Shopify Sidekick App Recommendations – A Nightmare for Developers Outside the Top 3?

Hello, with Shopify, the interests of the merchant come first—not those of the partner.

Hi Daniel,

You’re right—that’s precisely why I brought up the idea of ​​forming an alliance.

While thinking about it, a scenario came to mind:

Over the next three years, they’ll need people to clean up the low-quality content (the “slop”) generated by AI.
These individuals will enjoy the work and earn an income.
Once the time is right, the company will launch an AI tool—trained on the corrections made by these very people—designed to wipe out all that “slop” produced over the last five to ten years.
Suddenly, the forum will be flooded with complaints from these partners, who will find themselves shown the door without any prior warning.

This is just a hypothetical scenario, but something like it is quite likely to happen.

That is exactly why the proposal I’ve put forward would be to the benefit of all future partners.

@Paige-Shopify thank you this is useful context.

This explains the issue I have reported couple of times. It lists apps that match the words the user asked but not the feature they asked. Using just keywords like is undermining sidekick responses.
It should run a semantic search instead.

Why this is a problem?

With new AI based search (claude, chatgpt etc..) people are used the agent to turn their search into a semantic search. People are no longer searching for keywords. If the AI tool is based on keywords without them knowing it introduces a major searchability problem. Add to this that app listings have a limited number of keywords they can use it makes many searches impossible to land.

These days people are starting to blindly trust AI agents, if the agent tells them “here are the results”, they will probably not try further and stop at the first pass.
Especially most merchants are not technical people, so they don’t use AI tools as technical people would.


As an example this is happening a lot during my tests:

I behaved like I have a problem to solve with a store, I know what I want to do, but I don’t know anything about the existing app.

So I simply ask sidekick what I want to do.

  • Sometime it returns “this is not possible with shopify” look at the documentation,etc… and it does not offer apps.
  • Sometimes it searches for apps. It returns apps that matched the words I used. However very often they do not solve the problem that was asked. They’re just apps that use similar keyword in their listing despite serving other purpose.

Then I to refine the research. At that point if you’re not very accurate in what you want it’s hard to get relevant results.

Hi, @Umut-Dev

This is a good idea, but it wouldn’t come from Shopify and there is no guarantees it works. It be an initiative started by partners themselves. It would also requires to have enough trustworthy people who are not too busy to maintain it. I’d recommend opening a dedicated thread to talk about this.

G’day @Paige-Shopify

Thanks for your response!

I guess this is exactly where my issue lies. If all Sidekick does is turn a user’s question into a keyword search and return the top 3 results, then what’s the point?

This isn’t really AI-driven app discovery. It’s just a significantly worse version of the existing App Store search, one that only gives visibility to the top 3 apps.

Judging by the massive drop in install velocity many of us are seeing, merchants clearly used to explore beyond the first few results. Apps sitting at #4, #5, #6 and further down still had a chance to win merchants over with a well-written subtitle, a compelling listing, or simply by being a genuinely better fit for their needs.

Now, my apps are effectively invisible because we have this supposedly intelligent AI assistant that just recommends the same top 3 apps from a keyword search.

And I think the biggest problem is that merchants probably assume there’s some actual intelligence behind these recommendations. They trust that Sidekick has considered their requirements, compared the available options, and recommended the best matches. Why would they think to ask for more results when they believe AI has already done that work for them?

That’s what makes this so frustrating. Shopify has an opportunity to use AI to genuinely improve app discovery, helping merchants find the apps that best suit their specific needs, regardless of where they sit in the rankings.

Instead, we’ve seemingly ended up with a system that makes the existing rankings even more important, while removing the visibility and opportunities that smaller developers previously had.

I really struggle to see how that’s an improvement for either merchants or developers.

Often times it won’t recommend an app at all, it will recommend using metafields.

@Ollie_autoBlogger yes as I highlighted in my not-so-easy to read message… The problem is that the users expect a semantic search when they use AI tools. But in the end it just falls back to a good old keyword search. The user is not even informed of it.

They trust the intelligence of the model, as you said.

(of course that’s one top of the issue of showing only 3 results..)

Thanks @Paige-Shopify, I think everyone in this thread needs to read this ^^ part properly.

What Paige is describing is something I think we’re going to see a hell of a lot more of over the coming months.

We’re heading towards a seismic shift, and that a significant proportion of Shopify apps will become obsolete.

I worry that a lot of us are in denial about what’s coming. It’s easy to blame Sidekick for declining installs, and I’m sure it’s contributing, but I suspect the underlying problem is considerably bigger than that.

If your entire business depends on building apps on Shopify, I would seriously encourage you to start thinking about diversifying. We’re already having those conversations internally ourselves.

I don’t think we can afford to assume that fixing Sidekick’s recommendations is suddenly going to bring all those installs back.

@Alex_Hooper yes that’s true.

Regarding this point, I think there might be some denial, but there is also the problem that we’re not provided the numbers to make informed decisions.

We don’t know the intent of Shopify. We don’t know the usage of the users.

All that would be much easier if we didn’t have to guess.

I’m 100% sure is because something changed in the admin dashboard.

We have claude/chatgpt working with Shopify for at least 2 years, if merchants were using them to replace apps, we would see a gradual drop, instead we saw a wild 80% drop in some apps within 1~2 weeks during last update.

Hey all – jumping in here with a “sign of life” from my side :slight_smile:

@Paige-Shopify shared a lot of helpful info which is a good start. The feedback here has been really wide ranging (and sometimes heated, which is okay!) so I won’t be able to address everything individually – but the right people internally are reading this and taking it in.

On a personal level, I care a lot about the ecosystem having a diverse set of solutions for problems – I do agree that reinforcing the same top 3 apps for the rest of time is suboptimal for merchants, partners, and the ecosystem at large.

I know that what you want is for me to say that it’s all going to change right away, and I can’t say that – but I do expect this to keep improving. App discovery continues to be a hard problem that we’re working on.

Hey @Blair Thank you for taking the time to respond and for being open about this. It’s really appreciated!

I think we all hope that we’ll be able to mark this thread as solved in the near future. Fair app discovery is an important topic for all partners across the Shopify ecosystem.

If anyone has constructive suggestions or ideas on how to improve the current situation, I think this is a great place to share them. I’d just kindly ask everyone to keep the discussion respectful, constructive, and on equal footing.

Thanks again for listening and taking our feedback seriously. Looking forward to positive improvements! :slightly_smiling_face:

Thanks @Blair !

Shooting the shit here. But… What if Sidekick searched for the relevant keywords on behalf of the merchant, but instead of recommending just the top 3 apps within Sidekick itself, it provided a link to those search results in the App Store?

That way, Sidekick would still be helping merchants find relevant apps, but without presenting what is essentially a basic keyword search as an intelligent, personalised recommendation.

It could be the best of both worlds. Merchants would still benefit from Sidekick doing the searching for them, but they’d also have the opportunity to browse the full results, compare apps, and potentially discover something better suited to their needs.

I’m not saying this is necessarily the solution, but I’m just trying to come up with something that might work!

Something really needs to change here. For many of us independent developers, these apps are our livelihoods, and the impact of effectively making anything outside the top 3 invisible is pretty significant.

We also still have major issues with fake and purchased reviews influencing App Store rankings, including many in the top 3 of certain categories. By limiting Sidekick’s recommendations to those same top 3 results, this problem is also being compounded.

@Blair - This really isn’t good enough.

I’ve been doing some further testing of Sidekick, and the results are pretty concerning.

Basically, I asked it to recommend an app with a specific combination of features that I know my app has, and that none of the other apps in the category offer. All of these features are also clearly mentioned in my App Store listing.

Of course, it didn’t recommend my app, because it’s currently sitting in 6th position in the search results. Instead, it recommended apps that didn’t even meet all the requirements I’d given it.

When I pushed Sidekick to explain why, it openly admitted that it had relied on the top search results, hadn’t properly checked whether the recommended apps met my requirements, and had stopped searching too soon. It then acknowledged that my app should have been its first recommendation.

Shopify has specifically promoted Sidekick as a way to intelligently recommend apps based on merchants’ needs. But, once again, how is this intelligent or personalised when it’s effectively just returning the first three search results, regardless of whether those apps actually meet the merchant’s requirements?

The merchant isn’t being told that they’re only seeing the top three search results. They’re being presented with what appears to be a carefully considered AI recommendation, and they’re naturally going to trust that Sidekick has actually looked for the best app for their needs. This feel super misleading, especially with how Sidekick is being positioned.

Again - sidekick seems more aware then anyone?

Hey Ollie – yes, I understand your perspective and agree. Sidekick app recommendation should be better and will improve. I have nothing else I can share at this time.

Looks like there are some minor but important changes in the app search UI in the admin.

Last week, whenever you searched for an app here, the primary CTA after a couple of results was always Ask Sidekick Now it’s Search App Store.

Also, I’ve never seen this search return more than 3 results before, but now it shows more results with a “Load more” option. (Edit: The same keyword returns different numbers of results on different stores, so you might still see only 1 or 2 results)

Also, If the search doesn’t return any results, the Ask Sidekick CTA appears alongside Search App Store. Search App Stor wasn’t shown there earlier at all.


Thanks @Blair , I appreciate the response, and I understand ths is a evolving situation. Also, apoligies for sounding like a broken record!

But just to reiterate: A lot of us solo devs are really struggling with this one. My apps usually sit around 4th, 5th or 6th for their main keywords, which I thought wasn’t too bad, and which worked well with the old App Store. But since this update, my installs have dropped basically overnight to almost zero, and I’m getting virtually no traffic through Sidekick.

You’ll notice there’s nobody on here saying, “Woo, my traffic has shot up thanks to Sidekick!” And I think that says a lot.

The current setup is having a massive impact on small, honest developers, while seemingly benefiting only the biggest players and those already dominating the top 3.

I really hope this is something Shopify takes seriously, because for many of us, these apps are our livelihoods.

This looks like a good step!

It could be good to add a ‘Search App Store’ button beneath Sidekick’s recommendations, allowing merchants to easily explore more options. Along side a disclaimer making it clear that Sidekick results aren’t necessarily personalised recommendations, but simply the top three results for a particular keyword search. I feel the current setup is super misleading.

Please keep the feedback coming in! I’ve been forwarding the feedback to our SIdekick developers.

The side kick autocomplete recommendations is actually recommending apps that have 0 reviews, while other apps with BFS badge & 100+ reviews are not being recommended.