The demand to measure the return on investment (ROI) for marketing spending accurately has never been greater. Media channels continue to proliferate, and advertisers are being forced to sharpen their focus among an increasing number of options drawing from budgets already under pressure. Big data holds the keys to this kingdom, but harnessing and utilizing an overabundance of quality data has not historically been an easy feat.

Single-source databases are really the only way to tackle the challenge of proving ROI because they allow companies to match who is watching and not watching their ads, and who is buying and not buying their goods. Unfortunately, such databases have generally been constrained by low sample size, have offered limited purchase detail, were very hard to build, and were ultimately too expensive to be worthwhile.

Consequently, people have relied on other alternatives to assess the sales and profit generated by advertising campaigns. These tools have been essential in improving marketing and media planning. But they share at least one shortcoming: Because they can’t match any given individual’s degree of exposure to a campaign with purchase behavior, the marketers are forced to estimate ROI.

Today, however, big data—in the form of sufficient information that can be processed at an affordable cost—has addressed the problems that stymied single-source data in the past, enabling advertisers to know precisely the highest return programming for their advertising and publishers to know exactly which advertisers are best suited to sponsor their content.