If you’re wondering how to use Google Ads effectively to drive early traction for a new product, you’re not alone — we were in the same boat after launching CitySage. This post is a reflection on our first attempts at growth, what failed, what we learned, and how a small Google Ads campaign helped us uncover new insights.
The Frustration of Zero Traction
After building, designing, and iterating, we proudly launched the MVP of CitySage — a clean, intuitive, and genuinely useful tool for planning travel routes with the help of AI. We were excited. Everything felt ready. But once the product was live, silence. The first thing we tried was a Product Hunt launch. We followed the typical playbook: timed our post, shared with friends, even crafted what we thought was a compelling pitch. The result? Two signups. Not exactly the spark we hoped would ignite early momentum. Still hopeful, we shifted gears to social media. We committed to daily content across Instagram and Facebook — feature highlights, product shots, and even aspirational travel imagery. Days went by. Likes were sparse. Comments nonexistent. Traffic barely moved. Next, we focused on travel influencers. We handpicked profiles, personalized our outreach, and sent dozens of DMs. Most were left on read, others ignored entirely. At that point, the reality began to settle in: even with a product we believed in, finding those first users was proving to be brutally hard. We weren’t just lacking traffic — we were lacking signal. No feedback loops, no user behavior to learn from, just empty sessions and flat numbers. We knew we had to try a more structured and measurable approach — one that didn’t rely on luck or followers. That’s when we decided to give Google Ads a try.
What we learned
- Product Hunt is not plug-and-play — you need a warm community ready to engage on launch day.
- Social media without a base audience = shouting into the void.
- Influencer outreach needs trust, credibility, and something valuable to offer.
- Recommendation: Start manually - pick one platform and build a following by commenting, engaging, and following others consistently. Don’t post for the algorithm until you’ve earned attention.
Enter Google Ads: Learning by Doing
We didn’t expect Google Ads to be easy, but we underestimated just how complex the setup would be. Compared to social media or Product Hunt, this was a different beast entirely. We chose to run a “Maximize Conversions” campaign — mostly because it sounded promising, and we’d heard it was a good starting point for small teams without a lot of keyword experience. What we didn’t fully grasp at the time was how different it is from a traditional Search Term campaign. In Max Conversion mode, Google essentially takes over keyword optimization using its algorithm, requiring you only to input general themes or related search terms. The upside? Less manual work. The downside? You’re flying somewhat blind. To make sense of performance, we set up UTM tracking, connected LogRocket to view session replays, and monitored user flows daily. What surprised us early on was the automatically generated creative — poor-quality image assets and a strange auto-generated YouTube video that did our brand zero favors. Midway through the campaign, we scrambled to upload our own images and messaging to replace the defaults. Despite these hiccups, the results weren’t terrible. We ran the campaign for two weeks with a 400 CHF budget, generated 1000's of Impressions and clicks and acquiring a handful of users at an average of ~66 CHF per signup. Click-through rates hovered around 1%, and the cost-per-click remained reasonable. However, not everything added up. Through LogRocket, we discovered a large volume of ultra-short visits — sessions under one second — suggesting that scraping bots were likely eating into our budget. The campaign was a rollercoaster of insights, confusion, and cautious optimism.
What we learned:
- Maximize Conversions works best when you actively shape the creative and input keywords with intent.
- Google auto-generates terrible assets if you don’t supply them.
- UTM + LogRocket was essential to distinguish real users from bots.
- Scraping bots are real and expensive — especially for small budgets.
Recommendation on asset configuration:
- Prepare 3–4 strong headlines, 2–3 short descriptions, and at least 5 custom images or visuals.
- Use branded screenshots, clear call-to-action overlays, and a simple logo watermark.
- Don’t rely on Google to “figure it out” — control your message.
What We Learned and What We’d Do Differently
Despite the mixed results, running our first real Google Ads campaign gave us something we hadn’t had before: usable data. By watching session recordings in LogRocket, we could see how real (and occasionally fake) users interacted with our landing page. We noticed moments of hesitation in the UI, unclear value propositions, and weak calls to action — things we were blind to before. It also became clear that preparation is everything. Next time, we’ll launch with fully custom visuals and videos that better represent the product from the start. We’ll likely experiment with both Maximize Conversion and Search Term campaigns in parallel to regain some control over targeting and test specific hypotheses. On the technical side, we now know the importance of setting up bot filters and possibly using click fraud prevention tools to protect budget. We’d also improve our conversion tracking — not just signups, but smaller signals like time on page, scroll depth, or whether someone interacted with the AI. The biggest takeaway? Google Ads isn’t a silver bullet — but it’s one of the few tools that offers detailed, rapid feedback when you’re in the dark. For anyone else in the early stages: treat it like an experiment, not a growth engine. It’s not about perfection. It’s about progress.
What we’d do differently next time:
- Launch with polished image and video assets upfront.
- Test a Search Term campaign in parallel to Max Conversion for better targeting control.
- Use bot filtering tools or IP exclusions (something we will try next time).
- Track not just signups, but scroll depth, button clicks, and form field engagement.
Mini Pre-Launch Checklist:
- Create all assets (images, videos, headlines, CTA text) in advance
- Set up UTM links for campaign-level tracking
- Connect user behavior tools (e.g. LogRocket, Hotjar)
- Enable conversion tracking for both macro and micro events
- Use negative keywords or exclusions to filter junk traffic
- Plan 2–3 campaign types for A/B testing different formats
- Define your acceptable Cost Per Acquisition and track it live
Running Google Ads for CitySage didn’t unlock overnight success, but it gave us something far more valuable at this stage: real feedback, real behavior, and a direction forward. We learned what not to do, what to double down on, and how to approach digital campaigns with sharper eyes. For any founders navigating their first paid growth experiment, our advice is simple: prepare well, track everything, and treat every euro spent as a lesson bought.
Curious what we’ve been building behind the scenes? Try out CitySage today and plan your next trip with the help of AI at www.citysage.ch.