Built an AI pet costume generator — getting traffic but 0 paying users. What am I missing?
Hey IH,
I launched PawDressed about a week ago. It's an AI photo generator that dresses your pet in different costumes — knight, astronaut, pirate, 50+ styles — while keeping their actual face (most competitors morph pets into human-like figures, which felt wrong to me).
Where I'm at:
Traffic: coming in from TikTok and Instagram content
Signups: a handful
Paying users: 0
The funnel: upload a pet photo → pick a style → get a free watermarked result → pay to download HD without watermark. Pricing starts at $4.90 for one image.
What I've tried:
Free first image to lower friction
Watermark that shows the site URL (so people who share it do free marketing)
50 styles to give people reasons to come back
What I think might be happening (but not sure):
People like the free result, share it, but never intend to pay
$4.90 is either too high or the value isn't obvious at the paywall moment
The audience coming from short-video content is entertainment-only, not buyer-intent
My questions for you:
If you've seen a "free result → paid upgrade" flow work, what made users cross the line?
Is $4.90 per image reasonable, or should I be testing a different price/bundle?
Would you personally pay for a HD version of your pet in a knight costume? (honest answer please)
Happy to share more context. The watermarked results look genuinely good — that's not the problem I think. Something is broken between "this is cool" and "I'll pay."
Traffic with zero pay usually means wrong visitors or the page sells novelty not outcome.
What does "paid" look like for you — one-off render, subscription, credits? And where's the traffic from (TikTok, SEO, PH)? Very different fixes depending on answer.
Could not reply directly because of the thread limit, but one important point on the premium-name testing:
I would not test premium names only as creative options. I’d test only names where the .com can actually be secured.
Otherwise the Father’s Day data may validate the premium portrait/gifting direction, but the strongest name option may still be uncontrolled when you need to move fast.
Auryxa.com is one I control, and it fits the shift from novelty pet costumes to a polished pet portrait/gifting brand much better than PawDressed if the bundle test proves people are buying emotional keepsakes.
I also have a couple of alternative controlled .com directions depending on whether the winning buyer motivation is more premium, lifestyle-led, or emotional keepsake. But I would not scatter those publicly before knowing which direction the test is actually pointing toward.
So I’d treat Auryxa as the main controlled option, not just naming inspiration.
If you are seriously preparing premium alternatives now, it is worth discussing the controlled .com options before you build logo concepts or landing page tests around names you may not be able to own. I can keep the acquisition side simple and founder-friendly if the fit is real.
LinkedIn: https://www.linkedin.com/in/aryan-y-0163b0278/
Getting traffic is a good sign maybe the issue is converting curiosity into real value users will pay for. Have you tested a more niche audience or gifting use case?
Yes — gifting is actually the angle I'm betting on right now. Switched to bundle-only pricing last week with gift/keepsake framing, and Father's Day is the 4-week test window. On the niche audience side: so far traffic is mostly pet owners finding it via social shares, but I haven't deliberately targeted gift-buyers yet. That might be the next lever — if the bundle data shows people are buying as gifts, I should be reaching people who are actively looking for pet-themed gifts, not just people who stumbled on a cute dog photo.
The problem may not be the AI result quality. It sounds like the product is creating a fun “share moment,” but the paywall is asking for a utility purchase. That gap matters because people will happily share a funny pet costume, but paying for one HD download needs a stronger reason: gift, keepsake, printable poster, profile picture pack, birthday card, holiday bundle, or something with emotional value beyond “remove watermark.”
I’d probably test bundles before lowering the price. One image for $4.90 feels like a small impulse purchase, but it may also feel arbitrary. A “pet portrait pack” or “holiday costume set” could make the value clearer and give people a reason to pay instead of just enjoy the free version.
One thing I’d watch is the PawDressed name. It explains the use case, but it may keep the product feeling like a novelty costume tool. If this grows into a more premium AI pet portrait/gifting brand, Auryxa.com would carry the polished consumer angle better than a literal pet-costume name.
Hey, thanks for the sharp breakdown — this is genuinely useful.
The share moment vs. utility purchase gap is exactly what I suspect is happening. People upload, laugh, share the watermarked result, and the emotional peak is already gone by the time they hit the paywall. "Remove watermark" is too transactional a reason to pay $4.90.
Acting on your bundle suggestion right now — removing the single $4.90 option entirely and making the 3-pack the entry point, with copy that frames it as a printable keepsake / gift rather than an HD upgrade. Father's Day is in 4 weeks, so testing the gift angle has a real deadline.
On the name: I'm keeping PawDressed for now — the literal name actually helps with SEO and word-of-mouth ("I used this site called PawDressed"), but fair point that the brand positioning could lean more into portraits and less into costumes as it grows.
Will report back in 2 weeks on whether reframing the paywall changes conversion at all.
That makes sense. For the current wedge, PawDressed is probably helping because it is literal, easy to understand, and tied to the first use case.
The test I’d watch closely is whether the gift/keepsake framing changes who is buying.
If people start paying because it feels like a printable portrait, Father’s Day gift, profile set, or emotional keepsake, then the brand may naturally start moving beyond “pet costume generator.” At that point, the name question becomes less about SEO and more about whether the product is being perceived as novelty entertainment or a premium pet portrait/gifting brand.
So I would not force a rename now. I’d let the paywall test tell you what category customers are actually buying into.
If the 3-pack starts converting better, that is probably the signal to lean harder into the portrait/gift positioning instead of just the costume angle.
Agreed — not touching the name for now. The paywall is live with the gift/keepsake framing and bundle-only pricing as of today. Father's Day is 4 weeks out which is a clean natural test: if "perfect gift for pet lovers" messaging converts at a meaningfully higher rate than before, that's the market telling me what category this actually belongs in. Will report back with numbers.
That is the right test.
If the gift/keepsake framing wins, I would treat that as more than a paywall lesson. It would mean people are not just buying “funny pet costume images.” They are buying emotional value: something printable, giftable, memorable, and worth keeping.
That is the point where the naming question becomes serious again.
PawDressed is useful for the current wedge because it explains the costume use case quickly. But if the product starts converting as a premium pet portrait or gifting brand, the name may start anchoring it too low, more novelty than keepsake.
So I’d watch two things:
Are people buying more because of the bundle?
And are they buying because it feels like a gift/portrait rather than just an HD version of a funny image?
If both are true, I would not wait too long to revisit the brand layer. Once Father’s Day pages, paid users, examples, and word-of-mouth build around PawDressed, the category gets harder to move later.
Both metrics are now tracked. Paywall went live with bundle-only + gift framing yesterday. I'm watching the 3-pack conversion rate as the primary signal — if it moves, something shifted. The harder signal to read will be purchase motivation: I'll try to get that from post-purchase email replies ("what made you decide to buy today?"). Father's Day deadline in 4 weeks creates natural urgency, but also means I need to strip out the seasonal effect to see the baseline. Will post numbers either way.
That makes sense.
At this point, I’d let the test decide the category.
If buyers are mainly paying because it feels like a gift, keepsake, or premium pet portrait, then this becomes a brand decision, not just a paywall decision.
That is where PawDressed may start limiting the product, and a cleaner premium consumer name like Auryxa becomes worth considering seriously.
If the test does not show that shift, I would keep PawDressed and not overthink it.
So the real question is simple: if the gift/portrait framing wins, would you actually be open to locking in a more premium brand direction, or are you committed to PawDressed either way?
The name question is interesting timing — I've thought about it but I'm holding off until the positioning signal is clearer. "PawDressed" does anchor to costumes, which might work against the premium portrait angle. But renaming before I've validated the gift use case would be solving the wrong problem.
If the bundle + Father's Day window shows people are buying as gifts (not just downloading for fun), that's when it becomes worth paying the switching cost of a rebrand. Right now I'd rather keep the domain authority and SEO momentum while I test the positioning. The name can follow the product-market fit, not the other way around.
Will update in a few weeks with what the data shows.
That is the right way to think about it. I would not rename before the gift/portrait signal is real either.
The only thing I would separate is “do not rebrand yet” from “do not plan the brand path yet.”
Those are different decisions.
Keeping PawDressed while you test makes sense because it has SEO clarity and explains the costume wedge. But if the Father’s Day window proves people are buying this as a gift, keepsake, or premium portrait, then the brand question becomes time-sensitive very quickly.
At that point, every paid example, landing page, email, share, and customer memory starts reinforcing either the novelty costume frame or the premium portrait/gifting frame.
So I would not force the switch now, but I also would not leave the premium brand direction completely open until weeks later.
Auryxa is the kind of name I’d pressure-test quietly in parallel because it gives you room to move into a polished consumer/gifting brand without dragging the costume-frame with it.
If the test wins, you probably want the brand option already understood, not starting from zero after the signal is obvious.
Fair distinction. "Don't rebrand yet" shouldn't mean "don't think about it." I'll start quietly testing a few premium alternatives — domain availability, logo concepts, maybe a landing page — so if the conversion data comes back strong, I'm not starting from scratch on branding. "Auryxa" is interesting as a direction; I'll explore that angle alongside 2–3 others. The goal is to have options ready, not to be reactive in 4 weeks when I'm also reading Father's Day numbers.