We know from the questionnaires that the participants of our field test were much better informed after completion... but can this effect also be proven quantitatively with the return rate? Yes, it can!
We have evaluated the returns from phase I, II and II of the field test. Since the returns rate varies greatly over time, we compared the average returns rate across all participants in the field test with the same value for all other customers - in each case in the period in which the corresponding phase of the test was active. Phases I and II were about two months long, phase III three weeks. Christmas fell in the middle of Phase III, a factor that might explain the lower return rate in the control group - who returns a gift?
One can clearly see that the return rate of the control group already fluctuates strongly, from 41% to 36%, i.e. it is about one tenth smaller. In phase I, the control group is slightly better, but the difference is small and not significant due to the high variability. In phase II and III, the field test then has a significantly lower return rate of about 6%. Within the group participating in the field test, the return rate could be reduced from approx. 45% to approx. 30%.
What was tested in the three phases?
Phase I included machine learning strategies for optimized size information (based on size tables from production) and improved product information in the form of qualitative size features (such as figure playing, figure emphasizing, straight or casual).
Phase II included all strategies from Phase I and additional interventions of persuasive design, such as color and layout emphasis of size tables and product information, as well as the provision of sustainable delivery options.
Phase III included all strategies from the previous phases and interventions from gamification, such as rewards for sustainable buying behavior, the evaluation of comments, a personal level system, and banner information on delivery options.