The Click That Counts
The Click That Counts
The online path to purchase can take unexpected turns. Customers click through a banner ad, drop a product in their cart, but not buy it; then click through a search result and make the purchase. This creates mixed signals for marketers trying to determine which online channels and ads are contributing to sales, where the last ad clicked was probably not the only one responsible for the conversion.
“Last-click attribution methodology has been the standard for a long time and a lot of people understand intuitively that it's just not accurate,” says Ari Osur, general manager of eBay's analytics subsidiary ClearSaleing. “It's kind of rewarding the last swimmer of a relay race and forgetting the other three that got them there,” he adds.
ClearSaleing's research has shown that a customer has an average of 3.8 touchpoints with a brand before taking a desired action, which is why brands need a more comprehensive view of customers' purchase pathways. This requires pulling data from ad servers, search management platforms, and other sources, and then placing it in a hierarchy.
Knowing which click was most responsible for a sale strongly affects brands' online strategies and lets them budget and prioritize their marketing spend more efficiently. Air New Zealand, for instance, initially made marketing decisions based on the last click customers made before a purchase. This lead to some erroneous assumptions—for instance, the belief that its display advertising generated 1% of the company sales.
But when Air New Zealand installed a tag management solution from TagMan, it found that display advertising actually generated 30% of its sales. “When we actually looked at the data on interactions in the middle of the cycle, we saw that display had a huge impact, in terms of pulling our customers through the conversion cycle and pushing them into other channels,” says Chris Allison, online channel manager at Air New Zealand. “Customers were much more likely to convert when they had this interaction.”
Other misconceptions revolved around search—assumed to generate 35% of sales when the number was closer to 17%. Air New Zealand revised its budget to better reflect these attribution findings, putting more into display and less into search. The result was a 15% increase in revenue year-over-year.
The airline also shifted from a last-click to a flat attribution model, in which each impression is given an equal portion of the sale value; so a $1,000 sale in which the customer saw four search ads, four display ads, and two email promotions would give each of those 10 impressions credit for $100.
The nitty gritty around the value of clicks is often ignored by brands. “Marketing organizations are traditionally very channel and campaign focused,” says Forrester Research analyst Tina Moffett. This is partly due to individuals protecting their budgets and partly because those individuals don't fully understand other parts of the marketing mix.
Moffet recommends an advanced, cross-channel statistical approach to attribution, which goes beyond the flat attribution model. This includes drawing on loyalty program and CRM data, so the impact of various touchpoints can be weighted, as well.
Jon Baron, CEO and cofounder of TagMan, says he expects attribution to get more sophisticated in the years ahead, incorporating mobile and offline impressions into the attribution model. He points to companies doing “real-time econometric marketing” that considers TV ratings data, billboard, direct mail, and online data.
“The world's moving toward automation and online/offline coming together,” Baron says. “The benefit to the consumer is we won't be hit by 5,000 ads a day, of which 4,980 are not relevant to us.”