This report has been a long time coming, but I unfortunately havn’t had time lately to post on our results and experiments.
I’m not a blog writer, so I apologise if it’s long winded or boring, but I think some might be interested in some of the details.
(For those unfamiliar with us, Ambrosia Designs is a two-person “company” with a goal to generate sufficient income from android in order for us to eventually work on it full time)
On March 11 2012 we released our battery widget - Ambrosia Battery Widget. The concept is a battery level meter that has various skins. 3 skins are free, the rest can be unlocked by completing tapjoy offers or paying via Zooz to unlock all skins at once.
The widget was more an experiment than anything else. I did hope it would do better than it did - and if so we would then code more widgets of various types. My initial thought was that widgets are useful for ALL devices, ALL user demographics and therefore the potential for downloads should be massive.
Sadly something like a battery widget means I’m competing with about 100+ other similar apps so getting downloads is basically trying to grab a piece of the pie.
Initially we purchased some graphics off various sites that offer licenses in order to kick-start the number of skins. We have since added some of our own artwork to the list to increase the number of skins.
We also translated the app into Korean, Japanese and French about 3 weeks after releasing it. Translations didn’t seem to affect downloads much, if at all. We have a large number of Korean active installs but these were increasing even before we added localization.
On the 19th we paid for some admod advertising to see how it would affect downloads. I set it with a limit of $10 per day and paying $0.01 per click for 5 days ($50). Basically giving us 5,000 clicks over 5 days.
It gave the app a big boost, as you can see in the graph it jumped massively. We went from 15 active installs to 312 (note that there was a delay between day 1 and the remaining 4 days due to a change in the ad). In terms of stats it means we got about 0.05% install rate from the advertising - not particularly great.
I think it’s also worth noting that interestingly the first day of advertising gave us ~140 active installs from India, the remaining 4 days gave us a total of about ~160 from other countries, mostly indonesia. I didn’t try targeting specific countries, and at $0.01 per click it’s also likely we got poor placement of the ads. But this was for experiment and most importantly to see how many downloads we could get.
Now obviously what everyone is wanting to know is how much was earned? Unfortunately very little. Since March 11 we’ve made about $13. This has all been from Tapjoy offers so far.
There is a very big reason for this low number though. Most users do not know how to use widgets. This is a big problem. The percentage varies, but on average only about 60% of users who install the widget manage to find the settings page. This obviously results in a higher uninstall rate than one wants and a loss of potential revenue as well.
Below is the tapjoy graph. Our total downloads on Google Play market are 1,545, the app is also on other markets such as slideme, getjar, etc. Total new users on tapjoy is only 938 - tapjoy will only reflect a new user if they open the settings page, it will not initialize for the widget when it’s displaying on the desktop.
As an extra expiriment, we decided to try out the tapjoy offerwall advertising on 10th of this month. This proved to be very interesting. On admob $50 gave us 5,000 clicks but of those 5,000 only approx 450 actually downloaded it. So it basically cost us $0.11 per user of which approx. 60% found the settings so closer to $0.18 per “real” user.
We transferred $10 from our tapjoy earnings to the advertising budget and set it to run at minimum cost of $0.10 per install. Now the way tapjoy does the costing - for those who don’t know - you only pay for a user if they install your app AND run it - basically if your app runs the tapjoy initialize code and sends a deviceid to tapjoy only then do you pay for the install.
Well $10 went extremely quickly! Tapjoy support did advise us in email when the ad was approved that we should have a minimum budget of $1,000 but that’s not exactly the kind of budgets us indie developers have lying around
Within an hour there were already ~90 clicks with 22 of them resulting in paid installs. Tapjoy’s system slowly trickled the rest over a few hours with the 96th paid install happening just a few hours later. Obviously as mentioned previously since a low number of users manage to find the settings, we got more downloads than paid installs. I must admit a small part of me thought it’s a brilliant way to get high downloads for low cost, but the drawback is the large number of uninstalls (which happen very shortly after the install) which negatively affects your google play ranking.
The installs from the tapjoy offerwall have not generated any income as yet, but I was expecting that. Most users install an app from the offerwall in order to get coins/points for the current game they are busy with. They are not interested in following a chain of apps and offerwalls. However my hope is that they leave the widget on their screen and a few days/weeks later they look at it again after they’ve forgotten about it and unlock a skin.
On my list of changes for the next release is to add a rate my app popup and more importantly (after seeing a widget with this feature) - adding a shortcut to the homescreen when it installs that when run gives instructions on how to add the widget. This should increase the percentage of active installs and revenue as well.
When I make the change with the instructions I’ll give a followup report as it could be interesting.
I hope this report gives some developers useful info. If anyone has any questions I’m happy to answer.