Two years ago a merchant found your app by browsing a category. Today they arrive already knowing your name, and they are checking whether you are the one they think you are.
Discount Prime has been live on the Shopify App Store for two years this week. The most interesting change over those two years is not in the app. It is in how merchants decide to install it. Discovery moved upstream, evaluation got faster, and in the last few months a genuinely new install source started showing up. Here is the honest version of what we have watched.
Merchants arrive pre-researched now
Two years ago, a meaningful share of installs came from merchants browsing the discounts category, comparing listings side by side, reading their way to a decision inside the store. That still happens, but it is no longer the main road.
Most merchants now show up with a shortlist. They built it somewhere else: a search, a comparison article, a Reddit thread, a recommendation in a community. By the time they reach the listing, they are not asking "which app should I use." They are asking "is this the one I already think it is." The listing stopped being where the decision starts. It became where the decision gets confirmed.
That reframed how we think about the store page. A page written to win a cold browse and a page written to confirm a near-decision are different pages. The second one leads with the specific job, shows it working on a realistic catalog, and keeps the pricing legible. It answers the last doubt, not the first curiosity.
Reviews became verification, not discovery
Reviews used to do discovery work. A high star count pulled a merchant in. Now the star count is table stakes, and the job of a review is to verify a choice the merchant has mostly made.
That changes which reviews carry weight. A recent, specific review that names a real workflow ("we run wholesale tiers by customer tag and it just worked") does more than ten that say "great app, fast support." The specific one answers the question the merchant actually has: will this handle my case. We stopped chasing volume of reviews and started caring about whether recent ones described concrete use.
Evaluation windows got shorter
The second shift is speed. Merchants decide faster, often inside the first session. They install, they point the app at their real catalog, and they judge within minutes whether the core job works. An app that needs an afternoon to demonstrate value now loses to one that proves itself before the coffee is cold.
For us that meant treating the first ten minutes as the product. Setup that assumes a real store with messy data, a default that does something useful immediately, and a clear read on whether it is working. The analytics view matters here too: a merchant evaluating an app wants to see that it can show its own results, not just perform an action.
The new one: AI-referred installs
The genuinely new development this year is small but real. We have started seeing installs from merchants who found us through an AI assistant. They were researching a problem, an assistant summarized the options, and our app was in that summary.
These arrivals behave differently. They tend to come pre-qualified, because the assistant already matched the app to a stated need before the merchant ever clicked. They ask sharper questions. They are further along. It is early, and the volume is modest, but the direction is unmistakable: some of the discovery that used to happen through browsing and search is now happening through a model reading comparison content and handing back a shortlist.
The practical response is not a trick. It is to make sure the plain facts about what the app does, who it is for, and how it prices are stated clearly enough that a model can summarize them correctly. The same clarity that helps a rushed merchant helps an assistant reading on their behalf. Our comparison and category content exists partly for exactly this reader.
What we take from year two
Discovery moved upstream, so the listing confirms more than it convinces. Reviews verify instead of attract, so specificity beats volume. Evaluation compressed, so the first ten minutes is the pitch. And a new referrer arrived that reads before the merchant does.
For the longer story of what these two years cost to build, the one-year retrospective is where this thread started, and our note on the roadmap debt of a small app covers what we had to cut to get here. To every merchant who shortlisted us, installed, and then told us what was still confusing: that feedback is the product.
