Personalized marketing for retail is here to stay. Advances in big data processing and predictive algorithms are allowing today’s retailers to offer a highly customized experience. In return, they are getting increased loyalty and higher sales.
But what is often neglected is that there is a group of customers – sometimes a very large group – that simply don’t have enough transaction history to make personalization possible. They may be new customers or infrequent shoppers. They may be long-time customers who just recently signed up for a loyalty program. Or the retailer may simply be in a vertical where customers visit less frequently. We have recently worked with retailers where a typical customer might shop only once every two months or where more than one third of customers buy fewer than ten products during the year.
The fact is that personalization algorithms are made to work for customers whose data history is rich enough to make predictions. If I stop in your drugstore once in a while to get some painkillers to relieve a headache, you can’t exactly use those few transactions to dazzle me with personalized offers.
In our experience, retailers do not spend much time developing a strategy for engaging these “thin data” customers with their personalization programs. Many just let the algorithms run as usual. The output is, predictably, useless.
But there are some strategies that can help retailers extend their personalization programs even to customers without much history. Even better, they can move customers up the chain to become a frequent, valuable customers. In our experience, here are four of the strategies that work best.
1. Go Wide
If you don’t know the specific product a customer wants, a broader offer on a category or department is a great place to start. You may not know if a customer uses Crest Whitestrips. You do know that they brush their teeth (hopefully). An offer to spend $10 on anything in oral care is not only safer and appealing to the customer – but you can also learn something from their choice and be more personalized next time.
2. Use What You Know
At times, traditional market segmentation can be the enemy of personalization. Just because males aged 40-55 like a product in general, doesn’t mean I do or you will. But when you don’t have any personal shopping history to go by, segmentation can be a good fallback. For example, you may find that for drug and beauty customers not only does a customer’s age help determine which products they will likely buy (for example: 16-30 year olds are more likely to purchase makeup, 30-45 are more likely to buy nutrition and 45-65 year olds are more likely to purchase fragrance), but there might be a difference in discount sensitivity. Increasing the discount may not change things much for younger customers, but older customers might be highly responsive to savings—or vice versa. Combining these insights can go a long way towards getting the right offer to the right customer, even when there isn’t much purchase history to go on.
3. Create a Set of Default Offers
If you are committed to giving product level offers and don’t collect demographic data on your customers, you can still probably do better than the recommendation algorithms for customers with limited data. The output of the algorithm could be unpredictable. It is much better to create a set of “default” offers that you give to customers when you can’t personalize. The offer set should include best-selling items from a range of categories to make sure there is something for everyone. You will also want to avoid items that are sensitive or that people tend to have strong feelings about it.
4. Give Them a Choice
If you don’t know which product a customer wants, why not just let them choose? 10% off any product of your choice is a simple offer, great for the customer, and sure to be on target every single time. Some programs also let customers give input when they join about the kinds of offers, they want to receive. A few simple questions during an online sign-up process about a customer’s favorite items or categories is a quick way to get to know their preferences – and it sends a signal that they can expect a personal touch.
One last word of counsel. Whichever of these strategies you choose, it is important to communicate to the customer exactly what you are doing. If you have been shouting in your marketing about delivering personalized offers, customers have a right to expect them. If you get it wrong, they will notice. A simple message like “The more you shop, the more personalized your offers will be” will not only help set customer expectations. It will also get them moving in the right direction.
