Find the Sweet Spot for Ecommerce Loyalty Rewards

I set the first meaningful reward just beyond what a customer can earn from joining, referrals, or other non-purchase actions. That means they see progress immediately, but they still need to complete a purchase before real value is unlocked. I also prefer a minimum spend on redemption rather than making the reward larger, because it protects margin while giving the customer something attainable to work towards. The simple rule is that loyalty should encourage the next purchase, not subsidise the one they were already going to make.

My first instinct was to set reward thresholds based on average order value, but that penalized customers who buy one product at a time. So I pulled 90 days of purchase data and looked at how many orders a repeat buyer typically placed before going dormant. That number was three. When someone came back a third time and got nothing, they often stayed away on the fourth.

I pegged the first perk at a level reachable within two purchases. A small discount on the third order, enough to feel earned but low enough that my margin on a functional product with strong repeat demand could absorb it. The threshold sat just above the average cart total for my returning buyers, so most people had to add one item or place one more order to unlock it. That tiny gap kept engagement high and avoided training customers to wait for a discount before buying.

The first tier stayed easy to reach, and that kept the program working. Customers who found relief from their purchase already had a reason to come back, and the loyalty perk shortened the gap between their orders. I've tested higher thresholds and flashier perks at later tiers, but the modest reward on that third transaction still drives my most consistent retention.

We approached thresholds like staircase design. If the first step feels too high, people may never start climbing. If each step feels too shallow, the program can lose value before loyalty grows. So we built the opening threshold around a behavior we already wanted more of, which was making another purchase within a set period.

The model worked because we tied thresholds to real reorder patterns instead of guesswork. We also tracked how many members stopped just before each milestone. When too many members stalled at the same point, we knew the threshold needed a closer look. This simple check helped us keep progress clear for customers while protecting the program's value over time.

The threshold rule that held up for us is that the first reward has to be reachable on a customer's natural second purchase. Not a stretch, not a goal, just the thing they were going to do anyway.

Our first attempt got this wrong in the way most programmes do. We set the first tier where it looked meaningful on a spreadsheet, which was several orders out, and the effect was that members joined, looked once, understood they were nowhere near it and never thought about it again. Moving the threshold in so it lands on the second order took 55% of members over it, and that changed the psychology entirely, because a programme you have already won something from is a programme you keep an eye on.

The other decision that protected margin was making the first perk non-monetary. Early access to a new product, and a sample of the next thing we are making, rather than money off. Discounts as a first reward train people to wait for the discount and quietly reset what they think the product is worth. Access costs us almost nothing and it flatters people, which is most of what a loyalty programme is doing.

Points that expire are the part I would avoid. They convert a small reward into a small punishment at exactly the moment you were hoping to hear from somebody again.