How does Google Play ranking work? What are your conclusions about it right now?

Hello all,

I would like to know if any of you have any insight on how Google Play ranking works. I have the impression that is based on the number of active installations vs the total downloads and also on the star rating. What do you think?

Also let’s imagine you have a bad performing application that even after some tries you have not managed to rescue it properly. Is it better to unpublish it or just leave it underperform?

I write this because i have this impression. In March 2012 I published a game that was a instant success, with tens of thousands of downloads daily and really good ratings and comments. The problem it had was that it was absolutely not replayable. People were loving it, but after being in love and hooked for 4-5 hours and finishing it they would just uninstall it.

What happened? After an incredible peak where it was featured on dozens of countries it began its downfall until now it just gets lower and lower and no matter how many time i put on it I cannot recover it. The ratings and comments are still wonderful, the application is the same and even better but it has sank in the rankings and with less users every time, from 20000 downloads to just a few 50 users daily

So does this bad performance impact the developer general performance? (i think so but i am not sure. What do you think?). Is it better to just unpublish bad performers? Have you noticed a improvement in downloads? rating or somewhere after it?

What do you think?

Cheers and thanks for reading,

Interesting question - it is very related to the discussions here earlier about whether making App2 is better than keeping at App1.

Or issuing an app as Beta first - and then as 1.0 final …

We know that some publishers issue Game and then Game 2 - gets benefit of that first month of promotion etc.

Downside is you lose ranking and build up again to it.

However if a game is already going down it may benefit to issue a successor “Game 2” - or maybe even remove the first app from market (?). If you don’t remove, maybe first update the previous app with a screen you can turn on later that will point to the update - using AppBrain Remote Settings you can signal “The upgrade is now available” - the advantage is that all the current users (at least those who have auto updates on for your app) will update and find this message.

Regarding how Google ranks apps - the more I have seen of the stats and how they work, the more it seems that there can be any number of quite legitimate ways to rank apps - and I would not be surprised if Google is using some of those. Most naive writers who talk of rankings wind up using only language that refers to Total Downloads or such stuff.

However, once you have seen stats for your own apps - and how things lead to stagnation - esp. if there is a fixed new installs per day - but your Active User base is increasing - eventually leading to the attrition rate from your Active User base exceeding or matching the new installs - and that is STAGNATION.

So if I was Google there would be any number of ways one could start to analyze apps - and one way might be to include this stagnation metric. If an app is increasing Active User base AND also increasing new installs - that means that app IS SOMETHING and belongs higher up. If the app is stagnant - then let it slide and go below (if it still stabilizes i.e. new installs not going down then bring it back up to a stable ranking place etc.).

So there could be any number of quite legitimate ways to rank apps - and this type of ranking method would be very hard to “game” because it would be based on actually how the app is doing.

That is, if you can bring more installs per day to the table, you will go up (i.e. if you easily exceed your uninstalls per day from your Active User base).

If you have 10M downloads and you are still growing that means you are a seriously popular app.

Earlier threads have discussed the pros and cons of updating Game 1 to Game 2 - and some of the issues covered were whether the Descriptions are identical or similar and that may cause problems.

It could even be possible that Google does a comparison of the code base to see if it is a significant change (?).

I suppose one COULD first change the description of Game 1 to “Game Beta” and change the description - then ready “Game 2” with the new description and publish that.

I don’t have direct experience if there are pitfalls in this approach.

Thanks a lot adforandroidapps for replying and all your interesting ideas :slight_smile: The idea of moving from Game1 to Game2 and the beta version is interesting indeed.

According to Tijmen from Distimo support application performance is ranked based on the app itself and only for each country so it should not affect the performance of the rest of the applications on the same account (like having a small percent of impact on apps based on developer general performance).

Anyway, I am going to extract a csv with all the download information of the past 6 months of my top 5 apps (based on downloads), and perform a statistical corrrelation between them to see if i gain some insight on this.

If it is as mentioned there shouldn’t be any reason why to unpublish an application but that it might hurt your “trademark” somehow if it provides a bad image (like not unpublishing <3 star apps)

On the other side. When you have many apps, maybe could happen that the ocassional visitor that checks the “other apps from this developer” will be distracted by many apps and statistically choose to download other additional application that might not be working as great for you. Instead if you unpublish the bad performers you might turn those additional downloads to high performers. On the other side, the bad performers might attract people as well to your better applications.

Just some thoughts,

I would really appreciate any more insight on this from anybody.

Thanks again,

edit: on a second thought the correlation won’t help as it will only show that good days (or bad days) with many downloads are positive for most of your apps.

I don’t know how google rankings affects downloads (and vice versa) but I can’t see absolutely no logic in it even after months. Ok, I know that you will get more download when in 1st page than when on 5th page, but I think for “not top” apps it’s hard to find some logic. For example, my story - after 2-3 months on market, my downloads suddedly increased by 200-400% (from one day to another) without any promoting, any ads, any reviews - I searched If I’m not in some google charts, I also googled if some web pages doesn’t reviewed my app - and found nothing. I even didn’t update my app. Simple sudden, unexplainable and high downloads change from one day to another. Increased downloads were 2-3 weeks, and then they fall into normal.

Oh, and what’s even more interesting - I’m precisely tracking how my apps are ranked for all possible keywords every day (I’ve java program for that). My rankings - now hear carefully - didn’t absolutely change (maybe ±2 or 3 positions) for weeks (!!!). How can this be explained? For me it’s mystery :wink:

Hi @dna,

It is one big mystery :slight_smile: But being serious - there are some posts on other blogs about the factors impacting this ranking. I did a bit of a research when my company had ca. 10 games on Google Play as well. It seems that Google takes into consideration the following factors:

  • daily downloads
  • daily uninstalls
  • downloads velocity
  • rating
  • how often a given app is used
  • social proof, ie. +1s, comments on G+, etc.

It is a really difficult task to mimic this algorithm unless you have several apps within the same category.

I think people here are over thinking all of this. All I did was put my keyword in my title and a few times (less than 6) and now I am ranking in the 4th position. I went from being on like page two to being in the top 5. I did always have good downloads (500+ a day) and now I’m getting around 700+ with my new ranking. Money has been good too. I don’t think it has too much to do with rating, though I am currently I’m good standing. I’d say it deals with download count and keywords.

All of the things seem to be relevant - it seems if you have a good download rate and low uninstall rate (i.e. active user base is going up) - then that is generally good for rankings.

But if your active user base starts to shrink you can see some slide in the rankings.

However, for any specific app, choosing right words in the description and title also affect rankings - perhaps not overall rankings, but it will affect whether you appear or not when a user types a certain keyword.

For example file is a common keyword, so if your app is File File File And File - you may rate higher when people type “file” - but then may also get banned (!).

AppBrain has a free service called “Apptimizer” which shows how your description etc. words rank and so you can tailor your description slightly to improve it’s “findability” for the niche your app addresses.

i agree with adevinneed that people are overthinking this. from my research, search results are based on 2 factors. 1) total downloads 2) average rating
have a look for yourself. relatively new apps can jump past entrenched apps if their average rating is just .2 or .3 better.

this is for search terms of couse, the top charts are completely different.

Well, i think its simpler. I have an application that was with 10-15 downloads a day. One day it jumped to 300 . I searched and searched and then i found out why. It showed up on the suggestions "users that installed this app also installed that… ". Then i got enough downloads and that got me into the top 500 in a few countries which gave me more visibility. So its law of the jungle , the ones with more downloads in 30 days are the ones most visible. Ratings aren’t also that important, ive seen applications that sayd something like “free online tv” or something, to be on top 10 of the new free apps and beying completely fake, just had a black screen.

Try expanding your marketing reach beyond the Google Play Store. Social Media, and even a separate promotional web page for you app can help increase your internet footprint, and subsequently visibility… Especially if you feel that Google Play limits your ability to display all the features of the app. Take a look at this post for more information on expanding your app’s web presence.

I interviewed at Google (HQ) last year. This topic came up.

Google uses a combination of ratings, G+, download rate and some other things with some factors applied to each. This is replicated in various ‘zones’ (which may or may not match countries) so you can have a hit app in Korea and nowhere else for example. Nobody outside Google knows this info. Nobody inside Google is going to tell you. They certainly were rather cagey with me and it was just a 3rd round interview.

No, I did not get the job. I think that was a good thing, considering [WordHero] :slight_smile:

If you don’t think G+ has an effect on the ranking… take another look at 4pics1word.

Amazing info. Thanks!

Thanks for sharing this ranking secret with us:-)

I got rank 54 in Russia and Ukraine, so the statement, that it’s not country, but area related, seems true

Hello, an interesting topic.

Apptimizer also analyzes app update frequency. I’ve read recently that updating once per month is recommendable. But I see apps which weren’t updated for a year and still rank high. I guess it’s a combination of everything, but if nothing else, you at least “remind” the users about the app by updating it.

Do you try keep an update schedule or do you update your apps as needed?

Thanks.

OK guys what I now found out: uninstalling or raiting or sharing +s affects the ranking in NO WAY. The only thing that changes your ranking is the daily downloads. Why I am saying this?

I am some of the developers that keep fighting about the best ranking in the keyword “dubstep”. I split the little history into 3 parts.
In the first part, I got a few haters that gave my app 1* ratings so my rating dropped from 4.7 to 4.3. My daily downloads kept increasing and my ranking changed +2 (in the better direction). The second phase was a little mistake I made so the daily downloads began to decrease rapidly and my ranking dropped -5 even my rating went up to 4.4.
Now, the daily downloads increased again (and still do it) and my ranking is now between 5 and 9 (depending on the country, gender and age of the user).

Of course, google will “correct” your ranking if it sees, that you are buying downloads, but as you have “real” downloads, it will be the only important factor

Thanks for sharing!

Keywords in comments are no longer used when searching. I know this because WordHero’s description does not contain a particular keyword, but many reviews do contain the keyword. WordHero used to appear on the list of apps returned when searching for this keyword. It no longer appears.

I think you do not have enough data to run a meaningful comparison. Installs velocity is super important (as I have painfully learnt some time ago), as well as comments and +1s. Have a look at the last I/O videos, especially one by Ankit Jain about the app discovery.
Two weeks ago I had a chat with a Googler who told me that the algos check how responsive you are as a dev, ie. if you respond to reviews on G Play.