Advertisement

Here's How Facebook Plans to Help Brands Find Out When Their Ads Actually Worked

'Lift' addresses a nagging industry problem

Facebook hopes the industry moves beyond last-click attribution with the help of data.

Facebook is looking for an answer to that age-old marketing question: Which part of your advertising is actually working? The social network is launching a new product today to help brands assess which portion of their ad budgets are growing sales in-store and online.

The product is a measurement tool called Lift, and Facebook says it will tell brands when their ads on the social network actually boosted the bottom line. Lift works by studying groups of people who were exposed to the ad and people who were not. This is all part of Facebook's plans to gather the best data it can on ads running on its platform, and it thinks the new way of measuring their effectiveness could ultimately benefit the industry.

Of course, Facebook has a stake in finding proof that ads on the social network are effective. A lot of mobile-centric, non-search advertising fails to register as a deciding factor in purchase decisions, and not just on Facebook. Brands and their ad agencies often have trouble registering the impact of mobile campaigns, and that can lead to crediting the last click for any purchase or action, when exposure to ads earlier in the buying process may have had just as big an effect.

Facebook says this last-click attribution can be misleading. "It's like thinking the sign outside your store is responsible for foot traffic," said Brad Smallwood, Facebook's vp of measurement and insights.

So instead of investing in more effective advertising, a store just invests more in its sign, he said, continuing the analogy.

Facebook is uniquely positioned to connect ads served to actual purchases because it has an active user base of 1.3 billion people who use one log-in ID for all their devices. Also brands are able to link Facebook users to sales with their own databases that communicate anonymously with Facebook data through ad programs like Custom Audiences.

By no means has Facebook perfected the attribution, but it thinks the step of offering Lift analysis is a start. It has found that brands often mis-attribute, and in turn misallocate ad dollars. In fact, 90 percent of users who saw an ad on Facebook and then made a purchase didn't actually click on the ad, never registering credit for the social network. Meanwhile, perhaps the next click the consumer took—say, to search for a store—took all the glory.

This is a particularly hard problem in tying mobile ads to desktop transactions, because it's difficult to connect users' mobile activity to what they eventually do on desktop.

Lift will study the control group and the group exposed to ads to see if Facebook ads worked. Facebook is measuring the incremental increase in sales provided by its ads, so it is trying to highlight the effect that seeing an ad on the social network had on the end result, not the cumulative effect of all digital ads before the last click.

It does this by measuring the control group's tendency to buy a product versus the group exposed to the Facebook ad.

In one example, Land of Nod furniture sales rose 12 percent from its Facebook campaign. Smallwood said Facebook wants to give brands the ability to know if their ads are effective and spend accordingly, even if it means they find out the ads didn't work.

Facebook also thinks the new way of measuring impact of digital ads could lead the industry to shift more dollars online in general by showing their efficacy.

"We encourage the broader advertising industry to adopt Lift as the standard for measurement and believe it's the best method for proving the effectiveness of ads—whether they run on Facebook or someplace else," the company said in a statement today.

Advertisement
Advertisement
Adweek Blog Network