Pure Play vs. Full Stack: GroupM Digital Debate

Posted by Christian Manie On July 21, 2016

On July 20 Ashton Media hosted the GroupM Digital Debate at the Doltone House in Sydney to argue the merits of full-stack technology versus pure-play technology in regard to driving better results for advertisers. And first of all, when it comes to a field that moves as fast as programmatic advertising, a debate is a perfect way to go. As an industry we should hope to see more debates like this one, presenting both sides of a complex proposition and letting the audience decide.

The debate was chaired by Tim Whitfield of GroupM who first set about explaining the agreed upon definitions of full stack and pure play as the debaters had decided. Arguing for pure play were Sam Smith, managing director of TubeMogul, and Peter Hunter, commercial director at Quantcast. On the side of full stack we had Rhys Williams, head of media tech at DoubleClick, and James Diamond, managing director of Integral Ad Science.

Pure Play:

Smith opened the debate by clarifying that it was a debate about working with a generalist or a specialist, and that in his view, pure buy-side tech is built to drive the best result for advertisers.

“How on earth is it possible for a full tech stack to provide the highest return for a publisher and at the same time try and provide the best outcome for an advertiser?” Smith said. Pure buy-side “doesn’t own or preference any inventory.” He made the important point that pure buy-side welcomes third-party tech to verify performance – whether an ad has been seen or delivered to a targeted audience.

Smith also claimed that innovation is dependent on collaboration and full stack was limiting, primarily by stifling competition by limiting access to important industry resources and segregating data, which is not in the best interest of advertisers.

Quantcast’s Hunter followed up the pure play push by using an analogy in the construction industry, driving home the benefits of using a specialist over a generalist. Hunter feels that on the pure-play side the primary job is to drive the best result for the clients and it is about getting the right message to the right customer at the right time, not about how to drive down CPMs. Inventory is a non-issue, Hunter said. There is an endless supply of inventory, for both the buy- and sell-side, so it is more about “getting to the right inventory and putting the right message in front of the right person,” which, in his view, is something better achieved by a pure-play specialist.

In finishing his argument, he said it’s about the impossibility of serving two masters, and on the pure play side they serve only the client.

Full stack:

Williams from DoubleClick kicked off the full stack argument saying the debate wasn’t about the tech itself, but about the result it can deliver for the advertisers and marketers.

When considering the tech Lumascape, he said, “just understanding how those disparate and unconnected solutions compare is one thing, getting them to work together is a whole other thing. And it’s in this context that the value of the full stack becomes plainly obvious.”

Introducing the analogy of your marketing as a car, he said there are lots of components required to make that car work, but pure play will only provide you with one part. It’s up to the marketer to assemble of the various parts and make them work. But what if something changes – be it tech or a company goes out of business – and the marketer needs to start again? Williams says that’s not a problem when working with a full stack, as it just works as an end-to-end solution, allowing the marketers to focus on marketing.

He claimed a company loses about 10% accuracy every time you move an audience from a DNP, a SSP and a DSP, therefore reducing your reach. He quoted a BCG study which concluded moving to a full stack can generate a 33% productivity improvement across a campaign lifecycle. But most importantly, he said, a full stack gives marketers the ability to generate a single view of the customer.

Diamond from Integral Ad Science said that in every industry there are companies that start as specialists “and then over time, through mergers and through acquisitions, they evolve into companies that offer more of a full stack, companies that offer a more holistic set of unified technologies rather than just one.”

When discussing full stack, Diamond said it is important to emphasize the open stack – the industry needs to be standardized, a place where specialist can just plug into full stacks. But you don’t reach that openness by arguing the need for a greater number of more diverse specialists. He said that there are some 1900 hundred companies on the Lumascape, and in a world with less time to address tougher goals, marketers don’t need to be hanging out with those “nineteen-hundred tech providers, trying to get everything to work together.”

“We need fewer platforms with datasets that work together. We don’t need more logins, we don’t need more data from different platforms that don’t work together …” Diamond said.

In summation, both Rhys and Diamond emphasised that a full stack doesn’t equate to a closed stack. Most full stacks integrate with companies that have nothing to do with them but add value to the stack. Diamond also conceded that that many full stacks do need to do a better job integrating.


The audience was asked before and after the debate to vote in favour or pure play or full stack, and given the option of abstaining.

Pre-debate vote:
–          20% Pure Play
–          10% Full Stack
–          70% Abstained

Post-debate vote:
–          14% Pure Play
–          71% Full Stack
–          15% Abstained

So it seems clear, at least on the back of the debate, that there was a clear preference amongst the audience for full-stack tech, likely on the arguments of saving time, increasing efficiency and generating something more closely approximating a single view of the customer. But to say that full stack is the better choice would be wrong, negating the important role that pure play has in the industry, it’s relevance to many advertisers and the role it has played will continue to play in a more unified theory of open stack.

In the end it is up to the customer to decide what works best for the campaign: You need to take the best parts and the best tech that work in getting the right message to the right customer. It’s not a “one size fits all” solution. 

Christian Manie