Many people are familiar with the Latin phrase “caveat emptor” which means “buyer beware.” But after reading a number of misleading articles and reports in trade newsletters recently, I’m more inclined to say “caveat lector” whenever I read anything these days, which means “reader beware.”
To wit: this report from The Wise Marketer which references a study from a loyalty vendor called Thanx. The headline: “Most retail & restaurant customers never come back.” Let’s just start from that premise: is that likely? No, because as a generalization, there aren’t enough consumers in the world that if ‘most’ visit a retailer and never come back, that retailers could stay in business. Could they at least qualify it, maybe as in “most local retail customers never come back” or “most jewlery customers never come back” if that were the case? So let’s see where that headline comes from.
The key stats quoted are as follows:
- 70% of previously loyal customers are unlikely to return
- 3-4% are super-loyalty and visit at least monthly
- 25% of customers contribute 64% of revenue
- 49% make weekend-only purchases
This is too general. Is it likely that consumers have similar visit and spend patterns across all retailers and restaurants? No, not likely. Visit patterns to grocery stores are going to be different from general merchandise retailers like Target, different again from specialty retailers, different again from quick-service restaurants, and different again from fine-dining establishments. And over what time frame are they making this claim? Unlike to return ever, or just within some defined timespan?
Digging deeper and looking at the dataset from the source study, you find that they analyzed a random sampling of 10.3MM customers and 18.3MM transactions across 54 business that they support with 877 locations, using a 6-month timeframe. They further say that 3% of customers shop monthly, and 70% of customers haven’t revisited within four months. That leads to some further thoughts:
1. There were an average of 1.78 transactions per customer in the dataset. So did every customer in the dataset have at least one transaction? Or were there some on the customer file that were inactive during the period? What proportion? If every customer had at least one transaction, of course they show a very low repeat purchase rate, it is by definition!
2. 190K customers per business, or 11.7K per location, meaning about 65 transactions per location day. Dataset is too sparse to capture repeat business. How was their sampling done, did they randomly sample consumers, and then randomly sample transactions, or did they randomly sample consumers and then pull all associated transactions? Study design make a HUGE different in the output and the insights they derive.
3. Clearly mixing data from different business types is going to skew results. Is it representative of all retail categories? A heavy user of McDonald’s might visit 40 times across six months, while a heavy user of a French restaurant might visit once or maybe twice in six months. But they would still be loyal. A jewelry retailer might expect less than one visit on average over six months as well.
The source study is called “6 critical insights for customer loyalty” which sounds good, but these flaws in the methodology call each of their insights into question. And that seems to be a very common problem. I’m going to continue reading these types of reports with a “critical” eye.