Week 2 - Launching Our First Google Ads Campaign: $100 Budget, Real Expectations, and Hard Lessons

Why Paid Traffic Was the Only Realistic Start

Waiting for organic traffic to appear on its own rarely works for a brand-new tool — especially when you're competing in a space where Grammarly already owns mindshare.

To get real data fast, I decided to run a small, controlled Google Ads campaign right after launch.

Budget: CHF 100 over roughly one week.

Goal: Not sales (way too early for that), but honest answers to basic questions:

• Do people click through to the tool?
• Do they actually use it?
• Where do they drop off?

Setting Realistic Expectations

I kept expectations grounded:

• Zero expectation of immediate purchases — that would have been unrealistic.
• Hopeful benchmark: most visitors at least click the "Try it out" button on the landing page.
• Real target: some users actually engage with the grammar tool itself (paste text, see suggestions, maybe try a few sentences).

Even better (but not required): a handful of people creating free accounts.

Anything beyond that would be a bonus.

Tracking Everything with Google Tags

To measure properly I added Google Analytics events and conversion tags at every key step:

• Landing page views
• "Try it out" button clicks
• Arrival on the tool page (grammar-mentor.com)
• First text paste / correction attempt
• Account creation (if it happened)

This setup let me see the full funnel clearly. If nobody clicked "Try it out," the problem was likely the landing page copy, ad-to-page mismatch, or unmet expectations — not the tool itself.

If people reached the tool but bounced quickly, that pointed to UX friction we could fix fast.

Data > guessing.

The Biggest Unexpected Time Sink: Ad Perfectionism

I spent 3–4 hours obsessing over ad copy.

Endless variations. Emotional hooks vs. straightforward benefit statements. Testing what might trigger curiosity, frustration relief, or trust.

Conversations with AI to refine headlines, descriptions, extensions.

Then it hit me: without any baseline data, perfectionism is just procrastination in disguise.

You can't A/B test meaningfully with zero impressions and zero clicks yet. The only way to learn which copy actually works is to run the ads — not polish them forever.

So I shipped "good enough" versions and launched.

Campaign Guardrails That Kept Costs Low

I focused exclusively on the US market (highest search volume for "grammar checker" terms).

Daily budget cap: CHF 15

Max CPC bid: CHF 1.50

These are tiny numbers by Google Ads standards, but they let me gather meaningful data without burning through money.

Google's own AI recommendations (and my repeated double-checks) confirmed: start conservative, watch performance, adjust later.

What Really Matters at This Stage

The campaign wasn't about ROI yet — it was about learning.

• Which keywords convert to tool usage?
• What's the click-through rate from ad to landing page?
• What's the drop-off rate from landing page to tool?
• Do any visitors create accounts?

Every number tells a story. The faster we collect those numbers, the faster we can improve the funnel — whether that's rewriting headlines, simplifying onboarding, or tweaking the tool experience.

Perfection can wait. Momentum cannot.

Early Days, Honest Iteration

We're still in week one of real traffic.

The goal is simple: get people using the tool, gather feedback, fix what hurts, and repeat.

If you've ever felt stuck with writing — grammar doubts, tone worries, unclear sentences — and you're tired of tools that just auto-fix without explaining anything, that's exactly why we built AI Grammar Mentor.

No complicated setup. No long sales pitch.

Just paste your text, get clear suggestions that actually teach you something, and keep your own voice.

Give it a try today — I'd love to hear what you think.