Published July 6
Disney has grown into one of the most tech-forward, data-focused media companies in the world. Its audience graph counts 115 million households and 450 million unique identifiers.
Those figures are critical to connecting advertisers with a more complete view of measurement, which Disney has been sharpening for years now. Measurement was the central theme at its Tech + Data Showcase this January, and that continued at Cannes Lions.
The company’s senior vice president of data and measurement science, Dana McGraw, shared in an interview with The Current why advertisers should look at the whole picture to understand performance, rather than rely on the expedience of last-touch attribution. She also pushed back on the idea that more data is always better, warning that data without a clear objective becomes noise.
You’ve been pretty outspoken about the industry’s reliance on last-touch attribution. What replaces it in practice, not theory?
It’s certainly not just a single thing that replaces last touch.
What you’re getting is who the last person is to get the credit when the outcome happens. But you’re not looking at sort of all of the influence sphere that happened before that click.
It’s really important to think about all of the elements of [a] campaign on different platforms with different publishers in different contexts. How do those fit together to drive the outcome? What are all the behavioral elements that are happening in advance of the final step?
What’s the biggest human mistake advertisers make when they interpret their data?
I think it’s having too short of a time frame that you’re considering for influence. I think it’s trying to oversimplify, and we all do it. It’s human nature.
What is the simplest way for me to get to the answer? Sometimes we end up looking at one thing and we’re missing all the surrounding context. And as I said, missing that halo effect of influence. That was the forcing function for that final thing that you’re looking at. And I think sometimes you can lose sight and you’re only messaging to a very specific niche audience, and you’re missing the part where you’re building your brand.
Is it overconfidence in clean narratives like last-touch attribution?
I don’t know if it’s overconfidence. I think it’s expedience. I think that’s the most expedient thing to look at because it gets hard.
How do you build the math? What do you think about? How do you credit all of these other pieces of the puzzle?
So I don’t know that it’s overconfidence as much as it’s we’re all in a hurry. And so how do I expediently get to the answer I need and prove that the work that I’m doing is valuable work for the brand. It’s less about being overconfident and more about: How do we simplify this and bring all the complexity into something more simple, rather than just taking the simplest point of view and the easiest point of view?
If you stripped measurement down to just one, maybe two signals that you genuinely trust, what would they be and why?
I honestly don’t think you can. I think that’s a big problem that we’re having as an industry, and honestly, I think everyone’s addressing it.
We all understand that it’s an issue. I don’t think it can be dialed down. You have to take a portfolio view of how you’re measuring your success, rather than a simple campaign report. That’s the biggest thing, is that you can’t dial it down to one or two.
What’s your goal? What are you trying to accomplish in this instance? Then those are the metrics we need to look at. And they could be different from one campaign to the next.
What are people really asking for to make sure that we’re optimizing for outcomes, but in the planning phase?
I love this, this is one of my favorite parts of the job. They are asking us for the data in advance to understand the audiences, to have the insights before they even start.
The measurement and the targeting have to fit together. You want the targeting to be so robust and so insightful that the measurement happens before you even get to the end. You know what the outcomes are going to be before you get there, because you’ve done all the work in the beginning to plan correctly. The way those things fit together is the most exciting part, because measurement gets a lot easier if you did the work from the beginning.
Can there ever be too much data, especially in the planning phase?
I think so. I mean, data for data’s sake is not very useful. … Fulfilling your objectives and answering the questions you need to answer is good data. But sometimes, if it’s just all the data and it’s just a mess of things and you can’t make sense of it, it’s not very useful.