Follow the Crowd: The Science of Predicting Loyalty Enrolment

Ms Chan

Welcome to Loyalty Unlocked — the podcast where we pull back the curtain on what really makes loyalty programmes work.

Ms Chan

In each episode we dive into the strategies, psychology, and behind-the-scenes thinking that turn points into behaviour, and loyalty into impact.

Ms Chan

For full transparency, this podcast, it's script and your lovely co-hosts, are all AI generated.

Ms Chan

Alright, so let’s dive right in, Mark. What on earth does a Victorian experiment with bouncing beads have to do with loyalty programs?

Mark Sage

Good question. In the late 1800s, Sir Francis Galton came up with something called the Galton Box, or the bead box. Essentially, it’s this contraption where beads drop through a series of pegs and, as they fall, they scatter at random. But here’s the thing

Ms Chan

Wait, wait, let me guess. They don’t just pile up randomly at the bottom, do they?

Mark Sage

Exactly! The beads form a bell curve—a perfectly predictable pattern, every single time. Galton called this "the Law of Unreason." Even though the movement of any one bead seems random, when you get enough of them, the chaos turns into predictable order.

Ms Chan

Okay, that's already kind of blowing my mind. But how does that connect to loyalty programs?

Mark Sage

Well, launching a mass loyalty program is a bit like dropping those beads into the Galton Box. You can’t predict exactly when or how each individual customer will sign up, but when you look at the crowd as a whole? Behaviors become surprisingly consistent. It’s as if the randomness of individuals is smoothed out by the sheer volume.

Ms Chan

So you're saying... even though every customer thinks they’re acting independently, it all adds up to one big predictable pattern?

Mark Sage

Pretty much. And that predictability? It’s crucial. When we launched yuu Rewards in Hong Kong, the sheer scale demanded that we use these patterns to plan ahead. Otherwise, we’d risk being completely overwhelmed.

Ms Chan

Yeah, I mean, getting millions of people to sign up... How do you even start to prepare for that?

Mark Sage

You start by knowing your crowd. Hong Kong is densely populated, and grocery shopping is a frequent, habitual activity. That means we were dealing with a massive, highly active audience—perfect for a coalition program like yuu Rewards. But we didn’t just wing it; we leaned on data, historical trends from other programs, predictive modeling, any insight we could get our hands on.

Ms Chan

And this is where volume hides predictability, right?

Mark Sage

Exactly. For starters, we planned for key milestones—the "big waves." We knew, based on past programs, that a huge chunk of sign-ups would happen in the first few days. Early adopters always come in fast and furious. Then things slow down, but the patterns don’t change; they just scale with the size of your audience.

Ms Chan

So it’s like... you can’t pinpoint what any one customer will do, but you definitely know when the crowd will move. That’s so cool!

Mark Sage

It is very cool. Using tools like Galton’s findings and predictive data gave us confidence to tackle such a gigantic rollout, and to manage everything—marketing, operations, even technology capacity. Knowing how the "crowd" would behave meant we weren’t flying blind.

Ms Chan

Mark, you mentioned those "big waves" of early adopters just now. It’s making me think of those curves where groups gradually adopt something over time. Can you remind me how that applies to loyalty programs?

Mark Sage

Ah, you're probably thinking of Rogers’ Diffusion of Innovation. It’s a model that shows how any new idea or product is adopted by different groups in a population. The key here is to understand the different adopter types—like Innovators, Early Adopters, the Early Majority, and so on.

Ms Chan

Oh, right, like the bell curve! Innovators are the first ones jumping in, and then it spreads to the rest, right?

Mark Sage

Yes... and for loyalty programs, we’ve observed the same behavior over and over again. The enrolment curve mirrors this diffusion model because people adopt loyalty programs at different rates.

Ms Chan

But this enrolment curve doesn't really talk about time - how do you know when each group will be coming in?

Mark Sage

Thats very true. Rogers doesn't really help us with the 'when', but we can leverage the "4-40-400 Rule" to help here.

Ms Chan

Okay, hold up. The what-now rule?

Mark Sage

The 4-40-400 Rule. It’s a pattern we’ve seen consistently with large loyalty programs. It means that in the first 4 days or so, you get about 16% of your base—that’s your Innovators and Early Adopters. By day 40, you’re hitting the Early Majority and reaching around 50%. And by day 400, just over a year, you’ve onboarded roughly 85% of your total 2-year audience.

Ms Chan

Wait, wait. That’s super specific. Does it really hold up every time?

Mark Sage

Most of the time, yes. Across multiple markets and programs, like Nectar in the UK or yuu in Hong Kong, this rule has proven incredibly reliable. There are variables, of course, like market conditions or promotional strategies, but the overall pacing is remarkably consistent.

Ms Chan

Huh. That’s kind of wild. So why do these early numbers hit so fast? Is it just people rushing to be first?

Mark Sage

Not exactly. It’s about behavior. High-frequency shoppers—what we call "high loyals"—play a huge role here. These are the customers who visit frequently, spend a lot, and immediately see the value in joining. They show up early because, well, they’re already coming to the store regularly. It’s like they’re pre-positioned to move quickly.

Ms Chan

Ohhh, so these high loyals basically set the pace for everyone else?

Mark Sage

Yes - they form the initial wave. After that, the Medium Loyal customers—a bit less frequent, more divided in their loyalty—start coming in. They take longer because the value isn’t as obvious to them. And then finally, the Low Loyals trickle in over the rest of the adoption period - these laggards just take longer to convince.

Ms Chan

Okay, that totally makes sense now. It’s not just random; it’s their shopping habits driving when they join.

Ms Chan

Mark, I gotta say, understanding how these different shopper groups join at their own pace has been really eye-opening. But it’s got me wondering—when we dive into enrollment numbers, how do we even begin? How do you even start to estimate how many members you might have?

Ms Chan

Do you just base it on revenue or market share?

Mark Sage

Not quite. Market share is important, but it doesn't tell the full story. What really matters is penetration—how many customers you’re actually reaching within your market.

Ms Chan

Of course, that makes perfect sense. If you know how many customers you have, then you can simply estimate how many will join... but wait... how do you know how many customers you have?

Mark Sage

Thats the big challenge we all face - whilst market penetration is ultimately measuring how many customers in the market are shopping with you, its a really hard number to get hold of.

Ms Chan

So what - we're just stuck?

Mark Sage

Not stuck. We can come at it in a different way, and this is where Share of Category Requirements, or SCR, comes in. It’s, well, a way to measure customer loyalty based on their share of spending with you versus competitors.

Ms Chan

Okay, so SCR is like... the slice of their shopping they’re doing with you, right?

Mark Sage

Exactly. It’s a percentage of how much of their buying is happening in your stores. Once you know your market share, you can use SCR to "translate" that into market penetration.

Ms Chan

Oh - but hang on. That sounds like another tricky measure to get hold of? How do you find out your share of category?

Mark Sage

You're right - it is hard to get without using panel data or market research. But—here’s the fun part— the NBD-Dirichlet model helps us estimate this reliably using benchmarks.

Ms Chan

Wait, wait. You’re saying there’s like a cheat sheet for figuring this out?

Mark Sage

Well, you can create a cheat sheet. Basically, Dirichlet shows us that firstly, larger brands have more buyers and smaller brands have less loyalty - less buying frequency. Also SCR doesn't vary too much between buyers - it's pretty stable.

Mark Sage

This means that SCR increases with market share, and so we can link it to build a quick reference guide to let us estimate market penetration based on market share.

Ms Chan

Wow - so you've basically created a bridge from market share to market penetration.

Mark Sage

Pretty much! Think of it as a rule of thumb to make it easier to get an estimate.

Ms Chan

Does it really work though?

Mark Sage

Well, let’s take Sainsbury’s as an example. Back when they launched Nectar, they claimed 11 million enrollments within a few months. A lot of people doubted that number at the time, but when you break it down with SCR and their market share, it actually checks out.

Ms Chan

Whoa, so how does that math work?

Mark Sage

Alright, picture this. Sainsbury’s market share was around 17.5% back then. Using the SCR benchmarks, we can estimate that about 40% of their customers’ grocery budgets were spent at Sainsbury’s. Divide their market share by that, and you get a penetration rate of roughly 44% of UK grocery shoppers.

Ms Chan

Okay, that’s a lot of numbers. Break it down for me: what does 44% mean?

Mark Sage

It means 44% of the UK’s grocery shoppers likely visited and shopped at Sainsbury’s at least once that year. Given pretty much every adult in the UK is a grocery shopper, that was about 50.4 million at the time - so you’re looking at 22 million potential Sainsbury's customers.

Ms Chan

Wait... but they didn’t get all of them to sign up, right?

Mark Sage

No, not all. Typically, you only convert about 60% of those shoppers into loyalty members. So, applying that to the 22 million, Sainsbury’s could reasonably aim for 13 million enrollments within 12 months. Mapping that across 2 years - so accounting for new shoppers and churn - that number grows to about 16 million.

Ms Chan

But thats 13 million within a year - weren't they claiming 11 million within just 3 months? That feels super ambitious.

Mark Sage

It sounds ambitious - but it's actually quite achievable. Remember, the 4 40 400 Rule shows that we see 50% of the year 2 enrolment volume, within just 40 days - thats around 7.9 million for Nectar - which, amazingly is pretty much what they had.

Mark Sage

Month 3 just built on that to come in at just under 11 million - so their PR was correct, and the number matches well to their market share.

Ms Chan

Wow. It's fascinating how the numbers just seem to fall into place.

Mark Sage

It is. And what's really cool is that every part of this estimate comes from market share, and behavioral models like NBD-Dirichlet. It’s all connected.

Ms Chan

I love how it’s like a giant puzzle that just... fits together. So, you’re not guessing when these programs launch. You actually know what to expect.

Ms Chan

So, if I’m putting this puzzle together correctly, Mark—it seems like figuring out who’s likely to join is about veracity, their joining speed ties into velocity, and the overall number you can expect comes down to volume. Does that sound right?

Mark Sage

Perfectly put. Those three elements—veracity, velocity, and volume—they’re the backbone of setting realistic expectations and aligning your resources. You’re basically creating a roadmap for how the crowd will behave, and that means you can plan for call centers, servers, campaigns - even staff training.

Ms Chan

Yeah, and what really hit me was when you said, "Don’t guess the crowd, model it." Like, instead of hoping for the best, you’ve got all these tools—Rogers, the 4-40-400 Rule, Dirichlet theory—to stay ahead of the game.

Mark Sage

Exactly. And it’s not just about fancy models or numbers. This approach ensures that every part of your launch, from marketing to operations, runs smoothly. Plus, it’s about being prepared for those moments when things go better—or worse—than expected.

Ms Chan

Totally. And I think what really hit home for me is how, even in a crowd as massive as yuu’s 4 million members, there’s this beautiful predictability to it. Like, chaos turning into order. That’s wild.

Mark Sage

It really is. And it’s humbling, too—watching these universal patterns play out, no matter where or when. Grocery in Hong Kong moves just like grocery in the UK from 20 years ago. It’s a reminder that while every customer is unique, together they move in rhythm.

Ms Chan

Love that. Thanks for breaking it all down and I hope our listeners didn't get too lost in the numbers!

Mark Sage

My pleasure. And hey, always fun sharing the science behind the crowds. Thanks for having me on this journey.

Ms Chan

And that’s it for today’s episode of "Loyalty Unlocked." - If you enjoyed diving into the mechanics of customer behavior as much as I did, don’t forget to subscribe and share the pod. Catch you next time!