Is your programmatic buying really working? Can you prove it?

Most marketing professionals don’t realize it, but this hot, buzz-y technology too often overpromises and under-delivers.And that’s a big deal, because programmatic is an essential marketing tool that’s here to stay— not only due to its value in processing data, but because it’s so much easier for marketers and agencies than the old, manual way.

This leads to an important question for every marketer today: How can your company get on top of the programmatic juggernaut and get the most impact for your marketing spend?

The answer can be summed up in two words: manumatic and attribution.

First, some history. About 10 years ago, programmatic buying—which employs algorithms and automation to target an audience and direct ad spend—became available to replace the manual process, which required painstaking site selection, insertion orders, and collation of results by hand. Marketers embraced programmatic enthusiastically. Today, three quarters of digital display ad dollars in the United States transact programmatically, and it’s estimated that figure that will rise to 84% by 2019.

But there’s trouble in programmatic paradise. While providing important benefits, programmatic buying also takes away control. Marketers set parameters about whom they want to reach and where they want to be, but they don’t know whether those parameters are actually being adhered to.

Too often, programmatic buying is a spray-and-pray system. Sure, maybe you appeared on every newspaper site in the country. But maybe you were supposed to be in the online finance section, and ended up in the sports section, where the ad lost context. Maybe you got 3,000 calls or 10 million impressions, but people are calling about the wrong product—a low-margin one—or the people who are seeing the ad are in the wrong demographic groups.

Or worse. Since networks get paid for impressions, they want to be in as many sites as they can, so inevitably rogue sites creep in. I’ve seen blue-chip companies appear on sites that just weren’t proper for them to be on. But they don’t know it, because they are only getting information in the aggregate. Moreover, when an ad campaign succeeds, marketers don’t know exactly why. How many leads or calls came from Google vs. Bing vs. Facebook, from what geography, during what times of day? What impact did social media have?

When I ask companies those questions, the usually that don’t know the answers. In that case, how can they know when to spend more or less on any of those tactics?

Getting Under the Hood

Those of you who are auto enthusiasts, as I am, will recognize the term “manumatic.” Manumatic is a process that combines the best of human intervention and technological power.

If you rely fully on programmatic buying, your campaign is only as good as the person who created the algorithms that power the software. With a manumatic approach, a programmatic system will alert a human being—in real time—when results are falling significantly outside expected norms. Then, that person can make a judgment call regarding whether things are going really wrong or really right.

Here’s an example: My company was handling the ad spend for a major telecom, and Super Bowl Sunday came along. We were expecting a somewhat higher-than-normal volume of calls and clicks for that day. However, our system alerted us that we were getting an extremely high call count — as much as three or four times what was normal, particularly in the Southern California area.

In an all-programmatic system, the response would have been for the software to assume that more calls are good, and more money should be pushed to that geography. But when a human being looked at the keywords people were typing in, the geography that was lighting up on the map, and the content of customer service calls, we learned that the issue wasn’t increased demand — it was that the telecom had had an outage in Southern California on Super Bowl Sunday.

Because we were using a manumatic system, we were able to override the programmatic system and pause all the acquisition campaigns in that location. This saved the telecom about $33,000 in one hour, which could then be redirected toward more productive ad spend.

Perhaps it’s no surprise that teaming the human brain with programmatic technology leads to smarter decisions about marketing. Even so, those decisions can’t be optimized without the other leg of the stool—namely, full attribution.

Over the long term, true full-channel attribution can follow an individual buyer from his or her first exposure to a topic, product or service, all the way across IP addresses, domains and devices, to the point of conversion. It can tell you about the most effective sites, geographies, times and copy for your ads. You will know what tactics are actually succeeding. (For more information on how true full-channel attribution works, see “What You Don’t Know About Attribution is Costing You.”)

Ideally, your programmatic buying will show you all of this in real time — not with a model, not with a spreadsheet, and not three months after the fact.

And when that happens, you’ll have a programmatic buying system that really lives up to its promises.

Top takeaways:

  • A programmatic system should alert a human when results are falling outside expected norms.
  • Without full-channel attribution, marketers don’t know which parts of their campaign are working.
  • In order to maximize ad spend, your programmatic platform should show you, in real-time, where your ads are appearing and how they’re performing.

Sergio Alvarez is CEO and founder of Ai Media Group, a New York City – based media company that specializes in defining, managing, and executing online marketing strategies, US Programmatic Digital Display Ad Spending, 2015-2019