Six tips on measurement from the IAB Programmatic Marketplace


Not getting too complicated with metrics is just one important point covered in an IAB conference session all about attribution.

Earlier this month, the Interactive Advertising Bureau (IAB) held its annual Programmatic Marketplace in New York and here are the six key takeaways from our favourite presentation: “Cross Device Measurement and Attribution: Accuracy, Efficiency and Impact.”

1.Know your audience

According to eMarketer, 84 percent of U.S. adults use the Internet. Of that group, 81 percent do it from at least two devices.

The first thing you have to do is make sure your data collection is strong. In order to measure someone’s customer journey as they toggle devices, you first have to know who they are.

The second thing you have to do is make sure your user experience (UX) is equally strong. No matter who your customer is, you can be sure that they have no patience for slow-loading mobile pages or any of the other practices that compelled the IAB to launch its LEAN Ads Program back in October.


LEAN is an acronym for light, encrypted, ad choice supported and non-invasive. In other words, all the things your ads should be, if you want people to engage with them (and not block them).

2. …and know what you’re up against

Knowing their audience is just one challenge marketers face when it comes to attribution. There’s also the walled gardens of Google and Facebook, the fact that different consumers share the same devices, and the availability of multiple demand-side platforms (DSP), just to name a few.

Multiple DSPs means attributing them, in addition to impressions and conversions. GroupM’s strategy there involves a concept the agency calls “mega DSP,” says Nikos Tsagaroulis, director of programmatic optimizations, analytics and data at GroupM Connect.

“You basically have a centralized hub where you manage which DSP is good for which campaign and which audience,” continues Tsagaroulis. “This is where programmatic measurement is moving, as well.”

3. Don’t get too complicated

There’s no question that attribution is a complicated process. But Dan Murphy, senior vice president of audience measurement and analytics at Univision Communications, points out that it’s not the metrics that are complicated.

labyrinth 2

There are really only a handful that matter: reach, duration and frequency, for example. Marketers often confuse themselves by measuring too many things beyond those basics.

“What makes [attribution] complex is, we want to make sure it’s comparable. We also want to make sure its viewable and that its seen by a human being. We want to start introducing data to qualify the audience, so all these things make it complex,” says Murphy. “It’s like a sausage; how many people really want to know what’s in the sausage? But it tastes good.”

Duration is particularly important to Tsagaroulis, who recommends factoring exposure time into DSP algorithms.

“What that means is, there is a set of placements or a set of audiences or a combination of the two that has been proven – always in real time – to have high exposure time,” he says. “Through linkages provided, you can say those are definitely joined to drive high conversion rates and ROI.”

4. Find a measurement balance

“Always in real time” was an important distinction in Tsaragoulis’ point. Today’s technology allows marketer to understand the impact of their campaigns as they happen.

For example, you may have noticed that over the past few months, ClickZ has been less newsy, with more of a “best practices” focus. That decision was based on our Google Analytics, which let us know the kinds of posts that resonate most with our readers. For March, our most read so far teach you about executing Facebook ads and collecting data.

The benefit of real-time is that if you wait six months to measure something, by the time you’re ready to act, everything has changed anyway. But at the same time, it’s important not to get too deep in real time. Measuring on a minute-by-minute basis results in way too much noise.

5. Involve creative

Data and creative are often separate entities, but they shouldn’t have to be. In fact, they shouldn’t be. Dynamic creative – ads that change depending on who’s being targeted – is growing in popularity. Google even incorporated the practice into its Web Designer in November, following some very successful tests with a 580 increase in click-through rate (CTR) and a 75 percent drop in cost-per-click (CPC).

“On the fly, machines are getting smarter and smarter,” says Tsagaroulis. “You are basically reconstructing pieces of content that you have already, but now you’re tailoring it to the location, the audience, the weather.”


The rise of dynamic creative means that data dictates your creative. But it also makes it more measurable at the same time.

6. Road less traveled

Tsagaroulis recommends doing as much testing as possible because “everything is changing as we speak.” Murphy adds that it’s important to not to employ a tactic just because you’re comfortable with it.

For example, Last-click attribution is a fairly traditional method of analytics. But its star is fading, as marketers increasingly long for something more sophisticated.

“The road less traveled makes all the difference,” says Murphy. “We need to stay away from the easy short-term money when we know the short-term negative consequences.”

This article was originally published on our sister site ClickZ. We’re republishing a handful of their recent articles over the Bank Holiday weekend. Go give them some love.

Related reading

herding cats
Shift London
A screenshot of the Plonked search engine company page for itself (Plonked).


Source link