We’re very excited to announce that our iOS Download and Revenue estimates are now available in all +70 supported countries! But that’s not it. We’ve also drastically improved our precision, after collecting data for over 2 years. Find below all details regarding our Machine Learning Technology!
App Store Intelligence: +70 countries now available!
Did you ever wonder how many downloads a given app is driving daily and across countries? Ever wondered how much money your competitors are making on the Apple App Store? Or how many installs would you need to make it to the Top Charts?
…You’re not alone! These questions are very frequent and are actually very important to take into account in your App Marketing strategy.
Understanding the market, the past & current trends but also your competitors’ performance in terms of downloads and revenue can help you take better decisions.
It is crucial to compare your own performance with similar apps, across countries. You want to understand whether you’re outstanding your competitors, whether you’re part of the average or if you’re behind everyone. If you’re increasing in popularity or if your app is losing interest amongst its users.
Furthermore, seasonality is an important exogenous factor that can help explain certain trends and variants in your app’s success. Having a global understanding of your market will help you identify new opportunities and explain some discrepancies.
Although iTunes Connect provides useful insights of your own worldwide app’s performance over time (data available on AppTweak as well), the exact number of downloads and revenue for all apps is kept undisclosed.
Hopefully, various tools like AppTweak have worked hard to provide with estimations for any given app. A year ago, we launched iOS Download and Revenue estimates for 14 countries, relying on Machine Learning.
Today, we have extended the number of countries available. We now provide iOS download and revenue estimates worldwide, which is in all +70 countries currently supported on AppTweak!
Machine Learning: our method revealed!
Transparency is probably the most important key asset here at AppTweak. We are always open to share the way we collect, fetch and calculate data or indicators. This is why we have decided to show you the backstage of our Download and Revenue estimates calculation.
The technology we use to estimate app downloads and revenue is called Machine Learning.
The process is quite simple: based on the existing data that we already have (“real data”), we are able to build a neural network that is learning from this existing data in order to make predictions (“estimations”).
This existing data comes from over 2 years collection of iTunes Connect data synced to our tool thanks to… you! Each time an AppTweak user would integrate their iTunes Connect credentials - in order to have the data directly on AppTweak dashboard - our system would collect, encrypt and ingest all data.
Important: This of course does NOT mean that we share your data to any third party. Be 100% reassured, your data is anonymized, encrypted and safely stored.
Based on the number of app downloads correlated with the app rankings, we are able to determine a predictive curve of downloads & revenue estimates.
To make sure you understand, let’s take a concrete example!
The graph below represents both our “real data” (blue dots) vs “estimations” (red curve) for apps in “All Categories” in the United States.
Image Source: AppTweak
As you can see, the X axis is the spread of app rankings (from 1 to 1500) and the Y axis shows the amount of downloads (from 0 to 60,000).
Each blue point is a “real” and valid data point of an existing app, currently ranked in the U.S. in the “All Categories” chart. Each blue dot therefore gives us the amount of daily downloads and app ranking of each app.
As you can see, higher the ranking, higher the number of daily downloads. Inversely, lower the ranking, lower the amount of downloads.
All these blue dots are shaping a trending curve that we tried standardize into a more general predictive curve (red curve).
Indeed, thanks to the amount of information we have, we are now able to estimate the number of downloads for a given app in a given category for a given region, relying on our predictive curve.
Of course, the more blue dots we have, the more accurate our predictive curve will be for a given category in a given country.
3 Confidence Indicators: LO - MED - HI
For any app, our Machine Learning technology will screen any app’s category and rankings. From there, it will be able to provide with an estimate (more or less accurate) of its daily downloads (and revenue - using the same process).
We’re happy to announce that the accuracy of our download and revenue estimates has improved since a few days now!
Indeed, we have injected more data points (blue dots) in our Machine Learning, which had therefore more real data to rely on. Our predictive curve is thus much more accurate than before.
However, depending on the category and country, the accuracy of our estimations varies. Indeed, the amount of “real points” we have is not the same for each category and country. Again, the more blue dots we have, the more accurate our estimates will be.
Furthermore, the level of accuracy also depends on the ranking tier the app is in. E.g: is the app you’re tracking a very famous app (high ranking 1- 10) or is a newly launched app (low ranking 450 - 500)?
The ranking tier does influence the accuracy of our data as we may have more or less data points for a given group of rankings.
To avoid confusion and data misinterpretation, we have decided to explicitly disclose the level of accuracy for any given app, depending on its category, country and ranking tier.
The graph below represents our real data points vs estimations for the “Games” category in the U.S. As you see, it is a totally different set of data than from the one above (“All Categories”).
Image Source: AppTweak
You therefore need to take into account our confidence indicators while interpreting our estimates and take into account the level of accuracy for the app you’re looking at.
LO: The level of accuracy for the estimates you’re looking at is low. We don’t have a lot of real data points for the given category / country / ranking tier. It’s a rough estimation, that is more useful in terms of comparison between apps rather than providing an actual number of downloads or revenue.
MED: The level of accuracy for the estimates you’re looking at is medium. We have a fair amount of real data points for the given category / country / ranking tier. Although you still need to be careful while interpreting the data provided, these most likely have a good correlation with the exact data.
HI: The level of accuracy for the estimates you’re looking at is high. We have a lot of real data points for the given category / country / ranking tier. We guarantee that our estimates are highly accurate in this case. They should reflect very well the real number of downloads and/or revenue of this app you’re looking at.
Of course, our objective is to improve the level of accuracy of our estimates as much as possible. This is why we are going to make more regular updates of this feature in order to feed our predictive curve with more real data over time.
How to access the new Download and Revenue estimates?
Just like before, our Download and Revenue estimates are accessible via the “Analytics” tab of any iOS app you want to track.
1. Download Estimates
To view the daily app download estimates, click on “Download estimates”. This section provides you with the app’s daily download, for the selected period of time (“Line Chart”). You can easily customize the date range you want to track.
If you mouse-over each portion, you can see the exact data per day. Plus, the data presented on the graph is also displayed in a detailed table below.
Image Source: AppTweak
As you see, our Confidence Indicator (LO-MED-HI) is there to refine your analysis and take better decisions.
If you click on the “World Map” link, you can spy on the app’s downloads per country. Given that we now offer a worldwide coverage, you’ll be able to effectively monitor any app’s performance across all regions.
Here again, the data presented on the world map are also displayed in a detailed table right below.
Image Source: AppTweak
- Revenue Estimates
To view the daily app revenue estimates, click on “Revenue estimates”. Just like for the downloads, the first graph displays the app’s total and daily revenue (in USD) for a given period of time (customizable).
Image Source: AppTweak
The World Map enables you to view in which region the app’s making the most money. Here again, you can view this data in a separated and detailed table as well.
For both the Download and Revenue estimates, you can easily export the data to Excel.
Just like before, our Download and Revenue estimates are included in our Guru (299€ per month) and Enterprise (599€ per month) plans.
With the Guru plan, you can have access to 20 countries (while before it was 14):
- United Kingdom
- United States
- The Netherlands
On the Enterprise plan, all the supported countries (+70) are unlocked:
Argentina, Brazil, Canada, Chile, Colombia, Ecuador, Mexico, Peru, United States, Uruguay, Venezuela, Austria, Belarus, Belgium, Bulgaria, Croatia, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Norway, Poland, Portugal, Romania, Russian Federation, Serbia, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey, Ukraine, United Kingdom, Australia, China, Hong Kong, India, Indonesia, Japan, Korea, Kazakstan, Malaysia, New Zealand, Pakistan, Singapore, Taiwan, Province Of China, Thailand, Vietnam, Philippines, Israel, Jordan, Kuwait, Lebanon, Oman, Qatar, Saudi Arabia, United Arab Emirates, Algeria, Egypt, Nigeria, South Africa, Tunisia, Zimbabwe.
If you are on the Starter or Pro plan, you are more than welcome to upgrade in order to access this data right away.
Got questions? Reach out or schedule a quick call!