Defence
Master
Thesis
Defence
Master
Thesis
MSc Marketing
Intelligence
MSc Marketing
Intelligence
By George Radix S32444587th february 2019 First supervisor: prof. dr. Wieringa
Introduction:
- The combination of MSc
Thesis and an internship
- Lessons learned on the
office floor
Combining MSc thesis with an internship & lessons
learned
● Navigating the organisation ○ To whom do I go?
● Juggling company and academic interests
○ Most co-workers on different knowledge levels ● Being part of a cross-functional project team
○ Finding different truths between departments
Thesis: scope
“Determining what the effect of compensating customers for company errors is” ● Start with a table called financial mutations: containing 17 different types of errors that are
compensated
○ The accuracy of this table was not guaranteed and connection to customers was unreliable ● Move towards a dataset from the marketing department
○ A/B test data with a control group
● Zooming in on one very specific failure with reliable data
Methodology
● Main effects of the failure and two recovery strategies ● A unit-by-unit model approach for each customer segment ● Controlling for age and relationship length of the customers
Discussion: most important results
Effects of the failure:● Total of 9 cases where a negative effect is found
○ But also 4 cases with positive effect (3 in new segment)
● Mostly lower loyalty customers are negatively affected by the late delivery (exception: new customers)
● Coupon consistently stimulates buying behaviour through all DV’s in nearly all segments ● An apology is not impactful enough to have consistent significant effects
● No consistent significant differences found between levels of loyalty and respective effects of failures and recoveries (mixed signals found between the three models)
Implications 1/2
● Company errors (even as small as a delivery that is minutes late) will negatively influence the customer buying behaviour if left unchecked
● Using an effective recovery strategy, like a coupon, will decrease these negative effects ● The apology does not have satisfactory effects on the behaviour of customers
○ Is an apology without the personal contact still a true apology? ● Active customers are less bothered by the failure
○ Have experienced it more often? Or are reasonable enough to admit something could go wrong every so many times?
Implications 2/2
● The less active customers experience the delivery issue as negative
○ This could indicate that when orders are less frequent, problems with the delivery may be experienced as more severe
Limitations & Future research
Limitations:● Violation of the non-normality assumption for the error term ● Presence of heteroscedasticity
● Residuals are high
○ Fairly high amount of unexplained variance between observed and predicted ● R-square is low
○ No predictors of buying behaviour / explanatory variables included in the model ● Pooling of models was not investigated further
Future research: ● Buying behaviour ● Competitor effects
● Over time documentation of failures experienced ● Contact with customer service
● Product category information
Discussion: most important results
Delta amount spent:● Negative effect of failure for low loyalty customers (-) ● No effect for high loyalty customers
● Apology only significant for new customers (+)