I don’t have that much experience, but I think there is a bit of (as with most things entrepreneurial):
-
randomness (being at right-time/right-place)
-
having idea that is novel
-
having idea that sticks out as different from others (since you are competing for eyeballs)
-
conversely you can also target a genre that is known (from Google Play stats) to be something that always attracts eyeballs
-
none of these address the question of how to have an initial app get big or noticed (esp. in the first month where Google tries to be extra kind - or basically it’s ranking algorithms move apps up in the rankings - perhaps with the intention of finding the place where they will naturally fit i.e. if they are really good they will go up and then stay there).
But regarding marketing, probably the GREATEST impact is to begin with an idea that has an advantage to begin with (because of novelty etc. as above).
I say this because developers often start off wanting to make something THEY want to make (after all this is the motive behind wanting to make apps for a living as an alternative to a full time job). So presumably they would want some more independence.
However if you make something that you want, that is not a guarantee of success - or even of breakthrough of a new idea (that the market is not ready for) - for those types of ideas you need marketing (and even then the first mover may not be the beneficiary of eventual market acceptance), OR you need apps in your portfolio of apps which can drive users to the new app.
It also helps if you have an audience pre-aligned to download your app - so you have some initial downloads - for example if you are a high school student then you have an advantage that you can get your whole school to download the app by marketing it to them first - then time your release on Google Play so you don’t waste days out of that 1 month that Google gives you (to rise in the rankings).
After the first month is over, Google will bring down your app in the rankings - for this reason the first month seems to be essential - so you don’t want to be wasting days.
This suggests that app should not have many bugs to begin with - else get bad ratings and lower downloads/higher uninstalls (which I am assuming Google may include in how it ranks the app …).
One alternative we have discussed here - but no conclusive proof that it will work (well more likely no conclusive proof that it will NOT work - since it SHOULD work) - is that of making an app - getting all bugs out - then releasing an “App 2” version - which has ironed out all the bugs - and now you get a month AGAIN from Google to rise in the rankings - and will do much better than the first time (when you were plagued by bugs or user interface issues as you were tuning your app).
But again it seems that to retain whatever ranking Google takes you to - you eventually need to have users WANTING to use your app. Which means it should not be something that the developer only wants, but that many other people want as well.
All these issues are also related to “discovery”. Fortunately, the numbers (for android) as huge - and the issue is not the absence of users, but how wil users find your app from among thousands of apps.
This suggests there MAY be some value to making apps which target a niche or specialized section of users - because IF it gets known that such and such app is providing this essential service for that niche - then that info will spread through the channels used by that sub-group - and your app could become sought out by those folks - which may be a pretty big group.
So even if the niche is small - the universe of users is (now) so big - that even targeting niches may be of value.
Once one has enough apps out there that have moderate viewership - one could use house ads or in-app advertising/announcements (using perhaps AppBrain Remote Settings to set a flag which tells the app that some new app is available from the publisher etc.).
Conceivably if you have a lot of apps - you have the ability to drive high installs right in the first month of appearance on Google - and this has the potential to make your app rise during this very crucial period.
However an independent developer needs to perhaps think 50% of the time on the code and 50% on marketing - what will work (should I start this project or with that other project be “better” now).
Another thing to consider is efficiency - what amount of effort will get the same bang for the buck type of thing. Because the mobile world is odd in that way - many users may like a SIMPLE but clean idea better than some complicated one.
I have seen some pass-the-time type of games sometimes which include a number of different types of games - this is generally not where the mobile model is going (from my reading) - i.e. on mobile the users tend to want specialized apps which may do one thing well as opposed to all different sort of things (once you have app developers making swiss army knife type apps which do lot of different things - then they wind up being not that good in everything - and user is faced with 10 apps which are similar but each has different quirks). Contrast that with a situation where your app does ONE thing but does it better than all the others - this means users who want that feature will consider your app essential (to keep on the phone) - while the other (more complicated and more time spent by developer) will be seen as replaceable by another app of the same genre.
Marketing may also be difficult for coders/developers who are unaware of those types of issues - but it is possible that releasing holiday season-specific apps could get you noticed as well. Actually that is what I thought around Halloween when I saw a pumpkin carving app in the top charts (check out appannie for charts, ranking graphs). However it is not that simplistic - as a search for pumpkin or Halloween showed that there were TONS of other similar apps which were NOT in the top lists … so the analysis here is not that clear …
PLUS another area to really think about is monetization (if you want to earn some money from your apps as well - which always helps).
And some research (talking here about it) maybe helpful.
But I agree that if that is an issue which will have to be handled esp. if the app becomes big.
One thing to think about also is the possiblity that you include all the ads you want to in the first version of the app - so users are not pissed off at the worsening ad situation for your app - or if users will be unwilling to update to new version if the new versions have drastically worse ad nuisance.
Though some may suggest that worrying about this is over-thinking and if the app is successful one can always monetize it then (true - that for MOST users the old version of the app will not be available anymore or harder to get - so perhaps this issue is less relevant).
In fact thinking there is a STATIC base of users for your app maybe a misconception as well - as practically speaking you may see 50% of lower retention (i.e. 50% to 75% of new downloaders may uninstall the app within a short period).
Incorporate Flurry statistics - you can see the retention ratio graph i.e. how many users uninstalled by day 2, day 3 etc. …
A high exodus from the app (after initial download) winds up making the “retained users” (or “active user” base as reported by Google Developer Console) essentially wind up being dependent on the RECENT download numbers - rather than total downloads.
So you can see apps which have 1M downloads - HOWEVER their 'active user" base (or more practically their “Daily Active Users” - which is what affects ad revenue etc.) may not be going up that fast (because there is constant outflow from uninstalls).
For example from Flurry stats you maybe able to see that after 1 month of first install, only 2% to 10% of your users still have your app installs - if the attrition rate is that high, then the “active user” base will essentially be a moving average over say a 30 day period over your daily download numbers … or something like that.
It is conceivable you could have 1M installs, but only 40K “active user base” (for example if uninstall rate is nearly the same as install rate (and net user adds are very low per day).