icarda 2.0 :: Web Strategy
ICARDA Website Strategic Plan 2010-11
http://icarda2.benale.net :: project site
http://beta.icarda.org :: prototype
Version 0.5
Published July 7, 2010
Created by Bernhard Hack, CODIS
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
2 |35
ICARDA Website strategic plan 2010-2011
Table of contents
Change History ............................................................................................................... 3
Roles & responsibilities .................................................................................................. 4
Executive summary ........................................................................................................ 5
1. Introduction ........................................................................................................... 6
1.1. Website objectives .......................................................................................... 6
1.1. Site Mission statement.................................................................................... 7
1.2. Site goals ......................................................................................................... 7
1.3. Design Objectives ............................................................................................ 7
1.4. Project scope ................................................................................................... 8
2. User research ......................................................................................................... 9
2.1. Heuristic evaluation ........................................................................................ 9
2.2. Website statistics .......................................................................................... 12
2.3. Audience Analysis.......................................................................................... 13
2.4. User Profiles .................................................................................................. 13
3. Design Vision ........................................................................................................ 22
3.1. Functionality .................................................................................................. 22
3.2. Design ............................................................................................................ 23
3.3. Site concept ................................................................................................... 23
3.4. Compliance .................................................................................................... 24
3.5. High level site migration strategy of current sites ........................................ 24
4. Content Management .......................................................................................... 25
4.1. Rules .............................................................................................................. 25
4.2. Roles .............................................................................................................. 25
4.2.1. Functionality by roles ................................................................................ 26
4.3. Resources ...................................................................................................... 26
4.4. Example publishing process: Report ............................................................. 27
4.5. Content Management System ...................................................................... 29
4.6. Templates ...................................................................................................... 31
4.7. Metadata ....................................................................................................... 33
4.8. Content chunking .......................................................................................... 33
4.9. Search ............................................................................................................ 34
4.10.
Disaster Recovery ...................................................................................... 34
5. Glossary ................................................................................................................ 35
6. Timeline................................................................................................................ 35
7. Budget .................................................................................................................. 35
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
Change History
Version 0.1 :: January 24, 2010

Draft Project plan developed
Version 0.2 :: January 31, 2010

Draft Project plan revised

Edited Design goals

Added design objectives

Edited Reviewers
Version 0.3 :: February 24, 2010

Site vision and mission added

Design objectives revised

Results of content analysis added

Content Management drafted

Project Plan revised
Version 0.4 :: March 29, 2010

Added glossary

Updated roles & responsibilities matrix

Updated timeline
Version 0.5 :: July 7, 2010

Updated User Research, Design Vision, Content management

Updated timeline
3 |35
Roles & responsibilities
Reviewer
Core Team
Bernhard Hack
Manaf Hamam
Hazem Fattal
Ammar Halabi
Ahmad Al-Mously
Siba Darouzi
Email Address
Group
Role
Responsibilities
b.hack@cgiar.org
m.hamam@cgiar.org
h.fattal@cgiar.org
ammar.halabi@gmail.com
a.al-mously@cgiar.org
s.darouzi@cgiar.org
CODIS
CODIS
CODIS
CODIS
ITU
CODIS
Project manager
Webmaster
Assistant webmaster
Interaction designer
Intranet developer
Doc. repository advisor
Coordination, liaison, IA, UX
IA, interface design
Social media maintenance, event organizing
IA, interface & interaction design, user research, testing, analysis
Integration with intranet, user interface design templates
Ensure library integration
Extended Team
Dr Kamil Shideed
Dr. Zaid Abdul-Hadi
Colin Webster
k.shideed@cgiar.org
z.abdul-hadi@cgiar.org
c.webster@cgiar.org
Management
CBSU
CBSU
Project sponsor
Strategic advisor
Technology advisor
Advocate for the project in MC meetings, sign-off
Follow project documentation and give advice
Follow project documentation and give advice
Advisory WG
Chris Addison
Johannes Keizer
Valeria Pesce
Enrica Porcari
c.addison@cgiar.org
Johannes.Keizer@fao.org
Valeria.Pesce@fao.org
e.porcari@cgiar.org
IFPRI
FAO
FAO
ICT-KM
Web process advisor
Drupal advisor
AgriDrupal developer
ICT – KM advisor
Follow project documentation and give advice
Follow project documentation and give advice
Follow project documentation and give advice
Follow project documentation and give advice
Executive summary
ICARDA’s website requires re-development based on formative user research and business
requirements. In an effort to increase the impact of ICARDA’s knowledge products, the team looks to
improve the overall usability, look and feel of www.icarda.org. By designing a dynamic website based
on a cost-effective content management system overall usability will be improved.
Design Objectives
 Refresh homepage design
 Enable dynamic reuse of content by implementing a content management
system
 Develop templates to improve information design and distribute authorship
 Improve information architecture and navigation
 Enable online collaboration
 Improve and refine search
 Improve accessibility of website
Project Scope
 Research context, content and users
 Develop new information architecture
 Write web strategy
 Redesign Homepage
 Implement Content Management System
 Implement search function
 Bring www.icarda.org up to web 2.0 standards
 Build private collaboration space
 Rewrite and migrate content to new site
 Design blueprints and wireframes
 Develop 20 page templates (see Content management for details)
 Built and test prototype
 Provide training for use of CMS
Budget
Total project cost estimate
US$ 86,645
Requested from ICARDA
US$ 28,985
Of which for training U$ 13,740
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
6 |35
1. Introduction
Purpose of this document
The purpose of this document is to present a vision for ICARDA’s new website and to illustrate and
explain that vision to our stakeholders. It is intended as a bridge from the initial research phase to
actual design and to inform the conceptual design of icarda.org.
Target Audience
The intended audience for this documentation include:
 Senior Management: Input on objectives and priorities, stay abreast of project progress, and
sign-off.
 Design team: Confirm and collaborate on consensus design solution across roles.
 Focal Points: Form as a network in the process to establish continuous feedback loop, input
on design, provide content
1.1.
Website objectives
The website’s purpose needs to be in line with the overall vision and mission of
ICARDA and should, therefore, be derived from ICARDA’s business objectives.
ICARDA’s mission statement reads:
“To contribute to the improvement of livelihoods of the resource-poor in dry
areas by enhancing food security and alleviating poverty through research
and partnerships to achieve sustainable increases in agricultural productivity
and income, while ensuring the efficient and more equitable use and
conservation of natural resources (ICARDA’s Strategic Plan 2010-12).
ICARDA's mission is to improve the welfare of poor people and alleviate poverty
through research and training in dry areas of the developing world, by increasing the
production, productivity and nutritional quality of food, while preserving and
enhancing the natural resource base.
The environment in dry areas is harsh, stressful and variable, and agriculture in these
areas faces more complex challenges than in areas with adequate rainfall.
ICARDA is committed to the advancement of agricultural research; free exchange of
germplasm and information for research; protection of intellectual property rights,
including indigenous knowledge of farmers; human resources development; the
sustainable use of natural resources; and poverty alleviation, particularly among
women and children.
ICARDA pursues this mission in partnerships with national agricultural research
systems in developing countries and with advanced research institutes in
industrialized countries. The website’s mission needs to support ICARDA’s overall
mission:
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
1.1.
7 |35
Site Mission statement
www.icarda.org will be the go-to place for dryland agriculture information on the
Internet. As the dominant thought leader on dryland agricultural research on the
Internet, www.icarda.org will provide relevant, up-to-date information about
dryland agriculture to researchers, students, media, donors and any other users.
The primary focus of the site is to provide access to ICARDA's research ouputs and
knowledge. By more effectively communicating ICARDA's knowledge,
www.icarda.org will increase ICARDA's impact and, thus, contribute to improved
livelihoods of the resource-poor in the dry areas. The website will employ technology
that effectively enables conversations with ICARDA's key stakeholders, keeping them
informed and involved.
1.2.
Site goals

Improve ICARDA’s external presence: Establish the ICARDA brand as the
dominant thought leader on dryland agricultural research.

Ensure transparency to stakeholders: Open a window on ICARDA’s research
impact and demonstrate effective use of resources.

Increase usability: Provide effective access to ICARDA’s research outputs and
knowledge

Facilitate stakeholder interaction: Enable conversations and collaboration
with key stakeholders
1.3.
Design Objectives
Refresh homepage design
Updating the homepage design to incorporate recent trends and improvements in
web design will improve ICARDA’s web presence and reflect its standing in the world
as a thought leader in its field.
Enable dynamic reuse of content
By implementing a content management system (CMS) the site will move from static
to dynamic. Systematic reuse of content will reduce maintenance costs and improve
accuracy and consistency of content.
Develop templates to improve information design and distribute authorship
Templates for different content types will be developed to improve information
design at page level – layout, readability and look. Templates will also simplify the
publication process by decentralising content production while at the same time
maintaining a central editorial capacity.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
8 |35
Improve information architecture and navigation
We need a more coherent information architecture, providing information in
appropriate groupings and contexts and ensuring that information is easy to find for
a variety of users. An improved information architecture will support a simpler, more
usable navigation.
Enable online collaboration
Providing a seamless log-in area for researchers and other key stakeholders that is
private and features online collaboration tools such as wikis and blogs.
Improve and refine search
Improving the search facility to provide better results will assist users of the website
to find the information they require. Providing search results for different content
types such as publications, projects, news, and strategic documents will assist users.
Improve accessibility of website
Designing for accessibility issues:

Providing for low bandwidth users and other user segments

Offering an Arabic version of the information architecture

Cater to the increasing use of mobile devices

Complying with the Web Accessibility Initiative standard
1.4.
Project scope
The scope of the project is limited to these requirements. The implementation of
these requirements will be phased (see Project plan for details).













Research context, content and users
Develop new information architecture
Write web strategy
Redesign Homepage
Implement Content Management System
Implement search function
Bring www.icarda.org up to web 2.0 standards
Build private collaboration space
Migrate content to new site
Design blueprints and wireframes
Develop 20 page templates (see Content management for details)
Built and test prototype
Provide training for use of CMS
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
9 |35
2. User research
2.1.
Heuristic evaluation
The London-based web publishing company Green Inc. conducted an informal
heuristic assessment of www.icarda.org. Seven web designers evaluated ICARDA’s
website according to a set of heuristics they chose themselves, as part of a job
application procedure at Green Inc. The evaluators where asked to answer these
questions:
 Does the website communicate well?
 Does it, for example, make it easy for you to gain basic information about
ICARDA?
 What impressions does it give you of the organization?
 What advice would you give this client as to how to improve the website?
The heuristics that were used to assess the site can be summarised as navigation,
ease of use, content, design & layout and functionality. Below is a summary of the
more salient points raised:
Navigation
Navigation was described as clumsy and inconsistent and it was evaluated negatively
throughout. Concrete points raised were too much duplication, left hand navigation
bar is too long and complex, the Arabic menu is inconsistent with the English menu,
menu headings are not executed accurately (donors page not for donors but list of
donors, ICARDA in the news is actually media releases not selection of articles
published about ICARDA, new publications presents outdated results)
Recommendations
 Simplify information architecture
 Unify navigation
 Use drop down menus
 Reorganize content to create synergies and KM exchange opportunites with
fellow CGIAR centres
 Add category called “Our partners”
 Add small menu bar on the bottom: site map, privacy policy, terms of use
“It (the website) does not present the company as a dynamic progressive
research focused unit with strong academic leanings but rather, as an old
fashioned and inward-looking organization that is yet to embrace modern
technology.”
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
10 |35
Ease of use
Most assessments underlined that the homepage is too busy and cluttered and that
it is extremely difficult to find information. The website generally does not
communicate well.
Recommendations:
 Reduce homepage content by simplifying added content
 Remove redundant duplicate functions
 Highlight only latest events and media alerts as ‘teasers’ on front page
Content
Content is generally considered the actual strength of the site. The site is seen as rich
in resources, offering a wealth of information resources and that this is a good basis
to work with. The site is also seen as seeking to be exhaustive and encyclopaedic in
scope.
On the negative side, it is considered text heavy and too wordy, driving users away
quickly. The About page does not actually explain what ICARDA does and although
the front-page is heavily focused on international cooperation, it is unclear if these
are projects, general partnerships or donors.
Recommendations:
 Rewrite content according to web writing standards, more concise, crisp
 Reduce impression of complexity with a more simplistic approach
 Emphasize more effectively resources, projects and other timely information
“Gives impression of an organization that is a composite of many different
parts which are not coherent or even communicating with each other.”
Design & layout
Several evaluations pointed out that the design of the main and sub-pages is
incohesive. It was seen as unattractive, cluttered and too busy, weighed down by
typographical confusion, an overloaded front page, and a wildly varying look and feel
of pages.
Recommendations
 More consistency throughout
 Enlarge site to make use of white space left and right
 Use modern font families
 Settle on a one colour scheme
 Increase overall focus on branding
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
11 |35
Functionality
Here the absence of almost all web2.0 features was noted. All noted that the Search
function does not work.
Recommendations
 Integrate social media: online forums, twitter, delicious, online video and
image sharing, podcasts
 Set up RSS tab and alert system
 Add online stakeholder section / log-in areas
 Streamline the three news sections
 Repair search function
“ICARDA site and organization would benefit greatly from a radical re-design to
simplify both the look and usability. This would help clarify the organization’s
goals, integrate the site with overall marketing efforts and corporate image, aid
research, promote activities and enterprises, attract more users and, by changing
onsite commerce practices, even increase income.”
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
2.2.
12 |35
Website statistics
As a benchmarking exercise for a more general baseline we have identified the key statistical indicators listed
below. Tracking and monitoring this set of data on a monthly basis will show what quantitative changes the
redesign effected. Two types of data are of interest. The server logs tracking site visits and the search logs
tracking search queries entered in e.g. Google that brought visitors to the site. A simple way of gaining a quick
overview of the most important search terms is to display them as a search cloud, where size reflects
popularity.
Web-statrs for www.icarda.org (Source of data: Webusage CGnet)
Two key numbers 2009
Pages/visit
Time/visit (over 75% of visitors leave site in 30s or less)
Server log 23 Feb 2010
Hits
Entire Site (Successful)
Average Per Day
Home Page
Page Views
Page Views
Average Per Day
Document Views
Visitor Sessions
Visitor Sessions
Average Per Day
Average Visitor Session Length
International Visitor Sessions
Visitor Sessions of Unknown Origin
Visitor Sessions from United States
Visitors
Unique Visitors
Visitors Who Visited Once
Visitors Who Visited More Than Once
Search cloud
3.4
264 s
3,724,736
7,390
44,678
2,974,421
5,901
2,954,689
954,397
1,893
00:14:49
17.73%
30.75%
51.51%
256,944
207,887
49,057
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
2.3.
13 |35
Audience Analysis
To better understand our users’ information needs, we have done some research
with the aim to establish profiles. First we did some statistical analysis of the search
queries and the server logs. These are to be followed with stakeholder interviews, an
online survey and card sorting exercises. At this point in time these are ongoing.
Additional information was taken from the recent needs analysis on Communication
and Information Management conducted by Niche Consultants.
Key stakeholders of www.icarda.org include:
 ICARDA scientists
 ICARDA staff
 Outreach and regional staff
 NARS & Extensions Specialists
 Media
 Board members
 Donors
 Students
 CGIAR and wider scientific community
 Policy makers (government and others)
 Dry land farmers
 General public
Of these stakeholders, three key user groups were chosen as the main target
audiences of the website
1 Researchers
 ICARDA scientists
 Outreach and regional staff
 NARS & Extensions Specialists
 CGIAR and wider scientific community
2 Media
3 Impact Seekers
 Donors
 Policy Makers
2.4.
User Profiles
Profiling each of the key user groups improves our chances to target their
information needs and priorities. Additionally the design team will benefit from a
clearer understanding of who they are designing for.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
14 |35
Researchers
The needs analysis indentified the following web-related priorities for researchers,
NARS, outreach offices and the wider scientific community within and beyond the
CGIAR:
 Reduce email traffic
 Take advantage of Web 2.0 tools, such as blogs, Facebook, Twitter, etc
 Improve findability of documents and data
 They would value more e-communication
 They need brief, targeted, timely information
 Clearing house: Facilitate the sharing of ICARDA and NARS information
 Would appreciate a system to feed in stories to CODIS
 Help with media liaison and increased visibility
 Regional representation on the website is inadequate and needs to be
improved
 Researcher profiles with accurate contact data
 Improved internal communication and improved response rates
ICARDA’s information ecology
The diagram below shows a typical information ecology of a researcher in a medium
sized organization. The size of the cloud indicates the importance of the information
resource for a typical user.
The diagram shows that people view personal networks and colleagues as the most
important information resources, and see the current website and intra-net as
having relatively little value in their work lives. The diagram also presents a
fragmented information environment, in which artificial boundaries of technology
(media, format) or geography exist between pools of information (Information
architecture for the word wide web. p278).
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
15 |35
Media
Ongoing need for stories with a human face which requires readily available news
releases, a resource section, images and events calendar Journalists want an expert
guide that helps them locate ICARDA scientists. They like personal contact and
ongoing relationships that make work easier fro them.
Donors & Policy Makers
Funders are a critical stakeholder for ICARDA, and they expect an improvement in
how the organisation demonstrates the translation of research outputs into impacts
– expressed as measurable outcomes for the poor (Niche report: 3).
The needs analysis identified the following web-related priorities for donors and
policy makers:
 Compatibility with their particular strategic priorities, programs and projects
 Measurable return on investment – they value impact assessments, benefits
to poor, addressing strategic global issues
 Mutually beneficial partnerships
 Greater responsiveness and flexibility from ICARDA in responding to their
priorities
 Well-written, clear and current information about ICARDA and outcomes of
its research and development
 Targeted, relevant, impact-based reporting
 There is a need for “results”: good quality research and development
outcomes at value for money transaction costs
 Specifically want prompt, brief, non-scientific reports which specify outcomes
of research and development impacts
 At conclusion of a project, a final report that succinctly synthesises all
relevant outcomes
 The website should play an important role in providing current information
Method: Personas & Scenarios
Personas and scenarios are techniques for representing users and the things they do on the website.
Personas are fictitious characters that are created to represent a site or product's different user types
and their associated demographics and technographics. They are archetypes that represent actual
groups of users and their needs and they play out several scenarios using the site to fulfil their
information needs.
The technique turns user research data on key stakeholder groups into a few physical concepts of
users to create empathy among the design team, with the final aim of tailoring a product more closely
to how the personas will use it. Our persons are based on the data from the interviews, the online
survey and the statistical analysis of our search and server logs.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
16 |35
Persona 1: Hazel Purcell (Donor)
Hazel Purcell
Basic Information
Age
49
Occupation
Senior Grant Officer, responsible for CWANA and SE Asia, IDRC’s
Cairo office
Net usage
Always on, lives in her inbox.
Gear
Standard issue laptop that comes along wherever she goes. Nokia
smartphone, used mainly for calls, SMS, email only on the road.
Familiarity/Anxiety
Hazel has been using MS office applications all her life and knows
Excel inside out. If there’s a problem, she knows where to get tech
support.
Trigger for action
It’s the end of the year and she still has to push some funds out the
door.
Ultimate Goal
Get the grant to the right people. Keep the budget from last year.
People needed to interact
with
Project Managers, Directors, donors to her foundation
How Hazel Works
Business situation
Hazel is the boss of her department, she has 5 staff directly reporting
to her.
Working environment
Own office in busy downtown area. Sometimes limited connectivity.
Often travels to conferences and meetings.
Preferences for Interacting
with Others
Hazel is drowning in email and when it is really important she knows
calling gets you answers faster.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
17 |35
Persona 2: Malcolm Dandry (Journalist)
Malcolm Dandry
Basic Information
Age
36
Occupation
Radio and TV journalist for the BBC stationed in Dubai for the last 5
years.
Net usage
Online all day most days unless he is out filming interviews. For
email, as well as research for work. Gets online routinely while on the
road.
Gear
Latest net-book for travel, high end Acer latptop with docking station
in the office. Blackberry for email, addresses, notes, calendaring, and
time tracking.
Familiarity/Anxiety
Malcolm is a tech-savvy media professional who is comfortable with
computers, cameras and the web. He maintains a blog, is active on
various social media platforms and synchs most of his data up to the
cloud. He can do most of his own trouble shooting.
Trigger for action
The headquarter needs a story for the series about the drought they
were doing recently.
Ultimate Goal
Get that story out to inform the world, and then get paid.
People needed to interact
with
Story leads, communication professionals, interviewees, wiremen,
boss, colleagues
How Malcolm Works
Business situation
Part of a team but travels and works independently a lot.
Working environment
Half the time he sits in a busy newsroom in a glass and steel tower
opposite the Hilton where he’s got band-with to stream HD movies.
The other half he is on the road and stays in Four Star hotels.
Preferences for Interacting
with Others
Malcolm sends an email ahead and then usually picks up the phone.
He is a highly skilled interviewer with good listening skills who knows
when to ask the right questions.
Persona 3: Dr Alonso Giratesta (Researcher)
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
18 |35
Dr. Alonso Giratesta
Basic Information
Age
54
Occupation
Alpaca Sheep Breeder at ILRI
Net usage
Online all day most days unless traveling to meetings or conferences.
Gear
Dell laptop, standard issue, automatically backed up to the network.
Nokia Smartphone addresses, calendaring, and pictures, sometimes
email.
Familiarity/Anxiety
Frank has already been programming databases on mainframe
computers. Things need to work, not be beautiful. Uses skype to talk
to his daughter at uni in the US.
Trigger for action
Needs to find at least two more collaborating institutions to apply for
this project grant.
Ultimate Goal
Develop a breed for dual use (wool and meat) and publish it, put his
son and daughter through a good college, feed more poor people
People needed to interact
with
Researcher colleagues, secretary, funders, senior management,
sometimes farmers
How Malcolm Works
Business situation
His department has 12 research staff and two secretaries. He travels
too much, business class if possible.
Working environment
He has a small office full of books and papers. The internet
connection could be better, but we’re getting there.
Preferences for Interacting
with Others
Ideally would not have to interact quite so much and focus more on
research. Face to face meetings is best. Sends emails a lot, though,
makes calls, even conference calls.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
19 |35
Scenario 1: Hazel Purcell (Funder)
Persona:
Scenario:
Hazel Purcell
Finding a project
Hazel is a 49 year old Senior Grant Administrator at IDRC’s Cairo office
where she has been stationed for the last 8 years now.
Hazel has some funds left in a pot earmarked for substituting poppy with
other cash-crops in Afghanistan. The end of the year is near and she needs
to get some project proposals in fast to push the funds out the door, if she
does not want to have her budget cut next year.
She has done business with ICARDA before and althoght there were some
delays, she was satisfied with the delivery and the reporting. She also
knows that ICARDA has excellent working relations with the NARS in the
area.
Hazel googles for ICARDA and finds www.icarda.org as the third search results. She goes to the
homepage, ignores the tab that says donors and goes straight for the search function.
She types in Afghanistan and is offered a range of results from the projects database, several
publications, and two news releases. She clicks on an ongoing goat breeding project in Afghanistan,
reads the 100 word summary and quickly scans the other donors involved.
Then she turns to the staff working on the project and browses their personal profiles in separate
tabs, skipping back to the project page all the time.
As she looks at the pictures of ICARDA researchers, she recognizes Dr Barbara Kramer whom she met
at a conference in Cairo recently. Fortunately the skype icon in Barbara’s profile is green, showing
that she is online right now. Hazel puts on her headphones, clicks the icon and after three rings
Barbara picks up.
Although Barbara is a sheep breeder and cannot use Hazel’s funds for poppy substitution, she knows
of a colleague at ICARDA who has a proposal written and is looking for grants. She gives Hazel the
contact and then checks if her colleague is online. He is and she invites him into the skype
conversation, introducing him to Hazel.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
20 |35
Scenario 2: Malcolm Dandry (Journalist)
Persona:
Scenario:
Malcolm Dandry
Finding a story
Malcolm Dandry is a 36 year old Brit working for the BBC in
Dubai as a TV journalist.
His team is currently producing a series on the impact of the
drought in Central and Western Asia and he needs, both, some
sound facts as well as a human story. Best would be a farmer,
some footage of water harvesting and a couple of interviews in
the village.
Malcolm googles for “drought” and “impact” and finds
www.icarda.org among the top hits. He clicks on the link to the
homepage. First he checks the media releases and finds one
about a recent conference on drought.
The media release links to the conference website where he finds links to PDFs. He downloads those
for later. Then he checks the ICARDA weblog, filtering the entries using the water management tag.
He finds a story that has some good data on ground water depletion in Syria.
He likes the picture of the farmer pulling the bucket up from the ancient cistern in his backyard. He
clicks on it and finds himself in ICARDA’s photo gallery where he discovers a whole range of nice shots
tagged with water management.
He is intrigued by this serendipitous discovery, goes back to the homepage, clicks on the media
contacts tab and calls the media spokes person directly. Liz picks up, they agree on a time and she
arranges interviews with scientists as well as a field trip for him.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
21 |35
Scenario 3: Dr Alonso Giratesta (Researcher)
Persona:
Scenario:
Dr Alonso Giratesta
Finding collaborators
Dr Alonso Giratesta is a 54 year old Italian Alpaca Sheep
Breeder working at ILRI in Addis Abeba.
He is looking to apply for an EU grant for a multi-lateral project
on breeding sheep for dual use – wool and meat. The eligibility
criteria state that he has to have at least 7 advanced research
institutions in as many countries as collaborators.
He has a tried and tested network of colleagues at universities
in Europe and others throughout the CGIAR. He has been
working with ICARDA before and would like to bring the expertise of ICARDA’s scientists to his project.
He also knows ICARDA has recently won funding for similar work and would like help with writing his
proposal.
He calls up his friend Dr Mohammed Benshawi, senior sheep breeder at ICARDA and proposes his
idea. Dr Benshawi likes the project and sends Alonso an email with an invitation to join ICARDA’s
online collaboration area.
Alonso clicks on the link in the email and finds himself in a protected area where Dr Benshawi had
already set up a wiki for them to co-write the project proposal together. Other successful proposals
are stored in the resource section and they find that they can recycle quite a lot of the old content,
just touching it up here and there to make it fit the requirements of the current call. Alonso exports
the proposal to MS Word, downloads it and circulates it to his network so the other institutions can
fill in their parts.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
22 |35
3. Design Vision
3.1.
Functionality
Functionality
Search
Across sites and social media
platforms
Advanced search
Templates
Microsites / subsites
Drupal Module / Product
Google CSE
Content Construction Kit
Differential permissions
Navigation
Drop down menu
Breadcrumbs
Sitemap
Functional footer
Intuitive URLs
Multimedia display
Photo library
movies
slides
Pics
Apture plugin
Commenting
Community / Team sites
Organic groups
Blogs
Wikis
Forums
Polls
Scheduling
Calendar
Repository
Indexing
Taxonomy CSV Import
Searchable databases
Staff pages
Subscription
Project finder
AgriDrupal
AgroTagger / OpenCalais
NoSQL
Feed API
Newsletter
RSS
Social media
Interactive Map
RSS feed from ICT-KM Ongoing
Research
Accessibility
Browser compatibility
Multiple languages
Forms
File up & downloading
WYSIWYG interface
Secure login area
Print pages
Survey and Poll
Mobile version
MS Sharepoint
compatibility
Useful Error 404 page
Latest news
Group access to files
Core
FCKeditor
Drupal core
Core
From all modules
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
3.2.
23 |35
Design
We are deploying a user centred approach to design. However, recognising that findability precedes
usability the architecture of the site aims to provide several parallel pathways to every piece of
content. The guiding idea is to get to every document in three clicks, but we also acknowledge that is
not always possible nor is it even desirable.
Deviating from the original path of outsourcing design we are now using an agile approach to develop
rapid prototypes and test them in quick succession in-house. Each iteration of building and testing
lasts for two weeks, ensuring user feedback and buy in all the way through the design process.
3.3.
Site concept
3.3.1. Content
Most of ICADRD’s original content consists of research outputs which are created on the personal
(researcher) and on the research project level. The site’s aim is to capture thess on that level, i.e.
relating and linking them from the personal profiles as well as from project fact sheets. From there on
they can be aggregated to higher levels in the site such as program profiles, donor pages and finally
the front page.
Each new publication or poster uploaded to the document repository will be treated as a news story.
All this story-contents will the ‘bubble up’ from the personal and project level, flowing thought the
site onto the front page where they will be eventually superseded by newer publications and
announcements.
3.3.2. Taxonomy
The main site taxonomy draws from the strategic plan 2007-16 but also tries to account for the
shifting landscape of the CG system by allowing for retrofitting of the emerging Mega Programs.
Additional taxonomy inputs are drawn from the meta-tag of the original homepage and the meta data
developed for the Research Data Repository, which is a subset of AgroVoc. Proper indexing of
scientific publications is ensured by the library team, at first manually and at later stages via RSS
export.
3.3.3. Navigation
Several approaches to navigation should make the site as accessible as possible to its key users. First
there is role based navigation. Each of the three key stakeholder groups should instantly be able to
discern a way into the site to fulfil their information needs. Secondly there is subject based navigation
organized around research themes and quick links to them. Thirdly, there is the traditional menubased navigation according to standard menu items like ‘Who we are’, ‘Our work’, etc. And finally
search needs to be ‘better than Google’, providing advanced search and faceted results by document
type, date, owner, etc. Additionally globally deployed breadcrumbs and functional footer should aid
overall orientation. Menus should be kept simple, avoiding drop downs and other distracting moving
elements wherever possible.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
Proposed top level navigation for prototype testing
1. Home
Secondary Menu
2. Who we are
1. [ my cart ]
1.1 Mission Mandate etc
2. Staff
1.2 Staff
3. Pressroom
1.3 BOT
4. Careers
1.4 Donors
5. Training
1.5 Strategic documents
6. Events
1.6 Org chart
7. Contact us
1.6.1
Service units
1.7 Visitors guide
2 Our work
2.1 Snapshot
Block: Featured publications
2.2 Topics / Research areas
2.3 Programs
2.4 Research units
2.5 Projects
Block: Publications
Block: Countries
3 Our products
3.1 Publications
Block: Slides, Posters, Video, Audio
3.2 Datasets
3.3 Knowledge Products
3.3.1
Crop information Systems
3.3.2
Training materials
4 News
4.1 Events calendar
4.2 Media releases
4.3 Media contact
4.4 Subscription options (RSS Feeds)
3.4.
24 |35
Footer
1. Sitemap
2. Contact
3. Copyright
4. Breadcrumbs
5. Login
6. Copyright & Fair use
Compliance
To ensure compliance with ethical standards accessibility issues need to be taken into consideration
such ANAO and W3C. Adequate protection of private data needs to be ensured and users must be
given opt-in choices regarding its display, or opt-out as a minimum default standard.
3.5.
High level site migration strategy of current sites
A complete inventory indexing the URL, owner, ROT status (whether it is redundant, outdated or
trivial) of every page of the entire website has been conducted (available on project website). In this
way we will be able to keep track of every piece of content, as it moves from old to new location. In a
next step the focal points will conduct inventories of their subsites, to prepare them for migration and
to test the fit of their content with the existing wireframe templates.
Much of the existing content of the site is outdated and needs rewriting fort the web. This is a lengthy
process and will require some training to develop web-writing skills for staff.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
25 |35
4. Content Management
The point of dynamic content management is to keep content fresh and accurate by
enabling re-use while reducing overall maintenance costs. We envision a process
whereby staff submit content (text, pictures, video, etc) via an online form. This will
then pop up on the editor’s screen, and upon her one-click- approval the new
content appears in the dedicated section on the site. For this process several roles,
rules and resources are needed.
4.1.
Rules
This section should outline the processes whereby content is managed. We need to
create workflows followed by staff to create, publish and maintain content on the
site. This is contingent on work-flow mapping.
4.2.
Roles
There are several roles that need to be carried out for content to remain current and
accurate.
Editor
This is the one content vetting gateway through which all material on the site must
pass. Final editing responsibility of all content rests with the Head of CODIS. She
approves, edits or rejects content submitted for publication.
Website administrator
The webmaster is in charge of maintaining the site overall, customizing templates to
our specific needs, be the first port of call for web-site related enquiries and troubleshooting, and help train people to use the system.
Focal Points
Focal points are the interface between staff and the web-team. We envision a
network of focal points in each department, unit, program, and outreach office. The
focal points will ideally be tech-savvy individuals with some web-design affinity.
More importantly they also bring some subject area expertise to the job and the
ability to identify latest developments and trends within their team.
Their role could also be described as knowledge facilitators. They identify fresh
content among their team’s recent outputs, continuously remind their colleagues to
submit this content to CODIS and, thus, ensure a constant flow of fresh content to
the website.
The focal point is a highly visible role, bringing with it the intrinsic pay-offs of having
many contacts across units and departments. It has been said that knowledge
facilitators never need to worry about finding a job because they are so visible and
keep receiving offers.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
26 |35
4.2.1. Functionality by roles
All site visitors can
 Access information pages
 Leave comments on blog posts
 Sign up for a Website Account
 Use contact us form
All logged-in users can:
 Access discussion forums, wikis, blogs
Logged-in staff can
 Edit their profiles
 Write blog posts
 Upload content, files and pictures
 Access private and shared calendar
Focal points can
 Add new content of any type
 Edit Product Information pages
Website administrators can
 Designate which users are staff
 Edit All Information Pages
4.3.
Resources
The website is the final destination for the products of various workflows and
business processes such as the publishing workflow and the project life cycle. As
such the website’s CMS needs to integrate smoothly with the document
management system (DMS) as well as the library’s document repository.
The IT department is currently in the process of re-designing the intranet. Much
importance was given to integration and the ability to manage documents and
workflows. For a more detailed explanation of options how to integrate document
repository, DMS and CMS please consult the document ‘Intranet 3: Requirements
Analysis and Migration Plan’ recently issued by ITU.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
4.4.
27 |35
Example publishing process: Report
The following is adapted from a presentation by Peter Ballantyne, given to ILRI:This
content flow model follows the principle that content is captured once and re-used
often. This saves time and resources while increasing accuracy and consistency. The
sequence here is intended as an example and is played out with a report published
by an in-house researcher
Scientist produces reports
1. They go into the library
2. We produce a press release
3. We put it on the website
4. We put it on the intranet
5. Maybe link from blog
Now we add other media/outputs
6. Film and video
7. Photos and images
8. Slides and posters
9. Events
10. Podcasts and audio
Each of these content types go onto their own specialised platform. We keep track of all outputs and
use Google Custom Search Engine to search across all the sites.
RSS is the glue/catalyst that moves the content from all these services to users and onto our web and
intranet sites. It also publishes onto social and mobile platforms.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
Content Flow Model
28 |35
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
4.5.
29 |35
Content Management System
Requirements
From a content perspective, it is critical that the CMS provide:
 User-friendly and time efficient creation, editing (including spell-checking), and
linking of
 Secure, distributed content creation and management by authorised users;
authorised users are managed by role (e.g. departmental editor, information
custodian, administrator).
 Import, management (including version control) and display of files in multiple
formats (e.g. DOC, XLS, PDF), of images (e.g. JPEG, GIF and PNG), and of
multimedia in file formats such as MPEG, FLV, MOV, and multimedia ‘bundles’
(e.g. files produced by Adobe Presenter).
 Creation and management of customisable forms.
 Creation, editing, management and maintenance of links
 The ability to search, preview and replace text across the site or sites managed
by the CMS.
 Publishing of a single content item in multiple delivery environments (e.g.
intranet and internet) with the option of different delivery
requirements/templates
 The mapping and management of hierarchies and inter-relationships of
information
 A number of predefined and/or customisable templates and stylesheets
(including a print stylesheet) to control the visual presentation of content.
 Facilitate generation/review/approval of metadata at the same time as the
content item, as part of the one process/workflow.
 Enable automatic updating of navigation structures (e.g. the site structure) as
content items are updated.
 Automatically generate and update a site map for each website.
 Provide navigation menus that support multiple levels of navigation (e.g. pages
and sub-pages)
 Support the measurement of the use/effectiveness of each keyword search
query entered by a user in searching the website.
From a workflow perspective, it is critical that the CMS provides:
 The management of all aspects of the writing, editing, approval, publication and
management of content
 Assignment of roles and activities within a workflow to authorised individuals
 Assignment of workflow to classes of content (e.g. images, multimedia)
 A workflow approval process that supports sequential approval steps
 The ability to assign varying status (e.g. draft, pending approval etc) to a content
item during the authoring process
 Different work processes for different types and frequencies of content (e.g.
news items)
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)













30 |35
Notification/alert when data entered in a template is incomplete (e.g. form field
incomplete)
Notification to the relevant person when they need to perform a task
From a compliance perspective, it is critical that the CMS provide:
Provide or support standards-based development (eg XML, CSS, DHTML, XSLT
and other open standards).
Support current industry standard web programming frameworks (e.g. J2EE or
Microsoft .NET, PHP).
Support standard messaging protocols (e.g. SMTP, MIME).
Support data exchange involving open standards and protocols for web services.
This includes WSDL, UDDI, SOAP and REST
Interface with external directory services for security and authentication, using
industry standard protocols
From a technical requirements perspective, it is critical that the CMS can:
Provide cross-platform (Mac, Windows, Linux) multi-browser capability,
specifically, support for Internet Explorer 6 and above; support for W3C browsers
such as Firefox, Safari, Opera; and graceful fallback for older browsers.
Support development, testing / staging and production environments, and
control of content movement between each.
Support minimal response times in saving data, page load time and any workflow
action, ideally less than 10 seconds.
Provide reporting tools that give information and statistics on website usage by
clients.
From a support requirements perspective, it is critical that the CMS provide:
 Soft and/or hard copy user and technical documentation.
 Technical and specialist training for staff in areas such as site management,
template creation, business rules and workflows.
From: APPENDIX H – Content Management System - general requirements, Niche
report p57
Tool of choice: Drupal
The tool that best fits the requirements above is Drupal, an open source content
management system, probably the fastest developing one on the current market.
We are specifically adopting an FAO information product called AgriDrupal, which is
a custom adaptation geared towards the needs of agricultural research centres. In
this way we do not need to pay license fees, are ensured the expert advice of FAO,
and benefit from an evolving product with a healthy developer community. By
investing in building in-house capacity for development we ensure sustainability and
safe on outsourcing as well as future maintenance contracts.
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
4.6.
31 |35
Templates
Several templates will need to be designed for the Content Management System.
These were identified in the content audit.
Content analysis
A content analysis was conducted in three steps. First a content audit identified the
types of content and established content genres. Second, a complete inventory of
every URL was compiled noting content ownership and ROT status and assigning a
unique ID to each page, which will be used for site migration. Finally a content map
was drawn up, connecting content sources, with types of content and templates to
be designed (s. below).
Content Sources
Content types
Templates
Programmes (Res & Reg)
Project Reports
Program Profile
Project Profile
External websites
Field Guides
Proceedings (Conference & Workshops)
Program reports
About pages
Germplasm Catalogue
Mid Term Plans
Policy Docs
Material Transfer Agreements
Strategic Plans
Media Releases
Annual Report
Caravan
Bios (Board, Staff)
Ties that bind
Contact us
Contact adresses (Outreach offices)
Videos
Events Calendar
Photo Gallery
Publications database
Financial Statements
Peer Reviews (CCERs,...)
Manuals
E-learning courses
Vacancies
Visitor's Guide
Publication
Program Profile
Project Profile
Program / Project Profile
Publication
Publication
Publication
1st level page
Publication
Publication
Publication
Publication
Publication
Media Release
Publication
Publication
Bios
Publication
Form
Bios
Video Page
Events calendar
Photo gallery
Management
CODIS
Photo Lab
Library
Finance
External Reviewers
Capacity Building Unit
Human Resources
Visitor’s Services
Publication
Publication
Publication
The content map below illustrates how content sources, content type and content
templates are connected:
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
Content Map of www.icarda.org
32 |35
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
4.7.
33 |35
Metadata
Every content object is associated with some unique meta-data that will make it
manageable. Particularly keyword can dramatically increase the findability of an
object and Search Engine Optimisation relies heavily on them.
Metadata Element
Author
Publisher
Title
Date
Expiration date
Links
Document type
Subject area
Keywords
Related to
Geographic area
4.8.
Example
Max Johnson
Mike Ferrel
New publication on food security
Thu Feb 18 2010
18032010 12:02:23
/news/02182010/story.htm
News story, glossary term
Food security
Research, drought, food security, water
News stories, publications, food security
CWANA, Syria
Content chunking
Similarly to identifying metadata, ‘content chunking’ is another key element of the
re-use and re-cycle strategy. By identifying the smallest useful content chunks of a
given item it is possible to map them onto their future destinations and keep track of
them. This again reduces maintenance costs because content only has to be updated
once when it changes, saves on translation costs, and improves coherence and
accuracy.
Re-usable elements can be recombined to fit various media channels without having
to rewrite them everytime: e.g. a publications template on the website, a blog entry
on wordpress, a search result on Google CSE, or a Twitter message. Here is an
example of content chunks and where they might be re-used:
Content chunk
Author
Title
Date
100 word summary
Link to PDF
Relevant to
Picture
Website







Blog






Search result



Twitter




icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
4.9.
34 |35
Search
Google Custom Search Engine (CSE) is the most powerful tool available right now to
search across various URLs such as www.icarda.org http://icardablog.wordpress.com
and all the other future pages hosting dynamic content that we are going to set up.
This includes Flickr, various blogs, twitter, video-hosting sites such as blip.tv and
others (see Content Flow Model above). Google CSE was already implemented on
the existing site thus repairing the broken search function.
4.10.
Disaster Recovery
The website is mission critical and we should aim for it to be up 99.99% of the time.
ITU advised a two-pronged security strategy. One the one hand there needs to be a
system in place that protects data. On the other hand of measures are needed to
mitigate potential loss of data.
Strategies for data protection
 Daily backup of virtual server by ISP (CGNet)
 Daily backup of local LAN servers (on site)
 Daily backup archived in fireproof safe in separate building
 Weekly backup stored offsite in ICARDA Office 1
 Quarterly total backup archived in ICARDA Cairo office
Data loss mitigation techniques
 Disk failure protection technology (RAID)
 Surge protectors and uninterruptible power supply (UPS)
 Fire preventions — alarms, fire extinguishers
 Physical firewall
 Strong Authentication – Microsoft Active Directory use of Kerberos
 Encryption
 Layered Software protection:
- Proxy and Firewall
- Web Filtering and repudiation
- Antivirus software
- Anti-spam, Anti-adware
icarda.org :: web strategy
Version 0.5 published on 7 July by Bernhard Hack (b.hack@cgiar.org)
35 |35
5. Glossary
Content (From Polar Bear book p239)
We define content broadly as “the stuff in your web site.” This may include
documents, data, applications, e-services, images, audio and video files, personal
web pages, archived email messages, and more. And we include future stuff as well
as present stuff. Users need to be able to find content before they can use it—
findability precedes usability. And if you want to create findable objects, you must
spend some time studying those objects. You’ll need to identify what distinguishes
one object from another, and how document structure and metadata influence
findability.
Content management system (CMS)
Content management systems support the creation,management, distribution,
publishing, and discovery of corporate information. Also known as ‘web content
management’ (WCM), these systems typically focus on online content targeted at
either a corporate website or intranet.
Enterprise content management system (ECMS)
An enterprise content management system consistsof a core web content
management system,with additional capabilities to manage a broader range of
organisational information. This often consists of document management, records
management, digital asset management or collaboration features.
Document management system (DMS)
Document management systems are designed toassist organisations to manage the
creation and flow of documents through the provision of a centralised repository,
and workflow that encapsulates business rules and metadata. The focus of a DMS is
primarily on the storage and retrieval of self-contained electronic resources, in their
native (original) format.
Records management system (RMS)
The Australian Standard on Records Management(AS 4390) defines recordkeeping
systems as ‘information systems which capture, maintain and provide access to
records over time’. This includes managing both physical (paper) records and
electronic documents.
6. Timeline
s. below
7. Budget
s.below