Turn taught ads to think. Buyers had other ideas.

In 2009, Bill Demas gave the people at Turn a speech he later called “burn the boat.” The company would change direction. Selling advertising had brought it this far; software would take it further. It was an appealing declaration for a Silicon Valley company. Boats, however, are useful things. A few years later, Turn would discover that some customers still preferred to travel in one.

The boat they burned

Turn began with people who understood search. Co-founders Jim Barnett and John Ellis had worked at AltaVista. Their advertising business, founded in 2004, tried to bring automated targeting to a market where buying an audience could involve considerable guesswork. Its early network used action-based bidding: advertisers identified an outcome they wanted, and the system worked on finding likely responders.

That first business did not grow smoothly. In Demas’s account to Inc., sales flattened after the early years. He became chief executive in 2009 convinced that customers should be able to bring their own data into a platform, analyze it and run campaigns themselves. The customer would gain more control. Turn would become the machinery behind the decision.

A shorter trip from data to an ad

Consider an illustrative car campaign. A marketing team has website visitors, customer records and information bought from outside providers. An agency has advertising budgets and access to media. The practical difficulty is getting the audience description from the first group to change what the second group buys.

Turn connected those jobs. Its data management platform, launched in 2011, organized audience information, including online and offline inputs. Its demand-side platform automated purchases through advertising exchanges. Analytics helped marketers examine the results and adjust the next round. The Marketing Platform and Digital Hub framed these capabilities as a connected working environment.

This was enterprise infrastructure for marketing teams, agencies and trading desks. Singtel’s 2017 acquisition announcement named KraftHeinz, Toyota and L’Oréal USA as Turn customers. A Nielsen relationship added audience measurement; an ecosystem of partners supplied complementary data and tools. Turn’s expertise sat in making those ingredients usable together.

Its position was on the buyer’s side of advertising. Historical alternatives included MediaMath, DataXu and Google’s buying tools. Turn emphasized an independent platform and the connection between data and execution. Independence was its market position, rather than proof that its results would always beat a competitor’s.

The distinction between the two products mattered. A DMP answered questions about audiences; a DSP executed media purchases. Turn’s audience product could also support analysis involving another buying platform. Bringing the jobs together offered convenience, while keeping connections open gave agencies room to work with a wider set of tools.

The buyer kept the old receipt

In January 2014, Turn raised $80 million. The money supported a larger ambition, including international expansion. Yet a platform’s usefulness and its commercial packaging are separate experiments. A buyer can want the capability while resisting the contract that comes with it.

By late 2015, new CEO Bruce Falck was shifting emphasis away from monthly software licensing toward familiar media-linked arrangements. AdExchanger reported 57 layoffs from roughly 400 employees. Many clients preferred buying media through established processes. Falck’s diagnosis was that the market was not ready for Turn’s SaaS approach.

“There’s no need to reinvent the pricing model.”Bruce Falck, Turn CEO, December 2015

What did it cost? There was no single consumer-style sticker price. Historical reporting described software fees, managed services and percentage-of-media contracts. A December 2015 account put Turn’s technology fee at 13-14%, against a cited competitor’s 15%. That narrow difference had not persuaded enough buyers to change habits. These were reported historical terms, not a universal rate card.

For a team evaluating a similar platform, the useful calculation extends beyond the technology fee. Someone must prepare the data, operate campaigns and decide what counts as success. A seemingly cheaper contract can still ask the organization to change its workflow. That is a cost worth putting on the same page as the quoted price.

The person outside the dashboard

There was another limit to the system’s promise of better information. In January 2015, ProPublica reported that Turn used Verizon identifiers to restore deleted tracking cookies. Turn subsequently said it would suspend that practice. An advertiser might value a more persistent audience identifier. The person who deleted the cookie had expressed a different preference.

The FTC later alleged that Turn misrepresented how consumers could restrict tracking and how its opt-out worked for mobile applications. Its final 2017 consent order required an effective opt-out and prohibited misleading representations. This episode exposes a product problem beyond campaign performance: controls need to produce the behavior people reasonably expect.

Three names, two very different deals

Amobee agreed to acquire Turn in February 2017 at an enterprise value of $310 million. Singtel’s subsequent financial statements recorded $290 million in consideration after working-capital and net-debt adjustments. The announcement and accounting figures describe different measures of the same acquisition.

In 2022, Tremor International acquired the later Amobee business in a transaction announced at $239 million. Core products and platforms adopted the Nexxen brand in 2023. Amobee’s website now points visitors there. Turn’s standalone chapter had ended years earlier; later parent-company figures should stay attached to their own company.

Copy the loop. Study the purchase order.

The useful idea to borrow is the short path between learning and action. Organize audience information, make a buying decision, measure the result, then revise. That approach needs usable data, sufficient media activity and people equipped to interpret outcomes. With thin signals, little buying volume or weak measurement, connecting the tools cannot guarantee better decisions.

The commercial lesson is equally portable: learn how customers authorize spending before asking them to adopt a new payment model. Turn built technology that a larger advertising business wanted to own. Along the way, it showed how much an apparently technical product depends on ordinary human arrangements: a purchasing process, a permission, a promise someone expects you to keep.