I targeted the wrong keywords for six months

Six months. That is how long I optimized my meal planner for keywords it could never rank for. Not six months of laziness: keyword spreadsheets, competitor lists, a rewritten title every few weeks. My downloads stayed flat the whole time, and for most of it I blamed the app.

My reasoning was wrong at the root. I treated keywords as a shopping list where the biggest number wins. The biggest numbers in my category were "recipe app", "meal planner", and "cooking app". When people typed those words, the store showed the same five apps that had owned the category for years, while my listing sat around position 300, a polite way of saying invisible. I did not even know how invisible: I checked results on my own phone, where the store already knew my app. Run it from a fresh device and you get the honest picture.

Then came the mistake that felt clever. A cheap tool listed my biggest competitor's keywords, so I copied them. Nothing moved. What I could not copy was why those words worked for them: years of installs, a conversion rate their listing has earned, and the store's record that when people search this, they install that. Ranking is not text matching, it is a feedback loop, and I was trying to enter it from the outside with words alone. A new app at thirty downloads a week does not outrank an app the store has trusted for a decade, whatever sits in its keyword field.

Everything changed when I swapped the question. Instead of asking which keyword has the most traffic, I asked where my app could be the best answer. That gave me three filters. The phrase had to describe a job, not a category: "meal planner" is a category, "meal planner for two people" is a job, and my app was built for that job. It had to be long enough that big apps do not bother, because nobody fighting for "recipe app" cares about "dinner ideas with what you already have at home". And the results had to look beatable: old screenshots, keywords stuffed into titles, three star averages. I collected twelve candidates and worked them into the title, subtitle and keyword field, and waited.

Twenty days later I was in the top ten for two of them. Rank ten for a phrase five hundred people a month type earns more than rank three hundred for one two hundred thousand type. Downloads went from about twenty-five a week to two hundred and fifty over the next ten weeks, and the curve kept bending up, because installs made the following weeks easier. Paid installs stop the day the money stops, as I had learned the expensive way on Meta and TikTok. The organic loop was the first thing that compounded.

The other half of the lesson hid where I refused to look: languages. I had done all of that work in English, in the US store, on a planet where most downloads do not happen in English. When I ran the same exercise in other markets, the pattern repeated. In German, "Essensplaner" had real searches and weak competition. In Brazilian Portuguese, "planejador de refeições" had almost none of the apps I was afraid of. A long tail exists in every language, just shorter, because most developers never look. Some languages bring volume, like Brazil, Indonesia, or India. Others bring revenue, like Germany, Japan, or the Nordics. An English-only listing shows up for neither. The same logic applies to the niche: a cricket app has to speak the languages of the places where cricket is a religion. Every niche has a map and a vocabulary, and your listing has to live on that map.

I got plenty wrong on the way there. My first long tail list was absurd: "meal planner for couples who work late and want to lose weight" is eleven words, and nobody types eleven words. For a while I stuffed "meal planner" into my subtitle twice, until a friend said it read like spam, and he was right: the extra impressions never converted. And I ignored every non-English market for months, assuming that people who can read English search in English. They do not.

The lesson is that keyword choice is not a volume auction, it is a winnability test. You win where your app is the most relevant answer and the field is thin, then stack small wins until the bigger ones come within reach. My six months were not wasted, but two of them would have been enough with better targeting. If your downloads are flat, the fix is rarely a bigger budget. It starts with the question of where your app can actually win.

That mechanical work, the per language keywords, the title and subtitle rewrites, the translations, the screenshots that have to match each market, is exactly what AsoFast automates. It generates your App Store and Google Play metadata with AI, translates listings into more than 80 languages, provides screenshot templates, and publishes directly to App Store Connect and Google Play. Everything runs locally, so your credentials never leave your computer. It is open source under the MIT license and free, which matters when the alternatives start around thirty dollars a month and climb toward eighty. Start with the question of where your app can win. AsoFast handles the rest.