We Tried Google Ads for an Outdoor App. It Didn't Work.

This is the post we kept putting off writing. It’s easier to publish a feature announcement than to publish an experiment that didn’t pan out. But the lesson is worth more than the spend was, so here it is — a modest paid-ads experiment over a few months, what we tried, what we got, and why we turned it off before we’d spent everything we’d set aside.

Most marketing posts in our category are written by the people who won. This one is written about the thing that didn’t work for us, and we think that’s the more useful version. Indie outdoor builders are reading the same playbooks, hearing the same case studies, and quietly burning money in the same auctions. So here is the unvarnished version, including the parts that make us look slow.

What we tried and what we got

Over a few months we ran a small Google Ads experiment across search and a little display retargeting. The campaign produced some installs, a smaller number of trial starts, and a much smaller number of paying subscribers. The cost to land each paying subscriber settled well above what a $34.99 yearly subscription — or $7.99 a month — can justify, and that’s before the platform takes its cut. The arithmetic was unkind early on, and we kept the experiment running because we wanted it to work and kept hoping the next optimization pass would turn the corner. It didn’t.

The keyword traps we walked into

Generic outdoor keywords look like a goldmine on paper. “Fishing app,” “hunting GPS,” “topo maps,” “offline trail maps” — high volume, clear intent, recognizable to anyone shopping the category. In practice those terms are auctioned against incumbents with enormous war chests and a decade of conversion data. We were paying premium CPCs to compete for clicks from people who were going to install three apps and keep the one with the most reviews. Volume is not fit, and the auction was never going to be fair on a budget our size.

Why the clicks didn’t qualify

The clicks we bought were curious, not committed. Someone Googling “best fishing app” at 9pm on a Tuesday is in a different headspace than someone who just got a recommendation from a buddy at the boat ramp. Paid search caught browsers; it almost never caught buyers. Our highest-intent organic visitors arrived after reading a comment, a write-up, or a specific feature thread, already half-sold. Ad clicks arrived cold, scanned the landing page for a few seconds, and bounced before they ever saw what made us different.

The install-to-trial drop

We expected the funnel to leak; we didn’t expect it to leak this badly. A large share of paid installs never got past the first screen, and only a fraction of the rest started a trial. Organic installs converted to trial at a noticeably higher rate. The diagnosis was simple: paid users had no prior context for what we were, so the first thirty seconds had to do all of the persuasion work, and thirty seconds is not enough time to explain a niche outdoor tool to someone who showed up by accident.

The trial-to-paid drop

Trial-to-paid was the most disappointing part to look at. Paid-acquired trialists converted to subscribers at a meaningfully lower rate than organic trialists. The pattern was consistent: people who came in cold treated the trial like a free utility, used it for a single trip, and let it expire without a thought. People who came in warm — through a recommendation or a write-up — were already pre-sold on the category and used the trial to confirm a decision they’d already made before downloading.

Comparing the cohort to organic users

When we lined the two cohorts up side by side, the gap was not subtle. Organic users opened the app more often, saved more waypoints, returned after the trial ended at a higher rate, and were far more likely to leave a review or tell a friend. Paid users churned earlier, used fewer features, and rarely told anyone about us. The acquisition-cost math was bad on its own. The retention math made it worse, because every paid subscriber who lapsed was also one who never referred anyone. Whatever paid acquisition was selecting for, it was not the kind of user who stays through a season.

What we killed and when

We killed display retargeting first — it was the worst performer and the easiest call. Generic search keywords went next, in waves, as we accepted that no amount of negative-keyword pruning, ad-copy rewriting, or landing-page testing was going to fix the underlying intent problem. We kept brand-defensive bidding running the longest, then realized we were paying for clicks from people who had already decided to find us, and shut that off too. By the end of the experiment the account was paused. The relief was immediate, and the next week’s numbers barely moved.

What worked instead

After we turned ads off, one honest post in a regional fishing thread drove more qualified installs than the entire paid campaign had. A YouTube creator we’d never heard of mentioned us in a brief aside and we watched the trial conversions roll in for weeks afterward. The in-app feedback box, the Development Queue, and a handful of long DMs with power users produced more roadmap clarity than any landing-page test ever did. Word-of-mouth consistently converted far better than the paid clicks, and it cost us nothing but attention and follow-through. The pattern held across every channel that worked: a real person, talking to other real people, about a thing we’d shipped.

What we’d test if we ran paid again

If we ever go back, it won’t be with generic category keywords and a “best outdoor app” headline. It will be a small, specific test: one shipped feature, one regulation update, one regional cohort that we already know converts, told in a real person’s words on a landing page that matches. We’d cap the spend tightly, measure trial-to-paid rather than install volume, and shut it off the moment cohort retention dipped below our organic baseline. Probably we won’t run it again at all. The community channel is doing the work.

---

Paid acquisition is not universally broken. For products with broad, well-understood intent and a clean conversion path, it can be the right tool. For a niche outdoor app sold to people who care intensely about a specific region, a specific watershed, a specific season, the right tool is almost always another human telling them we exist. We learned that the slow way. Maybe this post saves someone else the lesson.

We also under-estimated how much of our early growth was already organic before we started paying for traffic. When we layered ad spend on top of word-of-mouth, the dashboard looked like the ads were working. They weren’t. They were riding on a baseline of organic interest we’d built by shipping features and answering DMs. When we shut the ads off, the organic line kept climbing on its own. That, more than any cost-per-acquisition calculation, told us the truth.

The thing we got right, almost by accident, was keeping the trial honest and the pricing simple. One feature set for everyone, fourteen days fully unlocked, no dark patterns, no upsells. That meant the funnel math was readable. If we’d been running a multi-tier price ladder we’d still be arguing about which step was broken. As it stands, the funnel was clear, and so was the verdict.

A niche outdoor app does not need a million users. It needs a few thousand of the right ones. Paid acquisition optimizes for volume by default, and “the right ones” is a constraint the auction does not know how to honor. Every dollar we spent nudged us toward the wrong shape of user base. Turning it off was a clarity decision as much as a cost decision.

The Development Queue is open — every shipped feature is the result of someone using the app and telling us what was missing. That, not Google Ads, is how Baseline Maps grew.