FRED LEACH: There are a couple ways we’re thinking about it. One is measuring the impact of advertising on mobile vs. traditional display. We’re doing a lot of work to understand what’s the difference between a campaign that includes a mobile component and one that doesn’t.
We’re also looking at how frequency impacts mobile vs. a more traditional display environment and what’s the right mix for a campaign to optimize ROI. The other thing that’s important is within the telco industry, measuring the impact that mobile device advertisers are having. We’re using some of the data Facebook collects to be able to understand when people switch phones.
With these outcome measurements, we’re able to tie advertising to the purchase of a new mobile device, and we can also understand when someone switches their mobile carrier, so it’s really great for mobile manufacturers.
Are you working with third parties to do device recognition?
There are third parties that we partner with that provide mapping services but the device recognition doesn’t involve third parties in the measurement process. Every device has an ID that maps to a name, and because what comes across the Web doesn’t necessarily say "iPhone 5," there are companies that provide that service to map information to the name that most of the world knows the phone by.
How does real-time data fit in the CPG space?
For CPGs, real time has been an interesting challenge. Because purchase cycles play such a significant role, real time on a second-by-second basis is less relevant on the measurement side.
Now CPG companies are starting to do things where they have people install apps, so to that extent, the real-time measurements are just as important. But in terms of its outcome, they’re less concerned about real time.
In terms of marketing, CPGs have this opportunity to optimize their marketing to fit with the zeitgeist moment by moment to have their messaging match what’s happening in real time. You saw this with the Super Bowl, and there were probably CPGs doing this last week during "Breaking Bad."
Having the ability to customize their marketing is important to CPGs but we don’t know yet whether real-time marketing actually leads to better ROI. I think in the CPG space, that’s one of the big questions that they’re being asked.
Closed-loop solutions like the work we’ve done with Datalogix are perfect for answering those questions because you can add campaigns that include a real-time marketing element to campaigns that are more traditional and set something up where you focus on a group with a real-time marketing component vs. traditional messaging and start to answer that question.
To what extent is real-time data being used in the CPG space?
I’d say it’s a single-digit percentage in terms of marketing being customized on the fly. In terms of [real-time data] being used for sales data or effectiveness data, I don’t know of any instances where that’s happening.
Does Facebook aspire to provide an end-to-end attribution solution?
Our operations are to provide what advertisers want, and if we’re asked to build out our services, we’ll certainly do it. What we’re mostly focused on though is participating in independent systems that are designed to develop standards, such that everyone is on an equal playing field and the same measurement system is being used not just for other digital platforms, but on traditional platforms as well. We’re doing that because every signal we’ve seen from advertisers indicates that’s what they want.
Going back to Facebook’s measurement operations, how did this division get started in the first place?
It started from wanting to answer the question, when you’re advertising online, what happens outside of the online environment? Prior to our team starting, most of the measurements were confined to the online space. What we recognized is with metrics like CTR, a lot of research has shown that if you try to use it to predict things offline, you might as well flip a coin since there’s zero correlation.
So we took a step back and looked at radio and TV and we identified one of the key things: There are standards of measurement. For radio and TV, there are agreed-upon measurements by which you know how many people saw a piece of advertising, such as by age and gender. And so we shaped our team around emulating that traditional approach to measurements.
We focus on three key areas. The first is reach. How many people saw a piece of advertising and how many times did they see it? That cuts across about every forms of advertising and is probably the most important one.
Then there’s resonance. After someone saw your ad, did you persuade them in any way? If you’re a movie studio, did they attend your movie? We look for leading indicators like that to see how well your ad might do.
Then there’s reaction. This is where the rubber meets the road. Did people buy your product, go to your store, sign up on your mailing list? One of the things we’re recognizing here is there are lots of different ways to measure reactions and even within the same industry there are companies and brands that care about reactions in different ways. For some it might be basket size or new customers and so we’ve worked to get a third-party system like Datalogix that can be an independent measurement of something that companies want to measure.