by Ray Schultz, Columnist, September 12, 2016
Marketers may be warming up to multi-channel attribution, but they’re not yet getting much out of the iteration known as algorithmic attribution.
What’s that? It’s a “scientific” measurement method. Instead of relying on the last click or touch, it applies statistical methodologies to assign a weight to each marketing touch. One benefit is that marketers can optimize their media buying using attributed metrics that are rebuilt every day.
Yet only a third of the firms using this discipline can make daily or faster buys of digital media, judging by a report commissioned by Visual IQ, a purveyor of algorithmic attribution services, and conducted by Forrester.
What’s stopping them? For 41%, it’s the fact that their marketing and media budgets are locked in place. And 37% say they have too much data.
“There are obstacles,” concedes Bill Muller, CMO at Visual IQ. “Some are self-inflicted by the brands, some are systemic to the marketplace and some are systemic to the partners that the brands work with.”
Miller adds that many firms can’t “shift their marketing budgets between addressable channels like display and search, and between Agency A and Agency B.”
That’s not the only challenge. Some brands are hampered by their technology and agency contracts — like their data, their agencies are siloed by channel and discipline. Others say employees are not incentivized to buy media more frequently.
To get this picture, Visual IQ and Forrester interviewed 51 marketing professionals in the U.S. and the U.K.
Who’s doing it right? If anyone, it’s the e-commerce people, whether they’re pure-play e-tail or multi-channel. “The marketplaces tend to be dynamic and fast-moving,” Muller says. “Sometimes it’s the nature of how quickly the competitors change or put new offers in. Sometimes its inventory based.” (Meaning product inventory.)
Yet most marketers see the need, in part because of the “speed and cadence” of campaigns, Muller explains. Almost 90% say daily optimization is at least somewhat important, and 27% feel it’s very important.
The smart ones know they have to “build their model daily and get freshly attributed metrics every day,” adds Michele Szabocsik, director of product marketing for Visual IQ. “You have to make sure your budgets are fluid.”
Why isn’t weekly good enough? Because the models decay over the course of a week. What’s more, an e-commerce firm with no stores can, by going to daily models, reduce its average error rate by 9%, Muller and Szabocsik claim.
At the same time, daily and weekly feeds can be enhanced with anonymous demographic and behavioral data from BlueKai and other sources.
The firms using algorithmic attribution are especially focused on developing a more precise reach strategy (75%); targeting more effectively (71%); adjusting their media spend across channels (67%) and adjusting their media spend within channels (63%).
Who’ll Do The Work?
That lead us to the biggest question: Who’s going to do the work required for more frequent optimized media contacts?
“You do need some human resources, but not necessarily net new resources,” Muller says. “It can be the same people who are currently buying and optimizing your media, whether at agencies or brands, but using a different set of metrics.”
In 2014, Visual IQ and the CMO Club surveyed 223 marketers. It found that algorithmic attribution ranked last in a field of four processes, with a 4.8% adoption rate. In contrast, 47.1% said they measured the last touch, and 13% performed rules-based attribution.
Sounds like Visual IQ and other vendors have a long haul in front of them. But Muller points to a Winterberry survey showing that cross-channel attribution is top of mind this year for 57.6% of all marketers. Second on the list, for 52.5%, is programmatic media buying for emerging formats.
Muller compares this period to the infancy of search marketing, when people had to get used to aggravations like putting key words in copy.
Let’s say you haven’t adopted Visual IQ’s SAS-based product, but want to adjust your media spend. Visual IQ urges readers of its paper to take these actions:
- Ensure flexible technologies are in place.
- Embrace a mix of human and automated marketing workflow processes.
- Adopt a daily attribution cadence.
- Create service-level agreements that ensure insights are passed to buying technologies.
MediaPost.com: Search Marketing Daily
(12)
