A creator refused to endorse my product. Here's what I fixed because of it.

The creator gave you a specific reason she couldn't stand behind the score, and you made the sources and scoring method inspectable. That's a better test than asking for general feedback again. I'd send the fixed version back to the same creator without asking for an endorsement - just ask whether the evidence is now checkable enough for her. What would count as a credible result of that retest for you, even if she still declines to endorse it?

exactly. totally agree with that. keep it up!

That creator gave you the best possible gift: a falsifiable objection.

I’ve started applying the same bar to competitive / market claims I make publicly. If I can’t click through to the original changelog, pricing page, status incident, or Reddit thread, it doesn’t go in the brief — period. Domain names and paraphrases are how “we think X shipped Y” turns into confident nonsense two weeks later.

Two small extras that help once sources are clickable: (1) timestamp the snapshot (pages move), and (2) separate “observed fact” from “our interpretation” in the UI so reviewers don’t have to reverse-engineer your confidence. Sounds like you’re already most of the way there.

TruthScore still at $0 after fixing the exact “can’t verify the claims” objection is such a hard quiet — the product got better, distribution didn’t. Soft free text-only week for that trust-fix→still-silent stall. If you want, reply with the one place you’d put a public scorecard if the next step was five skeptics who click a source, not another cold email.

The part that stuck with me was the contradiction you pointed out: a product built around skepticism was itself asking people to trust its conclusions.

That’s a much deeper product lesson than just “add clickable sources.” You basically found a gap between the promise of the product and the experience of using it.

I’m curious about one thing though: now that users can verify the evidence themselves, do you expect that to change the actual scoring experience too? For example, could showing the evidence make people disagree with the score more often, even if it makes them trust the product more?

A Reddit claim can be perfectly linked and still be weak or coordinated. I’d consider an evidence card showing the source type, retrieval date, independence, and exactly what the source supports versus what TruthScore inferred. Are you measuring whether people actually open the evidence before deciding the score is credible?

The rejection is more useful than another feature, but only if she can see the fix without redoing the whole evaluation. Send her one before-and-after on the exact claim she could not click: the old "Reddit reports say X" line next to the new link, plus the method-page section that says where that signal can be wrong. Ask her to check that one claim, not to endorse the product. If she still will not, that leftover objection is the one to build against, because "I cannot verify this" was the stated reason and you just removed it. A lead who asked for a write-up and went quiet often got a restatement of the product instead of an answer to the hard question they already asked.

the telling detail is that she didn't reject the score, she rejected the unverifiable citation. for trust tools, provenance is the product: the score is a claim, the clickable source is the evidence, and nobody repeats a claim they can't check.

the "where it can get it wrong" section is doing more work than the weights. a stated failure mode is the fastest credibility builder there is, because it tells the reader you already asked the question they're asking. tools that admit their edges get recommended. tools that claim none get tested.

Clickable sources fix the objection she raised, and "asking to be trusted blindly" is a sharp way to put it. The next skeptic will probably ask a different question, though: how often is the score itself wrong?

Have you run it on sites where you already know the answer, some confirmed scams and some confirmed legitimate ones, and counted the misses? For a scam tool, a wrong flag on a real business is the expensive error. If your methodology page said "we ran it on N known cases and got X wrong," I think that would do more for a skeptic than the source links.

Totally agree. It's easy to get distracted by premature scaling when the real bottleneck is distribution.

This is a great mindset. Whatever the numbers do next, you've got a before and after now, and that's worth more than guessing.

I went through something similar. I let an AI agent build my first product, then actually checked its work: 43 of 49 entries had errors. It stung, but it's the only reason the thing is worth anything now. Getting comfortable is the dangerous part. Growth is uncomfortable, and that's the point.

Curious: are you tracking anything besides paid conversions to see whether the trust fix is landing?

Your last line is the part more of us should copy: "I'd rather hear the real reason than nothing at all."

I'm doing a version of that this weekend. 31 signups, none reached the point where our product does its job, so I'm emailing each one with a single question and four numbered answers they can reply with. Lowering the effort to one character felt like the only way to get the silent ones to talk.

Did she reply after you sent her the fixed version?

One tactic worth stealing from this: when a skeptic says no with a specific reason, don't just fix it — send the fix back to them with one question: "Is this what you meant?" A critic who flips after naming the flaw endorses harder than a lukewarm yes, because they've already made the case against you in their head and lost it. Track every rejection in a sheet, too — after a dozen you'll see the same 2-3 objections repeat, and those are your real roadmap. One caution: she's a credibility lender, not a buyer. Make sure the source links and methodology page answer your paying users' version of "is this real" too, or you'll end up building for the critic instead of the customer.

Real respect for actually fixing something instead of just moving on to the next creator. The "no" that makes you better is rarer than it should be — most rejections just get shrugged off.

Two observations. First, go back to the same creator with the fixed version — a skeptic who flips after seeing her objection addressed is worth ten paid endorsements, and her audience is literally the trust-paranoid segment you're trying to reach. Second, clickable sources and a public methodology page are credibility table stakes, but they don't solve distribution: nobody discovers a trust tool at the moment they need it. The version of this that compounds is publishing your own public scorecards — score the 50 most-searched scams and put the results where Google can find them. That turns your methodology from something users take on faith into something they can inspect before they ever install.

that's a great rejection to get because she told you the exact thing to fix. we had a smaller version with our first testers, where nobody complained and nobody came back, and silence is so much harder to act on than a specific objection like hers. clickable sources feels like it should be the default for any tool whose whole pitch is trust. did she say she'd take another look now that it's fixed?

Link, since I kept it out of the post: https://bigideasdb.com. The TrustMRR section is where this came from. If you want to try the chat nobody uses, ask it for the revenue split in your own category and see whether the X pattern holds there. I'll run any cut someone asks for in the comments.

This is great work — reminds me of some of the calls I've had to make building Xstream4K. What would you do differently if you started over?

This is a great example of a rejection being the most useful feedback. "A tool that asks you not to trust blindly shouldn't ask to be trusted blindly" is a sharp insight. Making the method public is exactly what would get a skeptical creator to look twice. Did you go back to her after the changes?

The creator who turned you down is now the warmest lead you have, since she named the exact objection and you fixed exactly that — going back to her beats every cold outreach still on your list.

The creator's objection is the same one that separates useful AI tools from demos: show your work or get out. We built an SEO audit tool that scans 100+ ranking factors and the hardest design decision was identical — do you show a score and ask trust, or show the evidence trail and let people verify?

Iteration speed matters here. Recent model upgrades mean rebuilding a page component takes one pass instead of three or four debugging rounds because the model reads the actual error trace now. When a rejection names the exact fix, you can ship it in hours not weeks. That compresses the feedback loop enough that a single honest no becomes more valuable than ten polite yeses.

Revenue at $0 with a working product and a clear iteration signal is better than $0 with neither. The trust layer you just added is the foundation everything else stands on.

Have prospects responded differently since adding source-level verification, or are they still stopping at the same trust objection despite being able to verify the claims?