230% Increase in Customer Retention with AIš¤ Recommendation System
We are talking about the efficient message using Retenticsās recommendation system which leads customers to make a purchase.
Nov 14, 2022
āNew arrivals! Get them with a 10% discountā
If you receive this kind of message, how would you feel? If the message is the one you were waiting for, it would be worthy, but if not, you no longer want to get another one from the brand.
In this article, with the case of <Woodo>, we are talking about the efficient message using Retenticsās recommendation system which leads customers to make a purchase.
š Woodo operates a non-face-to-face book sharing service. It allows you to borrow and return books at the door. You can also consign them to sell.
1. Needs of personalized message
<Woodo> holds an event called āBook Partyā every other week, providing a 30% discount on borrowing books. The existing customers were well aware of it because they can get discounts for various items at a high rate. <Woodo> tried to keep customers who are likely to remain by this event. However, as they are already well aware of the benefits, the message of ā30% off for all booksā could not hold them back. They need more personalized messages.
2. From mass communication to personalization
Retentics learns the historical transaction from the first moment that the customers interacted and provides you with an item name that each customer is most likely to purchase next. With this pair, Retentics enables you to build up a personalized marketing message.
Our clients used to communicate mostly in rule-based ways to make their customers retain. However, these ways can be very critical if customers have various tendencies in frequency, monetary, and category range. The algorithm has already proved its performance in many different businesses in the last few years and has been finally advanced with our teamās hard-working.
3. Proven performance with the A/B testing
The A/B testing results of Retenticsās recommendation system are very incredible. We carefully chose the customers in the lapsing status and split them into two different groups. Some in the control group received the recommendation message with the most popular items but never purchased yet, and the others in the treatment group got the message using Retenticsās recommendation system. On average, 200% more customers in the treatment made a purchase than the others. Furthermore, the recent testing with Woodo, one of Retenticsās clients, induced 230% more customers to return and made 280% higher unit price than the others. Woodo also sent a message to the rest of the customers from the selection for testing, the difference in the returned customers was 920%, compared to the treatment group.
4. Feedback from <Woodo>
Makedash's AI solution is great, but the team's passion and dedication was even greater. We are highly impressed by the process of making results with their capabilities even for the parts that are difficult to define according to the characteristics of the business .
Retenticsās recommendation system is still being improved. You can filter the customers with the last purchase basis for now, but AI will detect the customers based on their moment such as lapsing, in golden-time, or almost loyal becoming, etc. Also, beyond the segmentation algorithm that clusters the customers into a few (3-6) groups with a similar transactional propensity, we will implement the hyper-segmentation that clusters them into tens of groups to detect the micro tendency in transactions by Q4, 2022.
Retentics does not focus only on increasing your revenue but enabling you to provide your customers with a better customer experience. In order to make your customers return, all you need is a day with Retentics.
Written by: Zack Lim
Share article