The Big Data-Driven Business

Chapter 10Measurement, Testing, and Attribution

It's an ancient complaint that it is difficult, if not impossible, to tell if advertising works—even though advertising has always been measured in some fashion. When ads ran, clients might offer their opinions on the creative side: thumbs up or thumbs down. Maybe the phone rang a little more often. There might be more sales opportunities—maybe, maybe not. And in some cases, revenue increased.

Historically, large marketers would attempt to advertise in single geographic locations and measure lift in revenue per store or similar metrics to see if the advertising was creating any measurable difference in a market where the population was exposed to the ads, compared to other markets where there was no exposure. But unless the test was very well designed, it was still often hard to tell if it was the advertising that was actually causing these results.

But now, with advanced uses of data, marketers and their advertising agency counterparts can tie marketing and advertising to revenue more tightly and reliably than ever before. With the right data, the right technology, and the right approach, marketers can test various initiatives, measure responses, and focus efforts on those initiatives that contribute directly to increased revenue.

Data has also enabled a more sophisticated and holistic view of how marketing works. Different marketing initiatives have different objectives. A branding campaign, for instance, generally delivers different ...