I built an AI to kill the 3-week Excel research grind. It got 8 PH upvotes. What did I do wrong?

Great perspective on the UX/design flow! Your point about spending a week on pixel-perfect screens only to get crickets resonates with many indie hackers. The emphasis on validating demand before perfecting the design is solid advice. Did you consider doing any pre-launch validation or user research to gauge interest in the concept before the full build?

Pixel-perfect screens + crickets" – felt that in my bones. Been there.

Bunzee solves the competitor + PRD grind. I solve a different part of the funnel: whether the idea has demand before the competitor deep dive.

I built TrendyRevenue – AI that checks search demand, revenue potential, and competitor saturation in 10 seconds. It's the "go/no-go" filter before I spend hours on PRDs.

Here's how I'd use your tool with mine:

Run idea through TrendyRevenue free tier (one analysis, no card). If demand exists and gaps are clear → green light.
Then feed those insights into Bunzee for the full competitor matrix + PRD.
That combo kills the 50-tab Excel grind and the "building a zero-demand product" risk.

If you're validating Bunzee's next feature or a new idea, try the free tier at https://trendyrevenue.com. See if the demand signal matches your gut. If it does, consider Pro ($39/mo) for deeper revenue modeling + source-cited gaps.

Appreciate you sharing the 8 upvotes story – honest launches help everyone.

Really cool tool , and honestly, this is exactly the kind of feature I was thinking of adding to my own startup's Growth Plan later. So I have a few genuine questions (not competing, just learning):

What does your pricing look like? Monthly subscription? One-time? Range?

How did you build the competitor analysis engine? Did you use an existing LLM API (GPT, Claude, etc.) or build something custom?

How many API calls does a typical analysis take? Trying to understand cost per user.

If someone wanted to integrate something similar into their own platform — would you recommend building from scratch or using an existing stack?

ran my own version of this post a week ago. small numbers, same
question, same instinct to call it a distribution problem.

the reframe that actually helped me: small upvote counts don't
measure pain-fit, they measure how many people in your warm network
are willing to do you a small favor. distribution is downstream of
that. the 8 who engaged probably weren't your ICP — most were
friends saying yes because they like you, not because they hurt.
the real signal isn't 8; it's how many of those 8 opened the
product, used it past the first session, came back. that gap is
the read.

two specific things worth sitting with for bunzee:

1. what % of the 8 upvotes actually used the tool? PH measures
   "i'll click upvote for a friendly stranger," not "i need this." the
   conversion from upvote → real use is the metric you care about.

2. PRD-as-output might be hidden friction. your target user is the
   "overwhelmed maker" — but a generated PRD is also kind of
   overwhelming. they wanted to skip the research, not get more
   documents. is there a version where the output is "build these 3
   features in this order, ignore the rest"? smaller surface, lower
   cognitive load, more trust.

the "using bunzee on bunzee" instinct is great but watch the loop —
if the product is mis-calibrated for overwhelmed makers, it'll
suggest positioning that's also mis-calibrated. external eyes (this
thread is doing it) are the corrective.

8 PH upvotes with a beautiful UI is the same lesson I learned. The product isn't the problem. The launch audience is.

PH audiences are mostly other founders and developers. Your actual users are the people drowning in that 50-tab Excel grind right now. Those people aren't browsing Product Hunt on launch day. They're in Slack groups complaining about competitor research, in agency Discord channels asking "does anyone have a template for this," or posting on LinkedIn about how long client research takes.

The fact that you're now using Bunzee to analyze your own competitors and pivot your messaging is actually the most convincing proof the product works. That's the story you should be telling. Not "here's our 4-step workflow" but "we launched, got 8 upvotes, realized our messaging was wrong, ran our own tool on ourselves, and here's what it found." That's a post that gets shared because it's honest and useful, not because it's a product announcement.

I'm in a similar spot. 95K lines of code, 39 tools, $0 MRR. My PH launch is deliberately delayed because I watched enough founders launch to an empty room and question the product instead of the strategy. Instead I'm spending months doing daily engagement in the communities where my users already hang out. No single day moves the needle but the compounding is starting to show. Whether it converts to revenue is still TBD but at least I'm not launching into silence.

One specific suggestion: your "Who Is Bunzee For" section lists solo founders, designers, developers, and agencies. That's everyone. Pick one. The agency use case is probably your fastest path to revenue because they do this research repeatedly for clients and would pay monthly. Solo founders do it once and churn.

8 PH upvotes actually isn't bad for a solo launch - the problem might be that Product Hunt traffic rarely converts directly to paying users without a clear "why now" hook.

A few things that might help next time:

• A specific pain point story lands better than the feature description. "I wasted 15 hours on this Excel task" is more compelling than "kill the 3-week grind"
• The hunting matters. Getting hunted by someone with followers (vs self-submitting) 3-5x the initial traffic
• Day-of outreach to your network matters more than everything else combined

That said - shipping and learning is the whole point. What's your hypothesis about the core user?

The 8 upvotes on ProductHunt despite great UI hit hard — went through something similar.
One thing I've noticed: the "what to build" problem is real, but most founders (including me) skip it because building feels productive and researching doesn't. Your own story proves it better than any feature list.
Question — does Bunzee work well for very niche B2B tools? I'm building for a specific audience (solo developers who live in Telegram) and most competitor analysis tools surface the same obvious players. Curious if the 1-star review mining actually finds the non-obvious gaps.

8 PH upvotes on launch day usually means the distribution was the problem, not the product. The 'feature matrix' pain point is genuinely real — I've seen FinTech and SaaS founders spend weeks on exactly this. The question I'd ask: who were the first 8 upvoters? If they were mostly friends/connections rather than people who stumbled on it via PH's feed, that's a distribution signal, not a product signal. The IH post-mortem approach you're taking here is actually smarter — IH readers ARE your target user. What feedback have you gotten from people who actually used Bunzee to do competitor research?

I completely feel your pain about the PH launch! Getting only 8 upvotes after weeks of hard work is tough, but using your own tool to pivot is super smart.

I usually just want to jump straight into an editor and start building, so doing that massive Excel research phase is totally exhausting for me. Automating the PRD and competitor matrix sounds exactly like what I need. Keep pushing!

I totally agree. AI has made the barrier to building almost zero, but customer acquisition is harder than ever. I’m curious though—has your own tool actually helped you discover any specific product directions that are truly worth the investment?

PH is a launch channel, not a discovery one. 8 upvotes usually means the launch ran without a pre-built audience to mobilize on day 1. The thing that worked for us wasn't optimizing the PH page , it was running content for 3-4 weeks before launch on the channel where our audience already lived (LinkedIn in our case). Without that warm crowd, PH is mostly noise. What channel are your target users actually on?

The 8 upvotes thing is brutal but honestly it's a PH distribution problem, not a product problem.

I had the same experience — launched something, crickets. The shift that actually worked was stopping chasing the launch spike and building search presence instead. Google and ChatGPT don't care about your launch day, they care about whether your content answers real questions people are searching for.

On Bunzee specifically — the messaging "kills the Excel grind" is strong but I'd push it further. Who loses the most sleep over this? Solo founders pre-launch? Agency PMs presenting to clients? The sharper you get on that one person, the better your next launch will go.

Also the irony of using Bunzee to fix Bunzee's own positioning is genuinely great — that's a story worth telling publicly.

Bunzee solves a real founder pain, but the name is where the seriousness drops.

The workflow is useful.
The category is painful.
The buyer is real.

But “Bunzee” sounds playful in a part of the stack where buyers are trying to reduce product risk, not discover a cute tool.

You’re asking founders to trust it with market validation, positioning, competitor analysis, and PRD logic before they commit weeks of build time.

That layer has to feel sharper than the current name does.

This is much closer to decision infrastructure than a lightweight research assistant.

Xevoa.com would carry this better if you lean harder into strategic product intelligence instead of keeping it framed like a friendly ideation tool.