App Store Optimization

Something published on the web can be found through many doors. Search engines index it, other sites link to it and send people through those links, and anyone can paste the address into a chat window and open it. A mobile app — an application installed on a phone — works differently. There is essentially one door, the app marketplace, and inside it the places where a person meets an app narrow down to the search box and a few recommendation lists. An app can be well known outside the marketplace and still get no installs if typing its name into the marketplace search box doesn't put it near the top.

Pagedeck's first month with a mobile app looked exactly like that. While Ravi was running paid ads the app picked up around 40 installs a day; when the budget ran out and the ads stopped, installs fell to 3 a day within three days. Searching "meeting notes" in the marketplace put the app somewhere around twentieth, and that phrase appeared nowhere in the app's name. Installs existed only while money was going in.

App Store Optimization (ASO) exists to widen that door with listing content instead of budget. The starting point is treating the name, the description, the screenshots, and the icon not as advertising copy but as the raw material search runs on and the screen that decides whether someone installs.

An install inside a marketplace passes two gates. Showing up in the search results list is the first; getting the person who tapped through to press the download button is the second. Different materials matter at each one.

The first gate turns on which words sit where. Words in the app name and in the one-line subtitle under it carry the most weight, followed by the description and a separate keyword field. The rules differ by marketplace. Apple gives you a keyword field that users never see and reads the words you put there; Google has no such field and pulls words out of the description text itself. The same app therefore gets different text written for each store.

What goes in those slots is decided by keyword research, the work of finding the phrases people actually type and choosing which ones to go after. What a team calls a collaborative document tool, people search for as "meeting notes app." The choice runs in two directions at once: a phrase searched 30,000 times a month is held by apps with years of installs behind them, while a phrase searched 500 times is usually typed by someone who already knows the problem and installs at a much higher rate.

The second gate turns on what a visitor sees. The icon, the first two or three screenshots, the star rating, and the content of recent reviews all land here. Someone scrolling a results list gives any single app a few seconds, so the fourth and fifth screenshots mostly go unseen. What goes in the first one drives conversion rate — the share of people arriving on a page who take the target action, here pressing download.

The two gates aren't independent. Marketplaces push apps that convert well relative to their impressions further up the list, so fixing the second gate lifts ranking at the first. Off-page SEO, the web technique of earning links from other sites as a reputation signal, drops out entirely: there are no such links inside a marketplace, and star ratings and review counts take over that role.

This work attaches at three moments.

1. When an app first goes up on a marketplace. Some of the name, subtitle, and screenshot choices need another review pass to change, so they get settled before launch. The name chosen here decides much of the search ranking six months later.
2. When paid acquisition gets expensive. People sent by ads still install through the marketplace listing, so lifting that page's conversion rate from 18 percent to 24 percent buys roughly a third more installs on the same budget. That's the reason to look at the destination before rewriting the ad.
3. When ratings move. A single update that drops the rating from 4.5 to 3.8 pulls ranking and conversion down together. Replying to reviews and writing what got fixed into the release notes is part of the job, because a visitor reading three one-star complaints about a bug that shipped two months ago has no way to tell it was fixed.

At Pagedeck, Ravi picks the phrases to target and Noa makes the icon and screenshots. This is also where the work shares material with on-page SEO on the web side. Page titles and app names get drawn from the same keyword list; one side unfolds into long prose and the other has to fit inside 30 characters.

"Let's put 'automatic meeting notes' in the subtitle after the app name. Eight thousand people a month search that phrase."

A proposal to move a target phrase into the slot that carries the most weight.

"Installs are flat but listing conversion went from 18 to 24 percent. All we changed was the screenshot order."

More installs out of the same impressions, earned without buying more traffic.

"That phrase isn't winnable through the listing — it's a paid slot."

A judgment that a competitor has held the term too long to displace organically.

Pagedeck reworks its marketplace listing once a quarter. Ravi brings the phrases that drove traffic last quarter along with installs per phrase, and Noa checks whether the top three of those phrases actually appear in the first screenshot. When what someone searched for and what the first screen shows don't line up, they leave right there.

The most common mistake is assuming more keywords in the app name is better. Once you learn that name text carries the most weight, it follows naturally. But the name shown in a results list gets cut off somewhere around 30 characters, and everything after that never reaches a human eye. A name like "Pagedeck - Meeting Notes Collaboration Docs Memo Work" may catch a few more queries while making the app harder to recognize, and marketplace review often rejects it.

Assuming web tactics transfer directly is nearly as common. The word "search" sits in both names, so the methods look interchangeable. In practice there is no link economy, so the entire off-site reputation half disappears, and writing longer text to climb the rankings doesn't work either. A handful of images decides installs instead. Teams that carry the web playbook over end up rewriting the description every few weeks while the icon and first screenshot stay exactly as they shipped a year ago.