I work at Claspo, we build popup and email capture forms. The numbers below come from the stores running them, so it’s client data in aggregate, and I don’t have a shop of my own in here. Putting it in front of people who have tested discount depth for real, because it messed with my head.
I expected the obvious thing, that a bigger offer pulls in more signups. 10% off converted at around 4%. At 20% it fell to roughly 2.5%. Past 20% nothing moved, it sat near 2.6%, so the extra generosity bought nothing.
Two guesses. I trust neither much. An offer that looks too generous might read as sketchy to cold traffic that has never heard of the brand, and instead of typing an email the person stops to hunt for the catch. Or the comparison is junk, the offers never ran against identical audiences and something upstream did the work.
If the effect is real it’s expensive in a dumb way, because a store cutting deeper into margin ends up paying more per signup and collecting fewer of them.
Your second guess is the right one, and I think it goes further than the offers just being mismatched. Those two numbers almost certainly come from two different sets of stores rather than the same stores tested twice, and that makes depth and store identity impossible to pull apart.
Think about which stores pick 20 percent in the first place. Usually the ones with weaker repeat demand, thinner brand recall, discount-trained traffic, categories where everything is permanently on sale. Those stores convert worse at any depth. The offer depth is a symptom of the store, so what you measured is store quality wearing a discount label.
There is also a distribution question sitting under the 4 percent and the 2.5 percent. How many stores in each bucket, and how much of the traffic does the largest one carry? If two or three high-volume stores dominate the 20 percent group then the aggregate is basically their rate with rounding attached. Put the median store rate next to the pooled rate for each bucket and you will know inside ten minutes whether you have an effect or one big client.
The clean version is probably already in your data. Find stores that changed depth on the same form, same traffic mix, same season, then compare before and after within each store. Fifteen of those beats an aggregate across hundreds, because the store is held constant and only the depth moves.
Do you have enough depth changes in the history to cut it that way?
Hi there @Taras_claspo
That is an interesting result, and I would be reluctant to assume discount depth was the only variable. The quality of the audence, the timing of the pop-up, the source of the traffic, the product mix, and the brand familiarity can all affect sign-up behavior. A good next step would be to test 10% against 20% with the same audience and traffic pattern, and compare not only signup rate but also eventual purchase rate. Even if the larger discount brings in a smaller number of signups, it may still lead to a different quality of customers, so you’ll want to look at downstream conversion and revenue per captured e-mail to get a much clearer read.
The margin point is probably more important than the raw signup rate. If 10% gets twice as many signups as 20%, there’s not much reason to give away the extra margin unless those 20% subscribers turn into significantly better customers later. I’d want to compare eventual purchase rate and revenue per subscriber, not just the email capture.
I would not assume depth was the only variable here. A 20% offer often reads as a store that has to discount, so people slow down and look for the catch instead of typing an email. Before cutting deeper into margin, test 10% against 20% on the same audience, same traffic source, same weeks. Then judge it on revenue per captured email rather than signup rate, since deeper offers usually pull in worse repeat behaviour. In practice timing moves opt-ins more than the percentage: firing the popup after real intent (second page, scroll on a product page, exit) beats making the number bigger. The other lever is a quieter second path for the people who will never fill a form, so list growth is not being bought with margin. I would keep a sane 10% popup and spend the effort on when and where it fires.