EMAC 40th Conference
“The Day After: Innovation, Inspiration, Implementation”
Using CLV Concept for Marketing
Budgets Allocation
Olga Oyner and Elena Panteleeva
(National Research University
“Higher School of Economics”)
Ljubljana, 2011
Theoretical background
– To make marketing measurable:
• More transparent from financial point of view (Kotler
2002; Kumar2004; Ambler 2006 )
• More efficient in terms of output/input (Shet, Sisodia
2002)
• To consider marketing results from long time
perspective: from costs to investment and from profit
to equity (Rust 2004; Ambler 2006; Strivastava 2001;
Doyle 2001)
• To estimate marketing impact on financial results and
firm value (Rust et al 2004; Ambler 2006; Shet, Sisodia
2002; Srivastava et al 2001)
2
Theoretical background (cont.)
• Marketing productivity chain – the marketing
impact on the firm’s value (Rust et al 2004;
Ambler 2006; Shet, Sisodia 2002; Srivastava et
al 2001)
• Customer equity and CLV (Pfeifer, Haskins, and
Conroy 2004; Rust and Lemon 2001; Rust,
Lemon, and Zeithaml 2004; Kumar and
Reinartz 2006)
3
Research object
• Large multinational business-to-business
(B2B) PC hardware manufacturer
• Business task: how to generate more value
from every client?
4
Model: chain of marketing productivity
5
Aims
• The proposed model is aimed at increasing the
efficiency of the intrafirm budget allocation
because it will help determine which marketing
instruments have the greatest impact on CLV and
sales
• Sales targets are set on individual level and don’t
take into account marketing budgets spent on
each single customer in order to achieve these
results. So we need to find out if there are any
differences in how marketing instruments
contribute to the sales volumes and CLV
6
Data characteristics
• Data on purchase volume and marketing
budgets spent on various marketing activities
for company’s clients
• 451 clients from one of the sub regions of
Russia and CIS countries
• 1471 datum
7
Hypotheses
• H1: discount amount positively influences the
sales volume, but doesn’t significantly add to the
CLV
• H2: account manager (communication level) has
positive and the greatest impact on sales
• H3: retail promotion has positive impact on sales
but this impact is less then account manager has
• H4: core product promotion has positive impact
on sales but less then all above has
• H5: other product promotions has positive
impact on sales but less then all above has
8
Variables and CLV
• Independent variables for the models were:
–
–
–
–
Account manager – investments in relationship marketing
Retail promotions – budgets for co-marketing with retailers
Discount – price reductions given to specific customers
EPSD – budgets for co-marketing activities with clients for server
products (seminars for end-users, trade-shows)
– UPSD – budgets for co-marketing activities with clients for
motherboards (seminars for end-users, trade-shows)
– DT - budgets for co-marketing activities with clients for core product
(seminars for end-users, trade-shows)
Sum of SO
Amt Disti
Cost
CLV
13073530
50692756
Discounts
Retail
promotions
Account
Manager
UPSD
EPSD
DT
376987
17000
63110
9000
4000
14700
9
Regression equation: sales
Unstandardized
Coefficients
Model
6
B
Std. Error
-2080,950
2844,633
Account
Manager
215,468
8,273
EPSD
594,232
UPSD
Standardized
Coefficients t
Sig.
Beta
95% Confidence Interval for
B
Lower Bound
Upper
Bound
(Constant)
-,732
,465
-7672,915
3511,016
,764
26,045
,000
199,205
231,731
19,225
,719
30,910
,000
556,440
632,023
-163,746
18,840
-,226
-8,691
,000
-200,783
-126,709
DT
55,565
8,672
,117
6,408
,000
38,518
72,612
Retail
promotions
-115,572
26,309
-,141
-4,393
,000
-167,290
-63,854
Discounts
1,548
,493
,062
3,142
,002
,580
2,517
Sales = -2080,950+215,468*Account manager +594,232*EPSD promo -163,746*UPSD
promo +55,565*DT promo - 115,572*retail promo+1,548*discount
10
Regression equation: CLV
Unstandardized Coefficients
Standardized
Coefficients
B
Std. Error
Beta
-43336,460
16671,892
Account
Manager
1204,753
48,275
EPSD
2261,849
UPSD
Model
5
t
Sig.
95% Confidence Interval for
B
Lower
Bound
Upper Bound
(Constant)
-2,599
,010
-76109,750
-10563,170
,904
24,956
,000
1109,855
1299,651
111,820
,579
20,228
,000
2042,036
2481,662
-775,462
109,376
-,227
-7,090
,000
-990,472
-560,453
Retail
promotions
-767,601
144,021
-,199
-5,330
,000
-1050,714
-484,487
DT
246,695
50,856
,110
4,851
,000
146,722
346,667
CLV (for 4 periods) = -43336,460+ 1204,753* Account manager +2261,849**EPSD
promo -775,462* UPSD promo -767,601* retail promo + 246,695 DT promo
+0*Discount
11
Regression equation: sales (2)
(Constant)
Unstandardized Coefficients
B
Std. Error
11311.215
Standardized
Coefficients
Beta
1320.071137
t
Sig.
8.56864
0.000
Account Manager*EPSD
0.247
0.003292838
0.745918229
74.92194
0.000
Account Manager2
0.056
0.001463787
0.845514497
38.16489
0.000
Discounts2
0.000
3.26565E-06
0.180181136
21.12752
0.000
-0.051
0.003019588
-0.19598328
-16.96065
0.000
0.034
0.005672027
0.261310706
6.070955
0.000
-51.218
6.833582779
-0.18154741
-7.495006
0.000
0.109
0.011900425
0.160109682
9.159051
0.000
-0.003
0.000535477
-0.08463594
-5.3837
0.000
-79.975
21.29279383
-0.16802509
-3.755958
0.000
Account Manager*UPSD
DT2
Account Manager
Retail promotions*DT
Discounts*EPSD
DT
In order to rationalize the budget allocations we need to determine the equation of the
non-linear multiple regression:
Sales = 11311.215 + 0.247Account Manager*EPSD+0.056 Account Manager2+0*
Discounts2+ -0.051Account Manager*UPSD+0.034 DT2-51.218 Account Manager+0.109
12
Retail promotions*DT-0.003 Discounts*EPSD-79.975 DT
Budget allocation optimization
Current budget
allocation
Optimized budget
allocation
Discounts
376987
0
Retail promotions
17000
9634
Account Manager
63110
69000
UPSD
9000
0
EPSD
4000
5000
DT
14700
24176
Company local budget
Company
overall
budget (incl discounts)
107810
107810
484797
107810
-376987
285515711
391037021
105521311
CLV
Difference
13
Framework limitations and future
research directions
• The study is based on one B2B hardware PC
company. Similar studies in identical
companies are needed to outline general
correlations
• It would be useful to calculate CLV and
customers’ profitability not only on an
aggregated level, but on individual ones in
order to tailor marketing mix tactics for each
customer specifically
14
Framework limitations and future
research directions (cont.)
• We evaluated only those budgets, that can be
attributed to each customer specifically. We
didn’t take into account corporate marketing
activities, such as PR, ATL, etc.
– how will the proportion between those
“customer-specific” and “general” budgets
influence the CLV?
– what if “general” budgets are more efficient on
the company level?
15
Framework limitations and future
research directions (cont.)
• We didn’t take into account competitor’s
response to the marketing actions of the
company. But for the oligopoly in which the
company in study is in, it is vitally important
16
THANK YOU!
Q&A
17