Reworked my app after your liability feedback. Now deciding: keep going or move on — what would you do?
A few weeks ago I posted here about Skinin, a Shopify app showing ingredient info on product pages. You all flagged that "safe/low concern" labels were a real regulatory liability risk, not just a UX issue. I took it seriously — rewrote every label to be disclosure-based instead of evaluative, audited 1,500+ ingredients and fixed 289 mislabeled entries, added proper source citations.
Here's where I actually am right now: I set myself a 1-2 week validation window that ends tomorrow, with three criteria for "keep going" — 3+ brands genuinely interested, 1 real review, or clear consumer-side signal that people care about ingredient transparency. I have 1 active install (a clean beauty brand I've been fixing bugs for directly), 0 reviews yet, and some early consumer interest I'm still reading the tea leaves on.
I'm not cleanly hitting all three. Genuinely asking: at this stage, with this data, would you call it "keep pushing" or "cut losses and move to the next idea"? What's your bar for making that call?
One thing not mentioned yet: the cleanup you did is a one-time event. The regulatory problem you're solving isn't. EU cosmetics regs update, FDA guidance shifts, ECHA adds substances to the concern list. Your stickiest users won't be the ones who needed correct labels today - they'll be the brands that need to stay correct.
If "auto-updates as regs change" isn't in your pitch yet, that's where I'd look. Apps solving recurring compliance problems have much lower churn than apps solving a one-time setup problem.
On the validation question: I've been running pre-order validation on a similar regulatory monitoring tool for small businesses (billwatch-landing.vercel.app - tracks legislative bills affecting SMBs). Same lesson as JohnMadison's - "interested" and "will pay" are completely different people. One filter I'd add to the 10-brand outreach: only target brands already doing manual compliance work themselves. Those people feel the ongoing pain. Everyone else just thinks compliance might be a problem someday.
The "one-time vs recurring" framing is the most useful thing in this thread for me — I've been thinking about the label fix as "the problem solved" rather than "the first instance of a problem that keeps happening." Pitching it as "stays correct as ECHA/EU/FDA lists change" instead of "we fixed our labels" is a genuinely different pitch, and probably a different buyer too.
Stealing your filter as well — restricting the 10-brand outreach to brands already doing manual compliance work themselves (not just brands that could theoretically care) should cut out a lot of the "interesting idea" noise before I even send anything.
Will check out billwatch — curious how pre-order validation goes for a problem that's similarly invisible until it bites you.
I've really enjoyed these conversations too. It's been refreshing to exchange ideas with someone who's focused on the underlying problems rather than just shipping features.
You can find me on X here: https://x.com/StacyWycof83995
Looking forward to staying in touch and comparing notes as we both keep building.
Hey @SkinInLabs i think platform has some thing like limit on comments and replies etc hahaha
I think that's a much stronger success metric.
Even if only one person asks a thoughtful question, you've learned that the positioning made someone stop thinking "interesting idea" and start thinking "how would this actually work for me?"That's a very different kind of signal than a higher reply count.
Ha, yeah I've been hitting walls left and right on here — should've expected a platform that values signal over noise to have limits on replies too 😄
And that one-question shift is exactly what I'm watching for. "Interesting" keeps you comfortable, "how would this work for me" actually means something. Follow-up just went out today — fingers crossed someone asks a real question instead of just being polite.
😄 Looks like we've managed to hit the reply limit.
I've really enjoyed these conversations over the last week. It's rare to find someone who's thinking this deeply about founder problems instead of just features.If you're active on X, send me your handle. I did love to stay connected and keep comparing notes there.
I actually respect founders who share moments like this publicly.
Sometimes the fastest progress comes from finding out what doesn't resonate.
What's the strongest argument that's making you consider continuing?
Honestly? The Luverta relationship. They installed during the messiest version of this thing, and surfaced three separate real bugs just through normal use — not a planned test, just someone actually relying on it. I fixed all of them directly with them. That's a different kind of signal than an "interesting idea" comment.
The second thing is narrower, and I only found out today: my own App Store listing and marketing page were quietly still making the exact safety-verdict claims my product doesn't make anymore. If that's been the pitch reaching merchants this whole time, the weak signal so far might partly be a "wrong pitch" problem, not a "no demand" problem. I genuinely don't know which it is — that's the actual reason I want one more narrow test instead of guessing in either direction.
I think that's a really healthy way to look at it.
The Luverta feedback feels different because they weren't evaluating an idea, they were relying on the product to do real work. Those are usually the conversations I trust the most because people stop talking in hypotheticals and start exposing the gaps that actually matter.Your second point caught my attention too. It's surprising how often positioning ends up being the hidden variable. Sometimes the product isn't failing—the story people hear about the product is.
I'd be really interested to see what happens after that messaging update. It feels like a much cleaner test than making big product changes before you know what the market is actually reacting to.
"The story people hear about the product is failing" is a much better way to put it than how I'd been thinking about it. Will report back once the Phase 2 outreach plays out either way — should be a pretty clean read on whether that's actually what was going on.
I think that's exactly the right approach.
One thing I've learned is that a positioning change often gives you a much clearer signal than another feature release. If the product solves a real problem but people are arriving with the wrong expectation, every conversation after that starts from the wrong place.I'm genuinely interested to see what you learn from Phase 2, even if the outcome isn't what you expect. Sometimes a clear "no" is just as valuable because it tells you where the real bottleneck isn't.Definitely share an update when you've run the outreach. I'll be curious to see whether the messaging changes the quality of conversations as much as the quantity.
The "tells you where the bottleneck isn't" framing is a good way to hold a "no" — makes it useful data instead of just a disappointment.
Will definitely report back. Honestly the quality-vs-quantity question is the one I'm most curious about myself — if the new positioning works, I'd expect fewer but more substantive replies, not just a higher reply rate. Quantity alone wouldn't really tell me much.
I think that's a subtle but important distinction.
Higher reply rates tell you people noticed the message. Higher-quality conversations tell you they understood it.If Phase 2 changes who responds and what they ask, that probably says more about the positioning than the raw numbers ever could
That's a sharper way to put it than I had. It also gives me something concrete to actually watch for — every outreach so far has gotten either silence or a generic "looks interesting," never a clarifying question. If Phase 2 gets even one person asking something specific about how it works, that alone would tell me more than five generic replies would.
I think the first clarifying question is often the moment interest becomes intent.
Generic replies keep the conversation at the surface. Specific questions usually mean someone is already trying to fit the product into their own world.At that point you're no longer measuring attention. You're measuring whether people can see themselves actually using what you've built.
That's a clean way to draw the line — a specific question means someone's trying it on, a generic reply means they're just being polite. Given everything so far has been either silence or vague "looks interesting," that bar suddenly feels a lot more useful than just counting replies. Phase 2 follow-up goes out tomorrow — curious whether anything clears that bar this time.
This comment was deleted a month ago.
Respecting your own criteria instead of quietly moving the goalposts is the hard part, and you're doing it. In trust sensitive categories the liability rewrite is often the thing that makes someone willing to install at all, since beauty and health buyers avoid anything making safety claims that could get them sued. Wean, my nicotine taper app, sits in that same trust heavy space and being careful about what I claimed mattered more than any feature. Have you put the new cited version in front of those clean beauty brands yet, or is the read still coming off the old listing?
Honest answer: no, not yet. The cited/disclosure-based copy only went live yesterday, and today's the deadline I originally set myself — so none of the existing installs or outreach actually happened against the version that exists right now. Everything so far reflects either the old evaluative-label listing, or no pitch at all.
Combined with the point a few comments up about the hero image still showing stale labels, I think that settles it: I haven't tested demand for the current product yet, just an earlier draft of it. Going to put the new framing in front of the target brands directly before treating any of this as a verdict.
The Wean parallel is helpful too — makes sense that claims-discipline matters more than feature count once a wrong claim is the actual risk, not just a UX nitpick.
Smart to reset the clock instead of calling it on data that never tested the real product. The one thing that moved the needle for me with cold outreach was making the ask tiny, so a yes barely costs them anything. Good luck with the brand pitches.
Appreciate that, especially coming from someone who's actually tested this. My ask is already close to that (install + 15 min feedback, or just a flat no) — but good reminder to make it even smaller if the first few brands hesitate. Thanks for following along on this.
Good to see the update, and the relabeling work sounds like exactly the right call. Also glad the "solo developer who replies personally, same day" line made it onto the listing. That's a real trust signal doing the work a review would, nice.
One thing on the new description though: you've made liability-avoidance the headline. "Without the liability risk of a safety rating" leads with caution and a developer's worry, not the merchant's desire. A store owner wants "this builds trust and sells," not "this protects me from a risk I didn't know I had." You fixed the product correctly by going disclosure-based. Just don't let the fix become the pitch. Lead with the upside, that transparent ingredients build shopper trust and confidence, and keep "avoids regulatory risk" as the reassuring bullet it already is further down.
And gently: the copy got a real upgrade, the images didn't. For a visual widget that gallery is still your biggest lever. One large shot or a short clip of it live on a product page, no click to expand, would do more than any line of text.
On your actual question, keep or cut: you set good criteria and you're not hitting them, which is honest signal worth respecting, not explaining away. But you ran that two-week test on a listing whose visuals never materially changed, so it wasn't a fully fair read on demand. I'd fix that one screenshot, give it a short extension, and then make the call, rather than killing an idea your own storefront never actually showed.
Really useful catch on the headline — you're right that "avoids the liability risk of a safety rating" solves my anxiety, not the merchant's problem. Moving the trust/upside framing to the front and pushing the liability line down to a supporting bullet where it actually belongs.
The image gap is real too — I'd actually flagged it myself a day ago and deprioritized it given the deadline, which I think was the wrong call in hindsight. The gallery's hero image still shows the old "Worth knowing about" labels that don't even match the rewritten copy underneath it anymore. Good catch landing on the same thing independently.
Taking the point on not treating this as a clean test. Going to fix that one hero image before making the call, rather than judging two weeks against a listing that was quietly working against itself the whole time.
This comment was deleted a month ago.
I'd give it one more very narrow test, but I would change the bar from "interest" to "behavior."
The liability cleanup is not just cleanup. In trust-sensitive categories, it can be the thing that makes someone willing to install at all. But I would not spend another open-ended month polishing.
My suggested next loop:
If you get 2 to 3 brands willing to actually try it, keep going. If it stays at "interesting, good idea," I would probably cut it or reposition.
This is exactly the reframe I needed — "interest" was letting me count things that felt like progress but weren't real decisions. Switching to "only count install + feedback, or a clear no" removes the wiggle room I'd been quietly giving myself.
Also appreciate you separating the liability cleanup from "is this idea good." That's been nagging at me all week — it's table stakes for being installable at all, not evidence the idea actually works.
Going to run this exact loop: 10 brands already positioning around ingredient transparency, one clear ask, count actions only. One question — does my existing install (where I've been doing free bug-fixing directly for them) count toward the 2-3, or should that be treated as a separate/biased data point since there's already a relationship there?
I’d probably not call this a clear “move on” yet, but I also wouldn’t treat it as fully validated.
The strongest signal here is that you took a real objection seriously and turned it into a more trustworthy product. That matters, especially in a category where liability and credibility are part of the value proposition.
My bar would be: one more short validation loop, but narrower. Instead of asking broadly whether people care about ingredient transparency, I’d focus on whether 3–5 clean beauty brands will take a specific next step: install, give a review, agree to a short pilot, or explain why they won’t.
If the answer is still vague interest after that, I’d move on. But one active install plus a much safer product feels like enough signal for one final focused test.
This is a really fair distinction — "took an objection seriously" being the strongest signal so far, separate from whether the core idea is validated. That reframe is helpful.
Funny enough, someone else in this thread independently landed on almost the same shape of advice — narrow the sample, count actions not interest, one more focused loop before deciding. Two people converging on that without coordinating makes me trust it more than either comment alone.
Going with 3-5 clean beauty brands, one specific ask each (install + review, or short pilot, or a clear no), and I'll exclude my current install from the count since I've already got a hands-on relationship there that'd bias the read. Appreciate you taking the time to actually engage with the specifics instead of just cheering the cleanup work.
Glad it helped — that sounds like a much cleaner validation loop.
I like the idea of counting actions instead of interest. A polite “sounds cool” can be misleading, but install/review, short pilot, or a clear no gives you much better signal.
One thing I’d also track is the reason behind each no. Sometimes the idea is right but the timing, wording, or onboarding step is the blocker.
Hope the 3–5 brand test gives you a clear signal — would be interested to hear what pattern shows up.
Tracking the reason behind each no is a good addition — I've been treating "no response" and "not a fit" as equivalent, but they're actually pointing at different problems. No response could be wording or timing; a clear "not interested" tells me something about the idea itself. Going to flag that distinction in how I log Phase 2 outcomes.
Follow-up goes out today. Will report back on what pattern shows up — curious whether the positioning change shows up in the quality of the nos as much as the yeses.
That makes sense. I like that you’re turning the next step into a clean test instead of just debating the idea forever.
Counting actual actions — install, review, pilot, or clear no — is much better than counting polite interest.
I’m learning the same thing with my own product. The hard part is not getting people to say “this sounds useful,” it’s getting them to take the next real step.
Curious to see what the 3–5 brand test shows. That should give a much clearer signal.
What jumped out at me wasn't the outcome.
It was the role the outcome is being asked to play.
Reading this, it feels like tomorrow's decision depends heavily on whether those three validation criteria are actually the right lens for interpreting what's happened so far.
A founder can miss a target and continue.
Another can hit the target and quit.
The visible result matters less than what the result is being treated as evidence of.
That's the part I found most interesting.
Fair callout, and it's making me actually think about this instead of just running the checklist. Honestly, the three criteria aren't a precise measurement — they're more of a forcing function. I know myself well enough to know that without a pre-committed bar, I'd rationalize either outcome after the fact: good signal → "see, worth continuing," weak signal → "just needs more time." So your comment is making me ask the better question: what was each criterion actually supposed to be evidence of?
A review = someone got enough value to bother writing about it unprompted. 3 brands = repeatable interest, not a fluke. Consumer reaction = the underlying problem resonates, not just my specific solution to it.
If tomorrow's numbers come back mixed, I think I'll weigh those "what is this evidence of" questions more than the raw count. Appreciate you pushing on this.
That’s a solid way of framing it.
If you want, I can take a closer look at how those criteria interact with actual user signals once you’ve got a bit more data — probably easier over email than in-thread if you’re open to it.
Appreciate the offer — happy to keep it here in the thread if anything comes up, easier for others following along too.