Are we on track for launch this year?
Yes. We’re making good progress. That said, we have a very tight sequence of processes ahead and there’s no mistaking how much there is to do before we can roll CFlight out.
CFlight is a pioneering organic data creation rather than a boxed product, so it’s not entirely predictable. But, unless we’re struck by unforeseeable events, it will launch in Q4 as planned.
What are the key milestones to reach before CFlight can launch?
Two specialist suppliers – RSMB and TechEdge – who are well known for their work with BARB, were commissioned to respectively design and build/operate CFlight. Most of RSMB’s design work was completed during 2020, resulting in a 62-page instruction manual of algorithms that TechEdge were handed early in 2021, and have been working on ever since.
They consist of a sequence of processing steps, and on completion of each step TechEdge creates a worked example using real campaign data for RSMB to then scrutinise and validate. Each step involves some fiendishly complex calculations.
It’s ironic given the apparent binary simplicity of combining linear spots with BVOD spots, but the two platforms are chalk and cheese at a technical and measurement level.
There’s also a lot demanded of the data in terms of how it is reported, whether now or in future phases, all of which has to be prepared for at this stage. But it’s all on track.
Do you have testing partners on board already or can people apply?
To be as even-handed and inclusive as possible we’ve turned to an industry group for assistance with testing. The IPA’s Television Research Advisory Group (TRAG) has representatives from across the agency world who have kindly volunteered to be the first CFlight testers outside the project participants.
Testing partners will have to conform to the criteria that applies for gaining access to the tool, which is that you have to be a buyer of BVOD and linear airtime.
What have been the biggest learnings from developing CFlight?
We’ve been developing a post-campaign cover evaluation tool as fast as it can be done, but it must be done well. To launch relatively quickly has meant not including every desirable feature it could possibly have, because that would take years longer to kit it out. Imagine developing an industry measure like BARB from scratch – how long would that take?! So we’ve learnt along the way what agencies and broadcasters want most (beyond having an operable tool), which has informed future phases.
It’s also fair to say that few fully realised the scale and complexity of such a big data collaboration. If you’re thinking of doing something comparable, I suggest you start this week.