ranting technical

Low Friction User Identification on Shared Devices

Long ago I worked on advertising when mobile advertising was not a solved problem and startups still had a play. Early on that work branched out to other possible advertising channels for CSPs. I spent much time studying and discussing the complications around advertising on “Shared Device”. The attraction of mobile advertising from the CSP point of view was the ability to know the consumer. Since there is a tight correlation between a user and a mobile phone, i.e. you don’t tend to share a mobile between two people, the profile a handset usage and other attributes is, or was and I assume still is, highly valuable as it enables better targeting of ads. The big discussions about shared devices came when started to extend the ad models to pay TV CSPs. Can an individual user be identified to better target ads and to avoid showing inappropriate ads to users? At the time the general answer was no. The primary reason being that outside the US TVs are a shared resource in most households (in the US the trend was more towards a one-to-one relationship due to the higher number of TVs per household). On a shared device a lowest common denominator needs to be taken when selecting ads. I’m massively oversimplifying here, and I assume the big brains at the successful advertising companies have a better solution here but let’s take an extreme example to illustrate the point.

Imagine a CSP build a profile for an end-customer based on their viewing habits via the CSP’s pay TV service. This particular profile shows that the end-customer consumes programing targeted at “kids and families” on a regular basis — cartoons — the consumer also consumes more adult, male oriented content — mixed martial arts — (not to be sexist here but teenage and twenty-something guys are the main audience right?). There is no good way on this information alone to know if the household includes kids and adults or just adults who like “kids cartoons” — see brony []. Assume for a second the TV is owned by a single male who does in fact enjoy watching “kids” cartoons. The advertisers for kids products are most likely wasting any advertising spend on this user.

Targeting and personalization is hard, and it harder on shared devices. It’s all about confidence and heuristics and blah blah blah… It’s much more complex than my example but anyway, I told you that story to tell you this one so stay with me.

I have an Apple TV and I recently downloaded the Vevo [] app. Vevo is an app version of MTV from the 80’s — it actually shows music videos. The cool feature that led me down the roundabout train of thought, is that you don’t need to log in on the Apple TV. Instead you just open the Vevo app on your mobile on the same network and presto! You’re in. It does not matter if the mobile is mine or my wife’s, just that the app is on the device and turned on (in focus). This means that, theoretically as I have not seen it in practice in my limited use, my wife’s profile can be different from mine and attached to her personal device and not the shared Apple TV. This could be a good way to link a shared device with an identified end-user. Bake this method of login into a CSPs set-top box and your can offer better personalization of content and a lower friction path to protecting user information.

For example, Netflix allows multiple user profiles, include kids specific profiles which block inappropriate content but users just have to select an icon to use the other profile so kids could select the parents account and watch all the zombie apocalypse they want. Now Netflix could, and maybe the do, add a PIN or password to secure the non-kids accounts but entering PINs and/or passwords via remote controls is limiting and downright annoying on the Apple TV remote with is lack of buttons. Linking the Netflix app on the parent’s mobile allows Netflix to rely on the user authentication mechanism on the device, such as PIN or Apple’s TouchID.

Anyway. It’s a half-formed idea and I see many things to confirm, clarify and comment on as I type this but it struck me as a cool feature I had not seen before that could have potential. And now it’s past my bedtime.