First Google Ads Results: What CHF 100 Taught Us About Early Traffic & User Behavior

One Week In: The Numbers Are In

After running our first Google Ads campaign for roughly 7–10 days (CHF 100 total budget), we finally have real user behavior data instead of guesses.

Quick context: the entire site was built in one week using basic HTML + JavaScript. Ads started in week two. No long runway, no big team — just a solo founder testing whether people would actually try a new grammar tool.

The Good News

• Strong initial interest: many people clicked the ad and then clicked the prominent "Try out" button on the landing page
• Real engagement: a solid number of visitors reached the tool page and actively used it — pasting text, seeing corrections, accepting some suggestions
• First free signup: one person created an account via Google login (completely unexpected at this stage — huge morale boost!)

Seeing that single account creation felt like crossing an invisible milestone. It proved someone trusted us enough to go beyond casual browsing.

The Friction Points We Can't Ignore

Not everything was smooth — and that's exactly why early testing matters.

Biggest issues spotted in the data:

• Extra click required: users must click "Try out" on the landing page before reaching the actual tool → adds friction and drop-off
• Login experience feels cheap and unpolished: quick implementation, but visually it doesn't inspire confidence
• High bounce after arrival: roughly 40–50% of ad traffic left immediately without any interaction (classic sign of expectation mismatch or trust gap)

These aren't fatal — they're fixable — but they clearly hurt conversion from visitor → active user.

The Big Technical Decision

That’s why we decided to switch entirely to Next.js—you can read about how we handled authentication without custom code here: Why I Skipped Building a Custom Login System (and Chose the Fast Path Instead).

The more I analyzed the data, the clearer it became: the current static HTML/JS setup was too limiting.

• No clean way to add a proper /blog section (critical for long-term SEO and content marketing)
• Scaling features (paid tiers, user accounts, better UX) would become painful fast

So I made the call: rebuild the entire site in Next.js. Yes, it means starting over in some ways — but it unlocks:

• Server-side rendering + static generation for better SEO
• Easy blog integration under the main domain
• Cleaner auth, faster performance, mobile-friendly foundation
• Future-proof architecture as we add more features

Short-term pain for long-term speed and growth. Classic founder trade-off.

Key Takeaways for Anyone Starting Google Ads

If you're about to run your first campaign on a new product, here's what actually worked (and what wasted time):

• Don't over-perfect the copy upfront: spend 3–4 hours max on headlines/descriptions. Ask multiple AIs for variations, pick what feels strongest, then launch
• Set conservative limits from day one: daily budget cap + max CPC (we used ~CHF 15/day and CHF 1.50 max CPC) — keeps spend predictable
• Launch even if it's "meh": you have zero data until impressions and clicks happen. Perfectionism before data is just procrastination
• Track the full funnel: button clicks, page views, tool interactions — not just ad CTR
• Expect iteration: the first campaign teaches you what to fix, not how to get rich

Data beats projection every time.

What's Next

We're now rebuilding with Next.js, polishing the login flow, removing unnecessary clicks, and preparing for round two of ads with better targeting and clearer value messaging.

The goal remains the same: get more people actively using the tool, collect honest feedback, and keep iterating toward something people love enough to return — and eventually pay for.

If you've ever felt frustrated by grammar tools that auto-correct without explanation, feel clunky on mobile, or force a generic tone on your writing — that's exactly the problem we're solving.

AI Grammar Mentor is still early, but it's built to teach you why changes matter so you improve for good — not just look polished for one moment.

Paste any text and try it yourself. No sign-up wall, no pressure.

Just clearer writing that still sounds like you.

I'd love to hear what you think.