Attribution Management: Credit Crisis
In its annual Christmas book released this fall, Neiman Marcus showcases the usual selection of over-the-top luxury gift items, such as a $250,000 houseboat for two. One of the more interesting developments regarding this year's catalog, however, is that for the first time it's also available on Apple's iPad. Using the wireless tablet, consumers can view the catalog, click on specific items to watch videos about featured gifts and/or place an order.
New marketing vehicles such as the iPad bring cross-channel retailers innovative ways to reach consumers. But they also complicate life, making it a challenge to get a clear picture of which marketing activities are driving the most traffic to a website. Without accurate information about which channels are performing best, multichannel retailers struggle with where to invest their marketing dollars and how to get the highest return on their investments.
The process of allocating orders in any channel to the advertising vehicle that caused them to happen is called attribution management. The idea has been around for a long time, and can also be called order allocation. Or, in the offline world, matchbacks. The need for better attribution management "has become more pressing as folks spend more money on paid search, display ads and retargeting," says George Michie, co-founder and CEO at search marketing firm Rimm-Kaufman Group. This is because inaccurate reporting can result in cross-channel retailers giving credit for a single order to both an offline and online vehicle.
"I think a lot of retailers know double counting is happening, but the challenge is how do they find the right solution to address the problem that doesn't cost $5 million and take years to develop," notes Al Bessin, consulting partner at direct marketing firm LENSER. Less-than-accurate reporting can result in management issues for some cross-channel retailers, with online and offline divisions each believing it deserves credit for driving the most orders — and therefore the bigger chunk of the marketing budget.
Bessin, Michie and other industry experts insist that while there's no perfect solution for allocating orders, retailers must attempt to gain a better view of their marketing activities if they want to improve conversion rates and ROI numbers. In some cases, the financial impact can be significant. Attribution management services provider ClearSaleing, for example, was able to help one client reduce its cost per action by an average of 43 percent, according to a recent report from Forrester. Another client reduced its ad spend by approximately $100,000, while increasing its conversion rate by 50 percent; and yet another saw savings of up to 20 percent per marketing channel.
When The Wine Enthusiast turned its focus to better attribution management last year, the benefits included saving money and finding a new way to reach customers, reports Glenn Edelman, vice president of marketing for the cross-channel seller of wine accessories, wine cellars, glassware and more.
The Wine Enthusiast signed up as one of the beta testers for Rimm-Kaufman's new attribution management software with the goal of finding out how efficient its online marketing activity was. "We had this incredible fear that for every order customers were touching every single marketing touch out there," says Edelman. The retailer's marketing efforts include catalogs, paid search, email, comparison shopping engines, affiliates and organic search.
The results of the 90-day-plus test both reassured and surprised Edelman. When 80 percent of its orders ended up having only one or two touches — with the second generally being a branded search on Google — Edelmen breathed a big sigh of relief. "The overlap wasn't as big as we'd thought it would be," he says. If a significant number of customers had been touching multiple marketing vehicles, this would have negatively impacted the retailer's ROI, he explains.
The surprising part was that for the approximate 20 percent of customers who did interact with three or more marketing touchpoints, the last click was generally an affiliate site. "That was eye-opening for us," says Edelman. Further research tracked most of those last clicks back to customers who went to Google to search for "Wine Enthusiast coupons" and wound up on an affiliate coupon site immediately before placing an order.
As a result of this research, The Wine Enthusiast bought the domain wineenthusiastcoupons.com and created its own coupon site. "I can't stop people from googling 'Wine Enthusiast coupons,' but I might as well get them to come to my site and not have to pay a commission to an affiliate," Edelman says. The money the retailer is saving in lower commission fees is being passed on to customers in the form of better savings, he adds.
The Wine Enthusiast's current attribution management focus is on trying to understand how its offline and online marketing interact.
"The Wine Enthusiast started as a catalog company, and the catalog is still a very important part of the business," says Edelman. However, because catalogs are significantly more expensive than some of the retailer's online marketing activities, it's trying to figure out which online orders are driven by a catalog and which ones would happen whether someone received a catalog or not.
One Size Doesn't Fit All
The only way to get at this type of information is by doing holdout tests, according to industry experts. A holdout test is when a retailer takes a segment of its customers, doesn't mail them a catalog and watches the results. "By not mailing a catalog or an email campaign, you truly learn whether these vehicles cause orders to happen," says Kevin Hillstrom, president at database marketing consultancy MineThatData. However, "most marketers aren't willing to do holdout tests" for fear of losing revenue, he continues.
Testing needs to be ongoing, too. "Testing for a month usually doesn't uncover what you're looking for," says Bessin. Any initial testing should last through one selling season, and follow-up testing has to be done because both marketing strategies and customer preferences change.
Because of the perceived complexity of attribution management, many retailers are hoping someone will come up with a one-size-fits-all formula for questions such as "How big a role does my catalog play in website sales?" This isn't likely to happen. "Not only does the degree of channel interaction vary widely for every different advertiser, but the type of interaction varies tremendously as well," explains Michie, about why coming up with one overarching formula to answer these questions isn't possible. "Every company is different, and you really have to be careful about applying generalities across different businesses," agrees Bessin.
Edelman agrees that retailers have to continually test their assumptions when it comes to allocating orders correctly. He's currently planning some follow-up tests to see how factors such as The Wine Enthusiast's new coupon website and social media activities are impacting where the brand's orders are coming from.
"The landscape is always changing, and we have to adapt," says Edelman. But, "as there are more marketing channels becoming available, it becomes that much more imperative to have accurate attribution data," he continues, "in order to know how your marketing activity is changing customer behavior."