I'm happy to jump in and follow up.
I have run three big departments before Treasury Board. There's huge diversity in the programs and actions they're taking.
Treasury Board, for example, would see a cement procurement. The data analysis and GBA+ done around a procurement like that is going to be completely different in its intensity and nature than, for example, the launch of a major new social program.
Data is critical to all of those, but with the data, I have found across departments that there's no one size fits all to these things. In many cases, you don't have data to start with. You have to devote resources to collecting it. Important investments have been made on that front to do that, but it's not that you can do it with just one spend. You have to customize it to each program to figure out what the data is. In some programs you may want to focus particularly on some aspects of the GBA+ where you know that it's much more important for the program.
When you ask if we can collect the data, another huge barrier historically has been privacy. I remember the time when parliamentarians criticized ESDC for creating what was then called a “big brother” database to try to link datasets—even though it was anonymized—to attempt to get better disaggregated data. That's why we have issued the guidance on how people can act in the privacy space.
Lastly, there may be cases where the communities and the GBA+ sectors themselves would be quite uncomfortable with us collecting the data. I will give a concrete example for you.
Parks Canada has a reservation service to reserve a park site. From a disaggregated data perspective, it would probably be very useful for us to know the racialized status and the sexual orientation of each individual who's applying, to understand if the service is being experienced the same by everyone. You can imagine that some of those communities might object to the notion that in order to reserve a park site, they have to provide all that information.
Those would be cases where it's actually not appropriate, at the end of the day, to collect the data. That will create some data gaps.
I think part of the challenge in the progress is that it has to be done on a program-by-program level in a quite customized way to be consistent. That's what has taken a lot of time.
That's why, as Andrew said, it's not a one-size-fits-all thing.
