What if all the ads you read online were written by robots, with copy tailored specifically to your tastes?
That’s the technology Culver City ad tech firm DataPop is releasing and what it thinks will be the future of online advertising.
The ongoing quest among online marketers is to make ads more relevant and therefore more likely to be clicked. There are already a variety of technologies that can send pre-made ads to individuals that are deemed ideal targets. Now, said Jason Lehmbeck, DataPop’s chief executive, his company is taking precision a step further by crunching all an advertiser’s data churning out the kind of keywords and writing that would appeal to specific consumers.
“There’s a lot of stuff that a machine can do better than a human, like being able to generate thousands of unique ads within the click of a button,” Lehmbeck said. “Our data machine knows how to write sentences as if a human wrote it.”
At the center of DataPop’s new play is “semantic technology,” which can examine the words on a page and determine its content, tone and intent. By analyzing the kinds of pages a person visits, the DataPop’s technology can pick out phrases that would appeal to that person’s habits.
For example, if a car maker has a handful of different models and financing options, DataPop’s program will try to figure out which combination of car and price will appeal to a specific consumer. Then it will write up a text ad with language that it determines will have the greatest impact.
There are, of course, some challenges to having algorithms churn out writing intended for humans. Some newspapers have tried to outsource routine article writing to machines with varying results (no comment here). Lacking editorial input can result in writing that’s garbled, spam-like or just slightly off.
Lehmbeck said he’s well aware of those challenges, but said writing ad copy is an easier task.
“Unlike the article writing, you’re dealing with a high volume of data with advertising,” Lehmbeck said. “We’re seeing the impressions and click rate training our data set to get smarter.”