Skip to Content



Driving Marketing Results with Big Data

In partnership withDigilant

For marketers trying to maximize their return on investment, predictive analytics based on big data is an exciting new tool.

In the digital world, predictive analytics based on big data holds the promise of creating a detailed view of what works, providing guidance that has never been available before for the fine tuning of advertising campaigns.

The promise of big data analytics is that marketers can analyze thousands of points of information about the digital activity of the purchaser—stripped of personally identifiable information—and combine it with their knowledge of television, radio, billboard, and print campaigns to tailor marketing messages and, ultimately, improve return on investment (ROI). With analysis, the numbers show how much lift each data point provided for each ad in each channel. With that data, marketers can make better decisions about how to allocate their ad budgets. Indeed, the analytics themselves will identify the smart choices.

A key challenge for any marketer is deciding what mix of media—TV, Internet, direct mail, radio, print—will best promote a product or service. “We can do media-mix modeling using big data and machine learning,” says Madan Bharadwaj, product marketing chief of Visual IQ, an analytics firm based in Needham, Massachusetts. “There are a lot of micro-efficiencies we can tap into. If you move a few thousand dollars here and there, you can get much more marketing efficiency,” in terms of ROI.

Historically, the most sophisticated marketers have relied on top-down ad campaign planning. They develop econometric models by looking at the distribution of the whole advertising budget. They analyze changes in allocation and one-time promotions and see how those changes affect their key performance indicators (KPIs), which may be making an in-store purchase, or opening a new account.

That paradigm is flipped in the digital world. Marketers rely on digital scoring of actions, starting from the bottom up with the KPI. “You try to work backwards to see the touch points along the consumers digital journey,” says Kim Riedell, senior vice president of product and marketing at Digilant, a customized programmatic media solutions company in Boston. Thanks to technologies such as cookies and browser pixels, marketers can now tell exactly where a specific buyer saw their ads. The data even shows how long that buyer watched a video or lingered on a page carrying the ad. It’s all found by backwards tracking from the point of sale of the product the person ultimately bought.

The world of perfect knowledge that was promised in the early days of digital advertising has proved illusory. Paying search engines for stimulating clicks that led to purchases was fine, but most consumers take a more circuitous route to their final decisions. The marketing funnel can be long, especially for big purchases such as automobiles, where people may do research for nine months before taking a test drive.

Advanced predictive analytics can now figure out what audiences have been most responsive to an ad. Then the same algorithms can find similar audiences on other websites and present the ads to them. With enough data, and a good algorithm, the analytics companies say they can determine just which ads made a difference.

Predictive analytics can’t incorporate everything. A favorable product review in Consumer Reports or a celebrity endorsement at the Oscars falls outside the algorithm. So does a plane crash that may hurt travel bookings. Sometimes, though, such events will cause a spike in discussion on social media, here they are monitored and even calculated into the equation.

Immediacy is one of the big benefits of analytics. Rather than waiting to see weekly sales results, and tweaking ad strategies in response, marketers can see online outcomes daily. With online KPIs, analytics firms can analyze ad performance in real time and recommend the most effective follow-up marketing the next day.

Jeff Zwelling, cofounder and CEO of AOL’s Convertro, an advertising analytics firm based in Santa Monica, California, says that his company sends advertisers reports on the performance of ads within 24 hours. “The day after the Super Bowl, we had reports to GoDaddy, Intuit, the NFL, by 8 A.M.”

Analytics can also help protect advertisers from fraud risks without forcing them to build their own investigative efforts. So long as the KPI is something that can’t be mimicked by bots—a purchase, for example—phony ads won’t show up as being effective. “All that is needed is a top-notch attribution solution and a pricing structure that aligns advertising performance to actual conversions (rather than clicks or impressions) to avoid fraud,” David Perez, Convertro’s chief marketing officer, wrote in a recent blog post.

In theory, the algorithms should be able to allocate budget to advertising networks that police their inventory to avoid phony ads. They should generate more key-performance indicators. By the same token, ads that aren’t viewable won’t drive KPIs. It isn’t clear whether that promise is being fulfilled.

Similarly, predictive analytics may discover correlations among categories of potential buyers that would be unlikely to occur to human marketers. For example, Digilant chief scientist Krishna Boppana recalls that while working for a financial services client, his company discovered that men who had been looking at boats in April and May were responsive to ads for 401(k) plans. He now hypothesizes that they might have received bonuses in April and planned to use the money for fun before remembering—or being reminded—that they should be funding their retirement accounts.

While working with a luxury cosmetics company, Boppana adds, Digilant discovered a correlation between women who were interested in exotic travel and those who bought Kashi cereal. While predictive analytics is often criticized for spotting correlation rather than causation, Boppana concludes that “advertising is all about correlation.”

Deep Dive


Capitalizing on machine learning with collaborative, structured enterprise tooling teams

Machine learning advances require an evolution of processes, tooling, and operations.

The race to destroy PFAS, the forever chemicals 

Scientists are showing these damaging compounds can be beat.

How scientists are being squeezed to take sides in the conflict between Israel and Palestine

Tensions over the war are flaring on social media—with real-life ramifications.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.