did you purchase the handmade product In? (Feel free to choose more than one.) 1IM 40% UMsty lios ftrs b1ig.SP Macor hoarn :M Is stars GU sinus apbi o Sne Mas r u-I 0M. lime CaMMg sm chitims tin ftottalngsto.e Feim3tom -4- '.14% Git a puanat ES 2MIL Ofltxu 15)3%6 2&.01% 01Op Ow Q9 Would you pay more for a handcrafted, artisan-made or fair trade product? 10% 0% Ye bowls = cfirtflus Depuds No kn, Resposes Answer CmuSo yre 47 11% 0egon0 o "We*1S 010 If so, how much more? 4001 UVO% ip eD 10% up to Lp O 20% 100% m oam um s Upeb n% UPtOS% meoe 4iu0% Upto Ik% mre Q1 1 Did the product you purchased have any association with a certain culture or region and was that part of the appeal? 00% 40% Ts N I Nanim Yes N0) Q12 Do you feel that handmade or arsanmade products have more value? 100% a"% A"% Vs* b N.aezams.. Aaawrma ILTUW. Q13 Have you every purchased Moroccaninspired home dbcor or clothing? Biat 40% Yes "a DbmI 2%% 71M-A MOW Q1 4 Does It mafter to you if MoroccanInspired design Is made In Morocco? vYs. aum-my- bNorl adb" iripwbrt mw 04 10% 2W% 23% 40% 60% tw% 7M% S0% 00% 1031 M..mer ca..es ND09 dewd quetY am0 impoom1t fd oei pYs ser Itwn ipar ll 242n to le Q15 Are you interested In leaming more about Morocco, Its culture, Its artisans and its products? sat sot 291% YrSn DWnI'N ma ft Mummer Obkess YOS 2l.l% 42.51% 016 Would you buy from a sit that shipped directly from Morocco rather than from a US internediarylsupplier? 100% 4D% 20%f 0% No Too hammer C~tms Vms t0 IL - 017 What typeslcategories of products would you be likely to buy from Morocco? (Feel free to chose more than one.) 100% 90% 90% 40% 20% Ow. ma. toa Fow -oem2 Furaium CS1q .awwey PameinW aemamt Ms easpewme ammer CielSma 13A% Fteii'mre 6.54% COdhing &IS0 Pamarg cam pducts Pammal cn proeim Q18 If you have purchased Moroccan products, where did you purchase them? (Feel free to chose more than one.) 40% OnilAe hi US rasg Armewr Oh hi NoMceOCau~Sill Cailg Fair Crll lhuso Andliv r 1111111 Isn 70.30% I, ii LO11111 svaMil odo* k A114111 4.93 In Mowco 2-55% Outsde IUS aO rct Mariocco Crt Fur 019 What Is your Impression of Morocco in five words? Hse Pivapoma Gwov bearhed' Soepin V. foo EflMc; 713112014 1245 pfetr uNeoU WICDy I1Mnemee pndsutt PM 7I1321411 00 AM 1131121441 AM Vhid coleassSai Nlot owedy PM 17311=14 1234 713120149:5" a SMlNrat ilh Molmc utek of 1a,6e, WOycolef0l4. Priq9eO And wLC *r Iltier AM 73120l4 :30 AM 713I1)4 1:4 PlSl ard illy efeeeably PIIeC AM 7312014 2:57 Atl awsc, nall. cau"ad. aUdk Vay colont place to live tfiagnuf hoedmee Cdility woO 103M014 pICW 11 732014 7M op4on fave Owd 15OPM 2014 1142 PM 1MM22141020PM 113120190 PM ?WW0149:40 PM MCA, dcy. 00l l 713?11314 (mnutl lmyftnaoc cebluoqataS aelc 8:3 PM 113014214:20 PM 7130(20 7:9PM allot Coamtm, Iltemega uile. Irftfgu oP , ci ms 7AI2)140 Mtily. pael. bntlC~a lleplaaloe318, C010A epeso M12214 9'S3 PM celfluil, lively,vibiout cio&ienhi, cS75, lobroa 2 PM 113012014033S PM and d0*dgr Ri in culkim arn :49 PM 1132014 8A lNdorapolelog Panan Fp :11 f2(1014 1:59 PM cil EMnu, Abicar I nPlly hav e 11:1PM 73020i4 7?30a1140:D2 1M2014 deig n. t u fuel, 73i PM 7:40 PM A2014 7:28 PM Q20 Have you visited Morocco? A8t ao ft vklt Answer -ag, CbcSes VAw I- g- -W NoA WD I" o Vill References Aral, S., Walker D. (2012). Identifying Influential and Susceptible Members of Social Networks in Science Aral, S., Walker D. (2011). Creating Social Contagion through product viral design: Randomized Trial of Peer Influence in Networks in Management Science Aral, S., Muchnik L., Sundarajan A. (2009). Distinguishing influence-based contagion from homophily-driven diffusion in dynamic networks Business Insider. (2015). Why a longtime Etsy seller decided to shut down her store Retrieved from http://www.businessinsider.com/ Colbert, F. (2003). Entrepreneurship and leadership in marketing the arts. International Journal of Arts Management, 6(1), 30-39. Crocker, J. (2003). Increasing value through process improvement. Retrieved November 15th, 2014 from the website: http://www.clermiston.com.au/ Crowdsourcing LLC. Retrieved April 29th, 2015 from the website: http://www.crowdsourcing.org Dave, D., Evans, M. (2001). The CODA Survey: The impact of crafts on the national economy. Retrieved from http://www.craftsreport.com/mayOl/codasurveyresults.html Etsy. (2008). Etsy customers survey results. Retrieved from http://www.etsy.com/ Fischer, M. (1997). What is the right supply chain for your product? Harvard Business Review Global Industry Analysts (2015). Global Home Furnishings Industry Report. Retrieved January 13th, 2015 from the website: http://www.reportlinker.com Hemer, J., (2011). A Shapshot on Microfunding. Fraunhofer Institute for Systems Ibrahim, N., Verliyantina (2012). The Model of Crowdfunding to Support Small and Micro Businesses in Indonesia Through a Web-based platform. Procedia Economics and Finance ( 2012 ) 390 - 397 Lambert, T., Schwienbacher, A., (2010). An Empirical Analysis of Crowdfunding. [Online] http://ssm.com/abstract=1578175 (April 23, 2012) Lambert, D. (2004). The eight essential supply chain management processes. Massolution LLC. Retrieved April 13th, 2015 from the website: http://www.massolution.com Ministry of Handicraft and Social Solidarity Economy. (2007). Vision 2015 pour le secteur de 1'artisanatd fort contenu culturel. Rabat : Morocco Ministry of Handicraft and Social Solidarity Economy. (2012). Rapport annuel de l'Observatoirenationalde l'Artisanat. Rabat : Morocco Ministry of Handicraft and Social Solidarity Economy. (2014). Customer survey for the American HandicraftMarket in collaborationwith Jola. Rabat : Morocco Rabinovish, E., Knemeyer, S. (2006). Logistics Service Providers in Internet Supply chains Stoddard, J. and Evans, R. (2012), Marketing Arts and Crafts: Exploring the Connection between Hedonic Consumption, Stoddard, J., Evans, M., & Dave, D. (2008). Sustainable tourism-the case of the Blue Ridge National Heritage area. Cornell Hospitality Quarterly, 49, 245-257 Trent, R. (2004). What everyone needs. Supply Chain Management Review, 85, 52-59 Wikipedia Kickstarter. (n.d.). Retrieved April 29th, 2014 from the Wikipedia Wiki: http://en.wikipedia.org/wiki/Kickstarter