“McKinsey says…” Isn’t Good Enough Anymore.

← Back to all posts

I recently dove into the sourcing for several loyalty metrics that are commonly cited (see here: https://www.linkedin.com/pulse/time-retire-commonly-cited-loyalty-metrics-john-c-keenan-pesne/), and on which companies may be basing marketing investment decisions. (And have previously taken a similar look into stats around personalization, here)

H/T to my friend Steve Bocska for summarizing it with this:

“Sloppy research isn't just annoying, it's dangerous to your business objectives. Low quality information leads to bad decision making.”

His response really got me thinking about how many of us throw around industry statistics without having any idea of their sourcing, and it came to mind again when I read a Retailwire article (“Is Tech the Best Lever to Drive Personalization and Relevance in Loyalty”) that in turn led to a DunnHumby publication (“How to Keep Hold of Your Customer”, may require registration) and so on down the rabbit hole. Until this:

“80% of companies tracking the ROI of their loyalty programs report an average return of 4.9x.”

No wonder every brand seems to have a loyalty program! But which exactly are seeing those kinds of returns? I’ve worked on many loyalty programs, including building out sophisticated financial models to provide justifications, and have seen positive ROI but nothing on the scale of 5x, and that is quoted as the average… Where did this stat come from, and what prompted a respected source like DunnHumby to use their authority to broadcast it widely without validation?

Not surprisingly, sloppy research. The cite in the DunnHumby article pointed to “The Top 215 Customer Loyalty Statistics for 2024 and Beyond” as published by Antavo. (Not trying to critique anyone in particular; there are numerous similar lists constantly being published, like this from businessdasher.com.)

So where did Antavo get this stat? From their own study, published in the “2023 Global Customer Loyalty Report”, and the same stat also appears in the 2025 edition. Digging deeper, it appears to be based on self-reported responses from approximately 2,600 respondents to a survey, with the methodology slide providing some breakdown by geography and type of business. But the numbers don’t add up, since 39% of respondents work for vendors, and we don’t know how many are duplicate reports from the same company, among other issues. If you’re in the US or Canada, by my calculation only on the order of 150 respondents would be brand-side from North America, distributed across a wide range of verticals, so you can’t project the findings with any confidence.

That said, this would still appear to be one of the most comprehensive surveys of loyalty practitioners out there, so it makes for interesting reading, but I see enough issues with the methodology that I wouldn’t rely on it to drive my decision making.

I suspect one could perform a deep-dive into each of the other 214 stats in this report and find similarly suspect sourcing, but that’s not necessary – just take any of these industry reports and rules-of-thumb with a grain of salt and don’t set expectations inappropriately as a result. Whether it’s “Antavo says…” or “McKinsey says…”, before you rely on any of these generally-accepted but thinly-sourced industry stats, do some checking. As the old journalism maxim goes, “Even if your mother says she loves you, check it out”.

View original on LinkedIn