Riley Chen makes Northpaw, a dog-walk tracker, from a spare room in Columbus. It records the route, the distance and, since last winter, which of the two household dogs pulled hardest. It has 9,400 paying subscribers and one employee.
For two years Riley kept a spreadsheet of keyword ranks. Every Sunday night: search the App Store for eleven phrases, scroll, count, type the number in. A second tab held release dates. Lining the two up was a job for a third tab that never quite worked.
The chart the spreadsheet was trying to be
Ascent tracks 50 keywords per app on the Indie plan and marks every release on the same timeline. The first time Riley opened it, the two-year spreadsheet was on screen as one chart, with a dot wherever a version had shipped.
One dot stood out by not mattering. Version 2.8 had added three phrases to the keyword field, including “dog walk tracker.” The rank for that phrase had gone from #34 to #31 and stopped.
One line of metadata
In version 3.1 Riley moved the same phrase out of the keyword field and into the subtitle, and changed nothing else about the listing. Apple gives the subtitle more weight than the keyword field, though it does not say how much.
The rank went from #31 to #8 over nine days and has stayed between #7 and #9 since. Downloads from App Store search rose 44% over the following month. Because 3.1 changed only that one thing, there was nothing else to credit.
One change per release
Riley now treats each release as a single experiment: one metadata change, two weeks of waiting, one sentence written down. The same move has since worked for “puppy walk log” and failed for “pet step counter,” which went nowhere.
“The failures are as useful,” Riley says. “I used to change six things and learn nothing. Now I know which word did it.” The spreadsheet has not been opened since April.
Keyword ranks next to release dates is the chart I tried to build in a spreadsheet for two years.
Riley Chen
Indie developer, Northpaw