YANN LECHELLE: When Apple decided to deprecate UDID, I came up with an open-source version of UDID called OpenUDID. About 17 ad networks started using it because people were worried about Apple removing something that is useful for the industry.
And here we are again. For no particular reason, Apple is again saying we can’t do certain things. So I took it upon myself to find another solution that will make it possible to do our jobs without depending on Apple. The goal is maximize efficiency and provide the best ads to users and OpenIDFA is an attempt to provide that workaround.
What is your time frame for completing OpenIDFA?
I’m hoping it will take a week or so to collect feedback, and then I’ll work on it some more and release it.
What has the demand been like for OpenIDFA?
So far more than 30 people are looking at the solution. This time I didn’t open source the code and I’m in the process of inviting 15 people, some of whom are from super high-end ad networks, to review the code.
Why didn’t you open the source code this time around?
I’d like to avoid the situation where, with OpenUDID, a number of actors decided to take the spirit of the code and turn it into their own alternatives. That just creates more noise in the community, when what we want is a centralized process.
The challenge is to find something that lets us do our job in a reasonable way while also preserving the spirit of privacy protection, so I built features in OpenIDFA that make it what I’m calling the Snapchat of device IDs.
What features, and what do you mean by the “Snapchat of device IDs?”
OpenIDFA will expire on its own. This will typically depend on a number of deterministic and non-deterministic factors. Whenever a user reboots the device, for example, this may invalidate the former IDFAs generated for that user.
With Apple’s IDFA, the user would have to manually go and reset the IDFA. Most people don’t even know about the IDFA, which renders the actual feature of invalidating it pointless. I’m taking it a step further and without the users’ consent, cleaning it out. OpenIDFA will prevent abuse by expiring relatively often.
Another component will make it harder for ad networks to collect too much data. Each day will have a tracking token that is only good for that day. Let’s say I’m app developer A and I want to know if my app was installed over the course of three days as the result of an action. What I have to do is request three tokens.
I’ll send those tokens to the server and, assuming that OpenIDFA has been distributed, application B could take its own OpenIDFA and compare it with a server. If there’s a match, then that means the transaction was completed as a result of the initial action.
How can clients install OpenIDFA?
In terms of implementation, OpenIDFA is a drop-in replacement code. It’s just another identifier that looks and feels the same, except its own dynamics are different.
Are you planning to sell OpenIDFA?
The goal is to keep it free. It’s released as a library but the code is not open. The library is free to use under the Creative Commons license with a few caveats: It needs to say that it comes from OpenIDFA, was created by Appsfire and adheres to the derivative work clause, meaning people can’t create derivative work with it.
Who do you see as your competitors in this?
Possibly big vendors will impose their own ID. They can create their own flavor of IDFA that is internal and bears no weight in the community. My goal is to create a standard for everyone to operate on.
What are your thoughts on Apple reigning in its IDFA?
Hopefully, we will be able to engage with Apple on a better way to do this. I’m an iOS fan in spite of Apple. They don’t treat their developer community like they should and [Apple’s actions] may backfire over time. Granted, they’re more focused on selling hardware and other things, but it’s not right to bully people around.