Oracle Advertising

» see oracle.com/advertising as it lived, on the wayback machine

~ the eulogy ~

Oracle Advertising was never a website you visited. It was a layer underneath other people's advertising: a set of audience segments, an identity graph, and a marketplace where a brand could buy the right to show its ad to people who had recently looked at refrigerators. The raw material was behavior. Where you browsed, what you bought in a shop with a loyalty card attached, which categories of thing you kept drifting back toward.

Oracle bought its way in rather than building. BlueKai, the data management platform that ran one of the largest third-party cookie pools on the web, was acquired in February 2014. Datalogix followed that December, bringing a file of offline purchase records built up since 2002, and that acquisition is where the unit got its name: Oracle Data Cloud. Moat arrived in 2017 to handle measurement and Grapeshot in 2018 to read the page an ad was about to land on. The press put the decade of buying somewhere north of four billion dollars. The division was renamed Oracle Advertising on the way into the 2020s, which was the last confident thing that happened to it.

At the top, around 2022, it was turning over roughly two billion dollars a year. Two years later Oracle shut it off. Everything that made the machine work had become a liability in the same window. GDPR arrived in 2018 and a wave of American state privacy laws followed it. Apple decided in 2021 that apps had to ask permission before tracking anyone, and Google spent those years publicly dismantling the third-party cookie. The supply of data got narrower and the legal exposure got wider.

There is no honest way for me to tell you whether I used this. Nobody used it. The only real question is whether my browser passed through it, and I have no way of finding out, which was the entire design of the product.

~ the dossier ~

~ the sentence that ended it ~

It died on an earnings call. Oracle reported its fiscal fourth quarter in June 2024 and chief executive Safra Catz put it in a single line: in Q4, we decided to exit the advertising business, which had declined to about $300 million in revenue in fiscal year 2024. A business that had been worth two billion dollars two years earlier was now worth a rounding error against Oracle's cloud numbers, and it was gone in a subordinate clause.

The customer-facing end came on 30 September 2024, when the DLX and BlueKai platforms were turned off. Agencies that had been buying Oracle segments for a decade spent that summer being told by consultancies which replacement vendor to move their audience targeting to. AddThis, the share-button widget Oracle had bought in 2016 and used as a data faucet on fifteen million websites, had already been switched off the year before. September was the last step of a decision Oracle had been taking in pieces for several years.

~ the one time it addressed you directly ~

In August 2022 Michael Katz-Lacabe and Jennifer Golbeck sued Oracle America in the Northern District of California, arguing that the company had captured and sold the personal information of hundreds of millions of people who had never agreed to any of it. In July 2024, weeks after Catz announced the exit, Oracle agreed to pay $115 million to settle. It admitted nothing. It also agreed to stop collecting certain user-generated data from referrer URLs and to run an audit program against its own privacy obligations.

The settlement class was estimated at around 220 million American residents, which is most of the adult population of the country. That is the one moment this product ever spoke to ordinary people, and it spoke in the form of a claims website. The class was residents of the United States, so everyone else whose browsing passed through the same pipes got the shutdown without the notice.

~ what it leaves ~

Oracle did not lose the advertising business to a better advertising business. It lost it to the legal and technical environment that advertising had to live in after 2018, and then it decided that a three hundred million dollar line item was not worth defending in court every year. The data brokers that remain are smaller and quieter, and the targeting money went where the data never has to leave the building, to Google and to Meta, each of which owns the place you are standing in when it watches you.

Four billion dollars of acquisitions, ten years, and the asset turned out to be a legal problem wearing a revenue line.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »