The days of ad buying's "dumb pipes" are numbered.

In its brief history, automated ad buying has spawned dozens -- if not hundreds -- of tech companies that built the sprawling passageways that have freed advertisers from the inefficiency of buying ads directly from publishers or ad networks. The need for those pipes has allowed those companies to raise billions of dollars through private fundraising and public offerings.

But with the pipes now in place -- and the relative ease with which they can be laid by others -- these ad-tech companies are seeing their power erode in favor of bigger players that own unique data.

"Data is the new oil," said Wells Fargo Securities senior analyst Peter Stabler in an interview. "Scale without data is becoming less valuable every day."

Today, programmatic ad buying mostly involves applying third-party data to large pools of inventory. But the emphasis on directly collected -- or first-party -- data is giving rise to a new group of digital-advertising companies and empowering marketers who can harness their own data and put it to work.

The latest and perhaps most formidable entrant is Facebook, which last month relaunched Atlas, an ad server built on its first-party data. In doing so, Facebook joined other global companies with massive troves of proprietary data that want a bigger chunk of the $50 billion digital-advertising industry, including Amazon, Walmart, Twitter, LinkedIn and Weather.

Each of these companies has either released or is producing ad platforms that will use their data to make ad buys across the web. Rakuten, a large Japanese e-commerce company, also introduced a U.S. ad business last month with the expectation that it will eventually use its first-party data to sell ads.
In addition to these giants, many marketers -- including Kraft, Netflix and Kellogg -- are using their own data in-house so they control their approach to automated ad buying.

By using their own data, marketers can reduce their dependence on ad-tech middlemen, which can charge 15% per transaction in many cases. For example, Kraft used its data as a negotiating chip to lower these rates, according to Bob Rupczynski, a VP leading the company's in-house data effort. "I think you'll see a correction," he said, speaking of ad-tech rates.

Kraft is also using the data to execute more effective automated ad buys, Mr. Rupczynski said. "There's only so much you can get out of the efficiency play, and buying cheap impressions doesn't build brands," he said. "You're going to see the industry make the transition."

At the same time, the new emphasis on first-party data is making marketers' lives more complicated. "They're going to have to become smarter about the data they collect," said Will Phung, director-media at M&C Saatchi Mobile. "Marketers now have to understand, 'O.K., I have this data, I have to find some way to translate it into an anonymous cookie that then can be matched in some sort of programmatic environment and shown an ad in the right brand-safe context.'"

Brian Lesser, CEO of Xaxis, WPP's programmatic unit, expressed a similar sentiment. "From an execution standpoint, it's getting harder, not easier." As for Facebook and others trying to leverage data, Mr. Phung said it's too soon to tell if they will dominate the digital ad industry. "It's not just having that data, it's also having some machine intelligence able to tease apart that data," he said.

In other words, first-party data isn't truly valuable without a layer that optimizes media buys based on it. "Weather data itself is not where our power lies. It is what we do with it that makes it truly interesting," said Vikram Somaya, general manager of WeatherFX, the Weather Co.'s data division.

Facebook does a great job with its data on its own site, but Mr. Phung said the question is whether it can do the same on the rest of the web. If it can match what it does in its news feed, "they're definitely going to get dollars not just from us but from agencies across the board," he said. Facebook has a running start on that front: its ad server can already tie together people's identities across screens, something the rest of the industry struggles with.

Up for grabs, according to Mr. Phung, could be the spending on a number of ad networks and programmatic platforms. "The opportunity is absolutely huge," he said.