[Many thanks to Stefan for allowing
me to write this guest post. My name is Jay Van Buiten, I’ve been a solo app
dev for about 5 years, and I am promoting Review Command, a beautiful new Mac app
that lets developers get all their app reviews & ratings, across multiple
apps, platforms, and countries, into one feed.]
How do you decide to use your 100
characters worth of App Store keywords? Probably a combination of looking at
competitors, tools like SensorTower and Mobile Action, and just guessing. But
why not find out what users actually search for when they have the problem your
app solves? It’s possible with Amazon’s Mechanical Turk, and in my experience
can 2X your downloads compared to traditional keyword optimization:
Mechanical Turk (mturk.com) is a marketplace for getting small,
online, repeatable tasks done by real people. It’s very cool.
For each localization, just follow
this process (I’ll use my app Video Merger as an
example):
1.
Create a survey on mturk.com
that describes the scenario/problem that would lead to a user needed your app.
In my case: “Imagine you have 2 videos on your phone, and you want to combine
them so you can post them to social media as 1 longer video”. Use something
like Unique Turker to make sure
each person can only respond once.
2.
Ask the user what they would search for on the App Store if they
were looking for an app to solve this problem. I like to ask for 10 search
possibilities from each user.
3.
Get about 10 responses (I usually pay about $1 per response). This
gives you 100 search phrases.
4.
Use a small script, or maybe Excel if you don’t code, to assign
points to each unique word in the responses. If a word appears in a user’s
first search, give it 10 points, if their second search, give it 9 points, etc.
5.
Rank the words by most points: Now that you have a list of the
most “powerful” search words, simply remove any words that already appear in
your title (Even better: add some of these words to your title), use the first
100 characters of what remains, and you have your keywords!
6.
Repeat for each localization (make sure to only target applicable
countries with your survey, and ask for responses in their native language.
Example: For a french localization, target only France on mturk, and ask
specifically for french responses.). Note that it may take longer to get
responses in some countries.
Here’s my downloads for the 2 weeks
before, and the 2 weeks after, I submitted an update with mturk optimized
keywords for Video Merger.
Hopefully you’ll have similar
results. If you’re looking for a way to find powerful keywords your competitors
won’t find on Sensor Tower, AppAnnie and other ASO tools, try asking real users! Enjoy!
If you
have trouble keeping up with your app reviews for different apps, platforms,
and countries, check out Review
Command. It’s a
beautiful Mac app that shows you everything in one feed - you can check reviews
across your entire app business, instantly. It’s 80% OFF for a few more days ($99
-> $19).