POV: Twitter App Graph


Twitter has announced the introduction of a new opt-out service “app graph”, which records the names of the apps users currently have downloaded on their devices. This data will be used by Twitter to hopefully improve ad targeting capabilities. It has already been implemented on iOS last week, with android likely to follow later this week.

Details and Implications

This new service will see Twitter looking at the apps you have downloaded to try and decipher your preferences, in order to deliver tailored advertising content. The list of apps will also be periodically updated, to account for the inevitable new additions and deletions. Although app graph will be an opt-out service, if you have previously chosen to limit ad tracking it will not be automatically activated. Also, the service is relatively easy to switch off, with both iOS and Android offering detailed instructions on how to do so.

Twitter’s prerogative seems clear, it is attempting to learn more about its user base in order to sell more advertising space at a higher premium. This seems a natural progression for Twitter, as it has already been capitalizing on algorithms during the past few months, in an attempt to identify Tweets its users may like and to recommend new people to follow.

With a business model based on targeted advertising “app profiling” seems a necessary addition for Twitter, as in comparison to Facebook’s copious amount of user data,   Twitter is still somewhat lacking in the amount of information it can ascertain from its users. As discernibly, only certain information can be gained from users, based on who/what they tweet and the people they choose to follow.

However, exactly how much of an improvement this information will make to advertisers is debatable. As Twitter has already explicitly stated that it will only collect data regarding the names of the apps downloaded and will have no access to any additional data generated within the apps. Furthermore, it seems likely that many individuals will have the same popular apps installed on their device, which may result in many generic user profiles without enough variation for successful ad targeting. Finally, just because an app is downloaded, does not mean it is used. A trend which is likely only to become more prevalent, given the increase in storage capacity on newer devices removing the impetus to delete apps which are rendered superfluous.


For advertisers this new service should definitely be a welcome addition, as any service providing consumer insights should hopefully reap some positive results as far as ad targeting is concerned. However, successful targeting is intrinsically reliant on the identification of differences and the aggregation of patterns, thus given the superficial nature of the data Twitter will be collecting, the level of targeting possible seems dubious. 

On the other hand App Graph does provide an excellent foundation for more extensive analysis in the future, such as the mapping of buying patterns and ROI predications. Especially following the introduction of the Twitter “buy now” button in September and more recently Twitter Offers a service which automatically adds offers to users’ debit cards to be redeemed in physical stores.