Building WinnerSpy: reducing noise in Facebook Ads Library product research
I built WinnerSpy because product research through Facebook Ads Library still feels too noisy and manual.
Most of the time, people end up scrolling through a huge number of ads and trying to guess which products are actually worth testing. Looking at ad count alone usually isn’t enough, and a lot of time gets wasted on weak signals.
With WinnerSpy, I’m trying to make that process faster and more structured by surfacing products worth testing using signals like:
ad longevity
multi-page activity
creative spread
store filtering
It’s still early, but it’s already helping me narrow down ideas much faster.
I’m sharing it here to document the build and get feedback from people doing ecommerce research, paid social, or media buying.
Curious: when you research products from Facebook Ads Library, what signals do you trust most?
Totally agree on ad count not being enough. I’ve seen plenty of products with lots of ads that still don’t convert. Longevity is probably the most underrated signal.
Exactly. Ad count can be useful, but on its own it’s a pretty weak signal.
I’ve seen the same thing — some products generate a lot of ad activity without turning into anything durable. Longevity starts to matter more when it’s paired with things like multi-page spread, creative variation, or reactivation after a pause.
The signal I've found most reliable is creative spread across multiple independent pages — it's the hardest to fake and usually means someone is actually scaling with real budget behind it. A single advertiser running the same ad for 30 days could just be a stubborn founder; multiple independent pages running the same product is a much stronger validation signal.
The store filtering angle is underrated too. A lot of the noise in the Ads Library comes from low-quality dropship stores that run ads briefly and disappear. Being able to filter those out quickly changes the quality of everything downstream.
One thing I'd be curious about: are you planning to track creative evolution over time? Sometimes the most interesting signal isn't the original ad but whether the advertiser is iterating on the creative — that usually means they're profitable enough to keep testing, which is a different kind of confidence than just longevity.
Exactly — that’s the core thesis. The advantage isn’t more raw ad data, it’s removing noise and surfacing signals that are actually actionable.
Good angle. The real value is not more ad data, but better filtering and clearer signals.
Appreciate that — that’s exactly the problem I’m trying to solve. Ads Library has a ton of data already, but most of the value comes from filtering the junk and making the useful patterns easier to spot.
WinnerSpy simplifies Facebook Ads Library research by filtering low-quality ads, highlighting winning products, tracking trends, and saving time, helping marketers make faster, smarter, data-driven decisions.
Really good points. I agree longevity is one of the strongest baseline signals, but I’m also interested in cases where the same product shows up across multiple independent accounts — that feels even harder to fake. And yes, the goal is definitely to reduce noise rather than just dump more data on people. Still early on the build, but I’m exploring the best way to structure the data layer around public Ads Library data.
Ad longevity is the signal I trust most. If someone's been running the same creative for 30+ days, they're profitable — nobody burns money on ads that don't convert for a month straight. Everything else can be gamed or misleading.
The multi-page activity signal is interesting though. Are you tracking whether the same product shows up across multiple ad accounts? That would be a stronger signal than one store running lots of ads — it would mean multiple independent sellers validated the same product.
Smart to focus on reducing noise rather than adding more data. The Facebook Ads Library is drowning in information — the value is in filtering, not in showing more. Same principle I'm applying in a different space — crypto converters show you a million numbers, but the one that matters is what your coins actually buy in real stuff.
What's your stack? And are you scraping the Ads Library API or building on top of their public data?
Really good points — I agree ad longevity is one of the strongest baseline signals, and multi-account / multi-page overlap is probably an even stronger one because it’s much harder to fake. That’s exactly the direction I’m interested in: reducing noise and surfacing higher-quality signals instead of just showing more raw data. Still early, but I’m exploring ways to track those patterns more clearly over time. Right now I’m building on top of public Ads Library data rather than relying on anything private.
signal quality vs signal interpretation is the hard part. ad longevity sometimes means the market is already crowded, not that the product is worth testing.
Totally agree — that’s the hard part.
A signal can be real without being actionable. Longevity by itself doesn’t automatically mean “test this.” In some cases it probably means the market is already saturated.
What I’m trying to get better at is separating stable demand from late-stage crowding, which is why I’m looking at combinations like longevity + multi-page activity + reactivation, instead of treating any one metric as enough on its own.
yeah that's the right frame. longevity confirms the category is real - doesn't tell you if there's still an opening.
You're asking the right question. ~
Accessing the Ads Library usually isn't the struggle. The struggle is knowing what is truly worth paying attention to inside the Library.
Many business owners interpret action as demand. Some of that IS demand. Some of it is that the business has good margins, a good post-click offering or a large enough budget to continue to run tests.
What you are trying to determine is if the information is a tool for someone to make a better decision.
Does it suggest there is something genuinely worthwhile?
Does it save them time?
Does it clarify what the next step is?
Those are valuable things.
One approach that might improve the product would be to illustrate why it was highlighted. Beyond a simple score or label, provide a simple explanation for the highlighting. People will be much more trusting of the tool if they know why it is displaying something.
It is also valuable to note which signals tend to be misleading. Knowing what not to follow can be just as informative as knowing what to follow.
This is a really thoughtful take, and I think you’re pointing at the actual challenge.
Access to the library isn’t the bottleneck anymore — interpretation is.
A lot of activity looks like demand on the surface, but it can also come from strong margins, better offers, or just enough budget to keep testing. So the goal isn’t just to surface movement, it’s to help people decide what’s actually worth a closer look.
I also agree on the explanation layer. A score by itself is useful, but trust goes up a lot when the tool can say why something was highlighted and which signals might be misleading.
That’s a big part of how I’m thinking about the product now: not just showing more data, but making the next decision clearer.
Ad longevity combined with multi-page activity is the right combination to focus on. A single page running one ad for 30 days could just be stubbornness but when you see the same creative spread across multiple pages it usually means someone is actually scaling it with budget behind it. That signal is hard to fake and harder to spot manually which is exactly why the library feels so noisy without something like this. The store filtering piece is underrated too since a lot of dropship noise disappears when you can filter by store type. Curious how you are handling ads that go dark and come back after a gap because that pattern sometimes signals a product being retested after a supplier change or a new market.
Exactly — that’s the signal I care about most.
A single long-running ad can mean something, but when the same product or creative starts showing up across multiple pages, that usually points to real scaling rather than just one advertiser keeping it alive.
And yes, store filtering helps a lot with cutting out low-quality dropship noise.
The “go dark and come back” pattern is something I’m looking at next. I don’t want to treat it the same as pure longevity because in a lot of cases it’s more of a reactivation signal than a stability signal.
Love the focus on reactivation—that’s a much stronger signal for a winner than just pure stability. Since you mentioned being in the early stages one thing to think about is the 'trust barrier' in this niche. Professional media buyers and e-com founders are usually pretty skeptical of new tools until they see them being talked about in high-authority spaces. I’ve noticed that when tools like this get even a bit of coverage in major tech or finance media it completely changes the perception of the data’s accuracy. It moves the needle from 'just another spy tool' to an industry-verified resource. Looking forward to seeing how that reactivation feature turns out!
Really good point.
I think trust is a huge part of this space, especially because most people have seen too many noisy or overpromising spy tools already. My focus right now is making the signals themselves feel reliable first, especially around reactivation, multi-page spread, and filtering out weaker store patterns.
You’re right though — once a tool starts getting discussed in more credible circles, people evaluate it very differently. But I’d rather earn that through data quality and repeatable results first, then let the broader coverage follow.
Appreciate you calling that out.
Makes total sense. Building a 'Product-Led' brand is much more sustainable than just hype. Most founders rush the PR and then fail when the product doesn't hold up under the spotlight—so it's refreshing to see you prioritizing the reliability of the signals first.
Good data usually speaks for itself anyway. I’ll be keeping an eye on WinnerSpy as it evolves. Whenever you feel the product is 'battle-tested' and you're ready to flip the switch on that authority building let me know. Happy to brainstorm more then!