Closing the Online/Offline Gap
How do you measure the offline impact of online marketing dollars? Solutions you can use now.
How do you measure the offline impact of online marketing dollars? Solutions you can use now.
Determining the offline impact of marketing dollars spent online has always been something of a holy grail of metrics achievement. After all, you can easily measure post-click behavior all the way to the sale, but who’s buying toothpaste online? (I do, but I am told I am an anomaly.)
When my colleague on the research side of our business, Gary Stein, looked at the state of the art of such measurement methods, his conclusion was fairly glum. A third of all advertisers want their online ad spending to influence offline sales, but “most measure this impact inefficiently and inaccurately.”
Great strides have recently been made in this area, and the further integration — and increasing importance — of online in the media mix will only serve to push such measurement strategies forward.
“It used to be that online budgets were run by the manager of online advertising. It was stuck in a corner and no one paid attention to it,” said Clark Kokich, president of Avenue A/Razorfish West Region. “Now, as chief marketing officers or VPs of marketing are taking a close look at online and asking ‘what can this do for my business?’ they are wanting a broader look at its impact.”
This week, I’ll talk about some of the more efficient and accurate methods out there to help handle this problem.
Old School
In his report, Gary advocates printable Web couponing. Coupons can be uniquely coded, and the advertiser can precisely track what impact the campaign had on offline purchasing. That works for certain products: consumer packaged goods (CPG) and apparel, for example. But it isn’t a suitable approach for others.
To address some of these other issues, Gary describes a “token strategy,” which uses some element to link online and offline interactions. An under-the-cap game could (and did) allow Pepsi or Coke to bring together the online and offline universes. Other possible strategies include providing an online tool (a “configurator,” for example), the results of which someone could print out and bring to a retail location.
Old School + New School (Homescan Online)
What I really want to focus on are emerging technologies that help make these measurements more precise. I talked with Mike Saxon, vice president of media products for Nielsen//NetRatings, to get up to speed on Homescan Online, a partnership between NetRatings and ACNielsen aimed at CPG customers, which launched in November of last year.
“In the CPG world, for a long time, the industry has been able to answer the question ‘how did I do?’ For the longest time, for online, the answer has been been ‘I’m not sure’,” Saxon said. Now, he says, one can say, “you spent this much online, and you got this much back in dollars spent.”
Here’s how it works: Homescan Online put together a panel of households that allow Nielsen//NetRatings to monitor their Web usage. These folks have also agreed to use a scanner to record (presumably) every product they buy bearing a UPC code. Right now, the panel consists of 9,000 households, but Saxon says the company hopes to expand it to 30,000 or 35,000 in the next few years.
So, Homescan can correlate what ads a family sees with what products eventually end up in their cabinets. Limitations include:
In its syndicated version (not the custom per-campaign research described above), Homescan Online could help marketers better target their media buys. In a recent press release, Homescan Online focused on the upcoming Labor Day holiday and said makers of barbecue charcoal should target their online advertising to Web sites that cater to NASCAR and pro wrestling fans. Heavy frankfurter-consuming households can apparently be found on gaming and music sites, such as Toon Disney, AOL Journals, eMusic.com and Noggin.
The syndicated version of Homescan Online goes for $50,000 a year if you’re a current Nielsen//NetRatings client, $75,000 if you’re not.
Retail-Centric (aQuantive)
AQuantive’s addition to the mix is called ChannelScope, launched in December of last year. Avenue A/Razorfish’s Kokich tells me “a handful” of its customers are using it, and others are testing ChannelScope.
Homescan Online is expressly for CPG customers, but ChannelScope won’t work for them. It’s aimed at companies with their own customer data about purchases. We’re talking retailers, financial services companies, automotive manufacturers, and telecommunications players.
ChannelScope uses a proprietary, patent-pending method to match up Atlas DMT ad serving data with advertisers’ own customer purchase databases. Kokich assures me it’s all completely anonymous. No personally identifiable information is compromised, nor does aQuantive have access to its clients’ customer data. This allows ChannelScope users to get a full picture of the sales online ad spending is driving, and it allows for the optimization of the media plan based upon that data.
“It has, built into it, a test and control methodology,” said Kokich. “We hold out a randomized set of our cookies that don’t see any advertising. Then we can do a lift analysis… and you can calculate the in-store ROI.”
Kokich says thus far, his clients have been surprised at how much impact online has on in-store sales. “Advertisers are actually getting more sales in the store than they are through e-commerce,” he said.
Yahoo-Centric (Yahoo)
Yahoo’s solution for CPG marketers is Consumer Direct, which launched last year with initial clients including Kraft, Unilever and Nestle Purina. The set-up is similar to that of Homescan Online, but instead of correlating scanner data with Web-wide surfing behavior, it matches scanner data to a Yahoo ID. Much larger than Homescan’s panel, Consumer Direct tracks 24,000 households.
Consumer Direct can be used as a targeting as well as a tracking tool. In its targeting incarnation, marketers determine an offline behavior they want to target, such as heavy usage of a category, or usage of a competitor’s product. Yahoo finds people who match that profile.
“Then we go out into the Yahoo universe and we find consumers that look just like those consumers,” said Blake Chandlee, the category development officer for CPG for Yahoo
After the campaign runs, Yahoo can determine whether the advertising impacted purchase behavior.
“All of the results have come back very positive on the impact that the Internet can have on offline behavior,” Chandlee said. When targeting was used on the front end, the average sales lifts have been in the 15 percent range, he said, with some programs delivering as high as 40 percent.