I vibe-coded a tool to see my website through my visitors' eyes

A few weeks ago I asked myself a simple question: is my own site actually any good?

Not "is the code clean" — I can measure that. But: does a tradesperson landing from Google immediately understand what I offer? Do they trust it? Do they take action?

I couldn't ask contacts to "look at my site" every time I made a change. So I did what I probably should have done from the start: I built a tool.

1. Why I built this tool

I'm a developer. When I have a recurring problem, my natural reflex is to automate it. This project — which I call Echo — is a good example of vibe coding: I had a clear need, I prototyped fast, I iterated.

The starting idea was simple. I wanted to simulate "realistic" visitors on my site — not generic bots that scan HTML, but personas with a profile, an arrival channel, a device, a decision maturity level. Visitors with a story who react the way a real potential client would.

The tool generates these personas, simulates their journeys, and produces a structured report: who converts, who hesitates, who leaves — and above all, why.

Once I'd built and tested the tool on my own sites, I decided to offer it to others. Not because it's a finished, polished product — it's still handcrafted, clearly — but because the insights I get from it are useful, and I think others could benefit too.

2. How Echo works

You give me your site URL. The tool generates a batch of personas and simulates their visits. Each persona has a detailed profile:

The output is a structured report: overall conversion rate, breakdown by channel and decision maturity, scores for experience / trust / clarity, list of identified friction points, and the 3 priority actions to improve.

The report is a dedicated web page, readable on mobile, that you can share. Not a 40-page PDF no one reads to the end.

3. What I discovered about marcm.fr

I ran the Pro version (100 visitors) on my own site. Here's what came back.

Echo Pro report for marcm.fr — view the full report

58 out of 100 visitors get in touch — which is pretty solid. But the real value was understanding why the other 42 leave or hesitate.

Not all acquisition channels are equal

The social media result at 25% stood out immediately. People arriving from Instagram or LinkedIn come without context, without clear intent — and the site doesn't warm them up fast enough. That's a friction point I wouldn't have identified on my own.

The main friction: a portfolio that isn't sector-diverse enough

35% of visitors weren't sure they were the right audience. The portfolio covers tradespeople and local businesses well, but not freelance professionals, healthcare practitioners, or service providers who need booking systems. These visitors hesitate or leave because they don't see themselves reflected in the work shown.

That's not a criticism I would have guessed intuitively. It's exactly the kind of insight that only surfaces when you look at your site from the outside.

Decent scores, but not excellent. Room to improve — and now I know exactly where to act.

4. Available plans

5. Try it on your site

Echo is still young. I consider it a well-functioning beta — useful, but I'm still learning what makes it most valuable depending on the type of site.

What I'm looking for now is concrete feedback: do the friction points identified match your gut feeling? Are the personas realistic for your sector? Is the report readable and actionable?

If you have a site — portfolio, shop, landing page, service site — and you want a structured external perspective, the Discovery plan is free. No commitment, results within 48 hours.