Month: May 2013

I’ve been obsessed by the field of App Store Optimization, as many of you know and recently I’ve been exploring the concept of app localization (and in particular keyword localization).  The opportunity of “free”, ongoing traffic is too attractive to ignore and that’s why I’ve been working constantly to improve my knowledge in the field…

Many of you may have heard about the theory of iPhone app localization.  Distimo, one of the ultimate publishers of app data and statistics, put out a publication in October 2012 title “The Implication of App Translations” (it’s publically available, you can download it here). To sum up their research, they found that, one week after adding a native language translation for an iPhone app, the following week downloads in that same country increased by 128% on average, with revenue increasing by 26%.  Interestingly, in the same study, they found almost zero effect on the iPad.

A thing to note about Distimo’s research results – this involved a very large sample size of 200 apps and so the data is highly significant.  What they didn’t include in their research: what does translation mean? Does it mean just localizing content within the app? Does it mean localizing the app name? What about keywords?

While the study leaves many questions unanswered, it definitely shed light on a potentially massive opportunity. Over 120% increase in downloads and over 20% increase in revenue PER country? Why not!

So, armed with this data, I set out to run my own little “experiment” on one of my apps.  Obviously this wasn’t carried out in lab conditions, but I think you will find the results quite interesting…

An Experiment In App Localization…

Here is the graph of the typical language distribution of a non-localized app of mine:

 

Data from “Harlem Shake Yourself”

 

What you can see here, is that over 76% of all the traffic for this app comes from English-speaking countries (including UK English).  The next most common language is Spanish (8.9%) and after that all the other languages are quite insignificant.

You can see the user distribution represented nicely on the image below:

The US in the image above is nicely lit up and pretty much the rest of the world is dark…

And then I localized one of my apps… You can see the stark contrast below:

Data from “Abe The Dragon” AFTER Keyword Localization

 

Suddenly, we can see that rest of the world is lit up! Apart from the US, we have other major loci of users, including France, Italy, Russia, Japan, China…   All of a sudden, we are getting traffic from all over the world and not just the US.

So is it worth it?  You bet!  Here is the pie chart, showing the language distribution:

Here you can see that in the app with the localized keywords English only makes up 10% (!) of the overall downloads (compared to 76% previously)… For this particular app in question, we had overall approximately 23,000 downloads in one month – and based on the 10% figure above, only approximately 2,300 users came from English speaking countries.

Put simply, without localization, we would have had around 3,000 downloads for that month.  With localization we had 23,000 downloads. Localizing this app’s keywords helped increase downloads by 767%! Yes, that’s over 7x the downloads we would have otherwise achieved.

These results are pretty startling… So of course I went on to localize my other apps and time after time the results were the same.

Should you be localizing? Yes! Otherwise you are missing out on potentially A LOT of traffic.

Overview Of The Localization Process

So what was involved exactly?  When I say “localization”, what I have done most of the times is to localize the App Name and the Keywords, with perhaps the top line in the App Description have a translated sentence prompting users to download the app.  In an ideal world I would also localize all the screenshots, description and language strings within the app.  However, with time and resource being limited, I’ve decided to make the change to the areas which will have the most bang for the buck…

I am certainly not the first to “discover” this.  Experienced marketers / developers such as Trey Smith (Secret HQ Games) and Gui Schvartsman (now founder & CEO of Revmob) have clearly cottoned on to this a long time ago, with their apps often being plastered over the top search results for popular localized terms such as “best games” and “fun games”.  However, from what I have seen from performing extensive keyword research in other languages, there is still significant opportunity with relatively few developers really using the power of keyword localization.  It is time to wake up and realize that there is a large and hungry market outside of the US.

Some things that I’ve learnt

French, Spanish, Italian & Russian have done very (very) well.  These three languages stand out from all the other languages in terms of the traffic I’ve managed to generate, often each alone bringing as much traffic as English speaking countries.  This is because in English speaking countries it is already very difficult to target high volume keywords, yet these countries still have a lot of potential for optimization.  I was particularly surprised by how well localizing the Russian keywords has done.

Greek & Korean have done very poorly.  This could be because my translation wasn’t done well or could be a reflection on the overall traffic these countries have.  I’ve since stopped prioritizing translating these languages (for now).

Any Keywords Localized In Spanish Will Be Indexed In The US Store.  This was an interesting loophole I discovered.  When using the Spanish localization (not Mexican Spanish) any keywords you use in the App Name or Keyword section will automatically be indexed in the US App Store (but not any other country).  This is clearly done in order to cater for the Hispanic community.  One, perhaps cynical, use for this loophole would be to include here keywords that you couldn’t “fit” into the English localization.

Exact Match Keywords In Title Carry A Lot of Weight.   This was true a long time ago and the importance has increased and decrease with various algorithm updates.  At the moment it seems that including the EXACT MATCH keyword in the app title significantly increases your chance of ranking for a particular terms (depending on the competition).  What this means, if you are trying to rank for the keyword “Racing Games”, you want to make sure that you have “Racing” and “Games” adjacent to each other, in that same order, with no other words in between.

Caveat

While there is no doubt that localizing keywords for the App Store dramatically increases downloads, I haven’t analysed revenue data which is a bit more complicated, since revenue is generated not only through in-app purchases (IAP) but also through advertising over various ad networks (where I don’t have a country by country breakdown).  Ad networks typically pay out highest in the US.  But, without analysing, I have no doubt that localizing your keywords with significantly increase your “bottom line”.

Tools To Find High Volume Keywords In Other Languages

If  you know how to perform keyword research in general, you follow the pretty same process when optimizing keywords in other languages.  As a primer, you can read my comprehensive post on how to choose keywords for apps.

As a little warning, for me at least, performing keyword research, especially in other languages, takes a lot of work!  It takes me on average 2-3 hour per language.  The most basic set of languages I aim to do for each app is French, Italian, Spanish, Russian, German, Portuguese, Chinese (Simplified) and Japanese… so JUST to localize to these languages takes me around 24 hours of work time!  I’m in the process of training my awesome VA Maria to do this, so slowly hopefully this will take at least some of this workload off my shoulders. (If someone has some clever suggestions here on how to perform highly effective keyword research in less time, I’m all ears!)

These are my favourite tools:

Google Adwords Keyword Tool (GKWT)

This is a fantastic tool to start off your keyword research.  Due to time constraints, I usually don’t go too deep in my research, but instead I pop the translated keyword for games into the GKWT; i.e. in Spanish I would enter the keyword “juegos” and then examine the results.  Make sure to set the GKWT to “exact match”, “mobile devices with full internet browsers”, select the country in question but I usually leave the language to “all languages”.  The reason for that is that I want to see how people search in that country, according to Google.

GKWT Search Results For “Juegos”

Google Trends

Another great keyword research tool is Google Trends.  Especially if you are optimizing for new, topical terms and trends, for example “gangnam style” or “harlem shake”, GKWT isn’t updated often enough.  In that case, Google Trends is a great alternative, since you can plug in your main keyword, select your country and then you’ll see how people are searching for that term in the local country.

Google Trends Results For “Harlem Shake” in France

 

App Store Search Suggest

This is probably one of the MOST important keyword research tools.  When you start typing a search query into the App Store search box, you’ll notice that users are shown “suggested” search queries.  For example, plugging in “games” into my iPad brings up suggested queries such as “games for free”, “games for girls”, “games for ipad” etc… While I don’t have any “proof”, from my own experience and from talking to other ASO experts, it seems quite clear that these results are sorted according to search term popularity.  There is NO better way to see how users are searching on the App Store than the search suggest terms, since while the Google tools are great they don’t reflect 100% the search patterns on the App Store.

To use this tool for app keyword localization, head over to iTunes, select the country in question and then type your seed keyword into the search box to bring up a list of related terms. Note: it is a compromise to use iTunes since results on devices are different – hopefully one of the ASO tool services will add this as a feature soon.

Also, on another note, some of the results returned are app names, some are “artist” names, while others are clearly terms that users are searching for.  The ones I look out for are the ones that are clearly search terms…

In order to help define the most popular queries overall in various countries, I asked my VA to go through the entire alphabet in each language and come up with a list of the most popular terms as defined by search suggest.  You can check out a sample of this keyword list here.  (If you want to receive updated keyword lists for all the various countries, leave a comment below or tweet me and if there is enough demand for this I will produce regular updates…)

Deciding Which Keywords To Target – Keyword Volume

So, the first stage was to find a list of keywords and search terms that users use in order to find apps.  Once you’ve come up with this basic research, you then need to define which of these you’ll want to target.  To do this, you will need to further define the keyword popularity / volume and the competition (do you have a chancing of ranking??).

There are several ways that I judge keyword popularity.

Google Adwords Keyword Tool – Monthly Search Volume

The volume on Google generally is well correlated with App Store search volume, but keep in mind that it is not 1 for 1.

AppStoreRankings.net – ASR Traffic Score

I’m still not sure how much weight to put on this, but ASR is proving to be one of the best tools on the market.  This stat is another figure I’ll take into account.

XYO Monthly Download Estimate

Xyo.net shows the monthly download estimate next to their search results for individual apps.  It gives a good ballpark figure, even though it’s not 100% accurate.  This is only relevant when you are targeting specific app names.

Top Charts Rankings

If lots of apps containing a certain keyword are in the Top Charts, this helps give weight to the importance of the particular keyword.  Luckily ASR shows this data (unfortunately this is limited to US only rankings at the moment.

Number of Ratings

This is probably the most important piece of data that I take into account.  If, for a certain keyword, the top apps returned in the search results all have lots of ratings, this probably means they are getting lots of downloads.

Deciding Which Keywords To Target – Competition

The main tool that I personally use when researching competition and keyword volume is ASR (mentioned above).  This is the BEST tool hands down for app keyword localization since they have an option to choose specific countries for keyword research.

The alternative to this is to use iTunes, however there are three disadvantages. One – iTunes does NOT always show the same results as shown on devices (ASR shows the results that are actually displayed on devices). Two – iTunes is slooooow. Three – ASR gives you handy stats at a glance, including the exact number of competitors for both iPhone and iPad, as well as scores for Keyword Difficulty and Traffic.

For example, if I type the keyword “meilleurs jeux”, “best games” in French (at least as per Google Translate), these are the results ASR comes up with:

You can instantly see how many competitors there are (91), how many reviews (all time and current version), when the app was last updated, whether the apps are free or paid and the US Category Rankings.  Pretty much all the information you need at the tip of your fingers!

There are some things I’d like you to note:

How many competitors are there?  This is one of the most important factors… When choosing keywords, you need to be pretty damned certain you will rank in approximately the top 3 results for iPhone apps (for iPad apps even if you ranked lower you will still have some visibility).  That means, mostly I aim to go for keywords that have, generally very (very) few competitors, unless I have some sort of special advantage, which I’ll go into in a minute…

Do the top competitors have exact or partial keywords matches in their titles?  As mentioned above, have an exact match in the title can give you a significant boost against competing apps.  In the example above, of “meilleurs jeux”, if the other competitors DO NOT have this as an exact match in the title and I do, I will actually have a really good chance to rank in the top (or even 1^st) results… On the other hand, if lots of the competitors do have exact or partial matches, my chances to rank will be low.

Are the top apps appearing in the search results paid or free? Paid apps generally have a much harder time to rank highly, since the rankings ARE based somewhat on download volume.  That means if my app is FREE and the top competitors are paid, I may have a chance to overtake them…

 

A Final Note…

 

In the end, what you should be doing is collating a large spreadsheet for each language.  The columns should be something as follows:

• Keyword (In Local Language)
• Translation
• KW Breakdown
• Number Of Competitors
• Number Of Ratings For #1 Ranked App
• ASR Traffic Score
• XYO Monthly Downloads
• GKWT  Monthly Search Volume
• Top Chart Rankings For Top Competitors?
• Exact or Partial Match In Top Competitor App Names?
• Top Competitors Free or Paid?
• Overall Keyword Volume Estimate

For each keyword I will fill out the relevant information and this gives me the best tool to decide and compare between various keywords and ultimately to choose the best keyword combinations. If you are serious about getting more traffic to your apps, you should do the same!

There is no doubt, from my experience, that:

As always, if you have questions or comments, chuck them in below.