Google’s launch Tuesday of Data-Driven Attribution for Google Analytics Premium users left some questions on the table, including how pertinent offline data would be to the product, where “earned” media fits into the mix and what implications there are for “non-Google” marketers and the entire attribution ecosystem.
Forrester Research, which released a Forrester Wave Report for Attribution Vendors last spring, at the time, recommended Google’s attribution tool solely for marketers that run Google shops from both an analytics and advertising perspective. Tina Moffett, Forrester analyst and one of the authors of the Wave, spoke with AdExchanger about Google’s latest attribution effort.
Any immediate thoughts on Google’s deeper attribution modeling?
It is, absolutely, the direction Google needed to go in. What they realize is though they have a powerful analytics tool right now, they need to build beyond the Multichannel Funnels and attribution tool, which is a good start. They knew for them to be serious contenders for marketers who are making significant investment in analytics, they needed to up their game, which means a few things for them.
It means building a more statistically grounded approach to attribution. They’re applying more statistical rigor around quantifying conversion paths and giving credit where credit is due. Second, they’re not only including a lot of the Google Analytics products, but they’re pulling in data that marketers can upload through universal analytics, so offline data, which is huge. When you make the connection of online to offline conversion paths, in addition, then you can bring non-Google data and information in from what I’m hearing.
How did they really prep for deeper attribution before the launch?
I think everyone in the analytics space knew that this was eventually coming from Google, but didn’t really know when. What I was kind of more surprised about was when we did the initial Wave about a year and a half ago, Google Analytics was in the Digital Attribution Wave. They did okay… but [I think everyone] knew Google had to amp up their game a little bit more and we’ll see in case studies if marketers are finding the benefit of using it.
What do you really see as a strength of the “pure-play” attribution vendors (Visual iQ, Convertro, etc.) vs. the larger cross-platform plays like IBM and Google that were evaluated?
I think with the pure plays like Visual iQ, Adometry, Convertro, DataSong [formerly Upstream software] they’re kind of positioning themselves as being data-agnostic. They really just want to input any type of data and do really good and interesting analysis on it, whether it be using attribution to understand marketing efficiency and effectiveness or using marketing attribution to really quantify conversion paths and identify like conversion paths and high-value conversion paths. They really take that positioning and it’s really appealing to buyers who have a lot of their data in different data sources because they can just pull all of that information in to one data source and run the model.
The pure plays are really well positioned in saying, ‘We’re an unbiased source of truth for you, Mr. Marketer.’ This is a challenge for Google, because I think there’s a perception out there that, ‘Google has a ton of my data right now and why am I going to give them more data?’ Even with the ability to input data to this universal analytics tool, I still think they’re going to have that kind of challenge. I don’t think that it’s in any way discounting their methodology or approach, I just think it’s reality.
Where do you think Google and other attribution vendors stand not only in assigning value, but helping alter the media buy in real time with those value insights? This whole argument about making attribution more actionable.
All of the measurement vendors are looking to close the loop and make this information actionable. It’s really imperative for these vendors to… automatically push out those values in to the buying system so that search bid platforms and the DSPs can put those true values in to their optimization system so that they know the minimum and maximum amount a marketer would pay for certain media and a certain search term.
But then there’s this question of, ‘What’s real time?’ I mean, every marketer says, ‘I need this in real time but the question then becomes, ‘Do you really need it in real time?’ ‘How much of an incremental impact would having this information in real time [have on making] real-time decisions?’ And that’s a real question that marketers should be asking and really demanding companies answer and prove out value.
I think marketers need to look at, ‘What is really real time?’ and do a study around, ‘What’s the incremental dollar-value benefit of making marketing and media decisions based on real time?’ I think, closing the loop in general around media buying and putting those values in to the media-buying system, is needed.
