Algorithms and Targeted Marketing, Part 1

Stephen DeAngelis

September 24, 2012

In a review of the book Automate This, Richard Waters writes, “Christopher Steiner makes a big claim for the mathematical formulas that underpin technology. ‘The ability to create algorithms that imitate, better, and eventually replace humans is the paramount skill of the next 100 years,’ he says. ‘As the people who can do this multiply, jobs will disappear, lives will change, and industries will be reborn.’ He may turn out to be right, though his subtitle, ‘How algorithms came to rule our world’, sounds decidedly premature.” [“An algorithmic future is almost here,” Financial Times, 29 August 2012] Even if Steiner’s claims sound a bit hyperbolic, the headline of Waters’ review is correct that algorithms are playing an increasingly large role in our lives. To read more of Steiner’s views on algorithms and the future, read my posts entitled Computers Get Creative, Part 1 and Part 2.

Waters is obviously skeptical about some of the claims that Steiner makes. “In the field of artificial intelligence,” he writes, “it has always paid to take the most expansive machine-replacing-man predictions with a large pinch of salt.” He admits, however, that history may indeed change thanks to two things that “have made the difference.” He explains:

“One is the rapid decline in the costs of computing and communications, making it economic to deploy programs that use heavy-duty num­ber-crunching in many more fields. Steiner cites a New York Times report that says combing through more than 1m documents in a legal discovery process can now be done for $100,000 using a computer, compared with the $5m it would have cost with paralegals on the case. The other difference is the dawning of ‘big data’ – huge data sets on which to ‘train’ algorithms, ref­ining their ability to find patterns and make deductions from a sea of information.”

I find it interesting that Waters writes about “training” algorithms to carry out desired processes. That makes it sound like a man with whip is standing over a computer cracking it loudly every time the algorithm makes a mistake. Most people in the field talk about algorithms “learning” as they gather and analyze data rather than being trained. Regardless, I agree with Waters that decreased costs of computing power and access to big data do create opportunities that have never before existed.

Waters reports that Steiner is “a for­mer reporter for Forbes magazine and now co-founder of an internet grocery company.” That means that Steiner is well aware of dichotomy that grocers face as taste differences between old and young grow wider. For more on that subject, read my post entitled Changing Tastes in Food are Challenging Food Providers. Identifying differences in tastes is one challenge that can be taken on by algorithms because they are superb at pattern identification.

Clearly the marketing campaigns necessary to reach customers with different tastes must themselves be different. In a previous post entitled New Directions for Retailing, Part 2, I cited an article by Jon Swartz [“Why shopping will never be the same,” USA Today, 9 August 2012]. For that article, Swartz interviewed David Fisch, director of platform partnerships at Facebook, who told him, “The first 15 years of online shopping was about making it easier for people to find and purchase items they were looking for. Now, it’s about helping you find what you may not know about, based on your social (media profile).” I also pointed out in that post that Marianne Timmons, a Senior Principal at Accenture, told attendees at the Grocery Manufacturers Association (GMA) Executive Conference in Colorado Springs, CO, that less than five percent of the $200 billion spent on consumer packaged goods (CPG) marketing is targeted the way that Fisch suggests. My company, Enterra Solutions, is working with clients to increase targeted marketing so their marketing money is better spent.

In fact, Targeted marketing is likely to be the next big thing in retailing. For example, Tim Bradshaw reports, “Twitter and Facebook have introduced new capabilities to allow advertisers to more finely target their users, as the two social media sites look for new ways to extract revenue from their audiences.” [“Twitter ads to target interests,” Financial Times, 30 August 2012] Reporting on the same story, Shira Ovide writes, “Twitter said it will let advertisers push their marketing messages to users based on indications of what they like. A company that sells a sports drink, for example, could elect to show paid ads to Twitter users who are fans of professional football.” [“Twitter to Target Ads Based on Interests,” Wall Street Journal, 30 August 2012] Bradshaw continues:

“This ‘interest targeting’, billed as the largest upgrade to its ad technology in a year, is Twitter’s response to Facebook’s longstanding ability to allow advertisers to tap what members of the social network say they like and do online. At the same time, Facebook revealed that it will now allow its larger advertisers to use data they have collected about their customers, including email addresses and phone numbers, to target ads on the social network. ‘Ultimately we want to deliver more relevant advertising to people who already have relationships with those people,’ Facebook said. The development adds an additional layer of ad targeting beyond what it currently sells to advertisers, which includes targeting based on the age, location and interests indicated on a user’s profile. This ‘custom audiences’ service represents a major development in Facebook allowing brands to include data from outside Facebook into its targeting mechanism, in a way that individuals might not have expected when they originally handed over their information to an advertiser.”

As I wrote in my post about new directions in retailing, “The elephant in the room when it comes to targeted marketing is privacy. Journalist Charles Duhigg told Swartz, ‘There is a trade-off between privacy and convenience, which I think will only accelerate. People always choose convenience and don’t realize the cost of privacy.'” Bradshaw agrees that privacy remains an issue. He explains:

“A process of anonymisation, known as hashing, prevents both Facebook and each advertiser from exchanging more information than the email address. Users can opt out of advertisements as they appear, one advertiser at a time, but not prevent the targeting entirely. The move further blurs the line between the tracking of consumer behaviour online and in the physical world. In its attempts to help marketers understand the effectiveness of their ad campaigns, Facebook recently started working with outside data companies, such as Datalogix, to track hundreds of details about consumers, ranging from the products they buy at brick-and-mortar stores to the cars they drive. The development is likely to raise concerns among privacy advocates. … Twitter’s new targeting capabilities also risk irking some users, who are not able to see or manage what topics the site thinks they are interested in.”

Although consumers will likely have privacy concerns, they must also recognize that the benefits of social sites like Facebook and Twitter will only be available if the companies are financially successful. The financial future of both companies rests on finding a viable advertising model that generates revenue. Targeted marketing represents their best hope. Ovide reports:

“Twitter’s appeal to advertisers is crucial for the company to prove it has staying power as a big business. The San Francisco company has more than 140 million active monthly users but has been slower than Facebook Inc. in showing it can translate those fans into ad dollars. Some advertisers have said it is difficult to know if their Twitter ad messages are reaching the right people. And Twitter executives have said they are focused on tinkering with the signals used to ensure advertisers show their messages to the most interested Twitter users.”

Despite consumer concerns, Bradshaw reports, “Both innovations … are likely to be popular with marketers who want to improve the effectiveness of their spending on the two sites. Social media remains an unproven format for many, putting pressure on Twitter and Facebook to find new ways to advertise to their users.” Since most marketing is currently untargeted, the potential upside for Facebook and Twitter is significant. I agree with Bradshaw that targeted marketing is going to be popular with marketers because the return on investment is likely to be much higher than it is with current marketing efforts. Tomorrow I’ll discuss what Gary Hawkins, CEO of Hawkins Strategic, has to say on the subject he calls “personalized marketing.”