The Exchange Lab - Powered By Proteus

Menu
News / Events

Chris Dobson Explains How to be Fleet of Foot in Programmatic

June 10 2014

On why automation is important

In some respects, I think agencies need to go programmatic because they can’t cope manually. Generally, in a pre-programmatic world, global deals were a nightmare.

The beauty of automation is that you can make calls market by market, almost district by district. The more technology allows you to do that, the more it starts to achieve things you could never have planned if you were trying to do it manually. Our view is we still absolutely need human intuition and intelligence in the system — but you’ve got to let the machine make those humans super-human. You can’t just keep on adding traders; the machine’s got to make one of those traders do the work of 10 efficiently.

On global media deals

The bigger agencies do big deals with global players like Facebook. You saw Omnicom did a deal with Twitter. I was just a bit confused about that, because if you’re in the programmatic space what you want is to buy the right audience, at the right time, where that audience is and how that audience is behaving — which is going to be different depending on who the audience is. And of course, the price will be determined by what you’re trying to achieve, as well as the time of day and all sorts of stuff. So how do you do a $230 million guaranteed-price deal with Twitter, and then put that into a programmatic space?

We really believe that being agnostic in the marketplace is key to success. The old-world way of doing big deals that channel customers’ money to a particular media owner (albeit for a good price) doesn’t fit the world of programmatic, where media is bought component by component in real time.

On brand safety

Brand safety is the other thing I don’t think the programmatic industry has cared enough about — but it is absolutely top of the agenda for CMOs. They are the custodians of multibillion-dollar brands, they’ve spent decades building their relationship with customers. You can’t allow apps and algorithms to throw it away.

It’s not only brands, it’s publishers too. Because publishers face as much of a risk of putting inappropriate ads in front of the wrong audiences on their sites.

At BBC we were anal about it. We would have people who actually clicked through an ad to see the website it went to, and then see whether the claims in the ad were supportable. If they said they were number one, and they weren’t, we wouldn’t carry the ad. We ended up with these really clunky, manual pre-approvals on a monthly basis — which doesn’t scale. So we’ve got to figure that out.

I think what programmatic needs to realize is that environment still counts. The brand juxtaposition is not just about finding a BBC.com audience on some other site. CMOs still want to have a BBC.com audience on BBC.com, because there’s an appropriateness around the content and the ad placement.

On using multiple tech platforms

The thing that we’ve learned by taking a very wide view of the market is that it really is “one size fits one.” Depending on where you are in the world and who the audience is, it takes myriad different solutions to deliver whatever the KPI is for that client.

When we plan campaigns, we always start with multiple DSPs and multiple tactics, to test and then refine as the campaign goes on. We pull down the CPA, or the CPC, or whatever metric. No two are ever the same. If you go to Asia, for instance, you’re going to be all over the mobile. There’ll be hardly any display.

You’ve got to have what we believe in and what we’re building: a trading platform that allows you to plug everything in and change it as the universe evolves. Because it’s going to evolve — every day, every week, every month. Nobody can pick a winner yet. So why bother?

You need to be able to dip in to see who’s holding the best inventory, in the same way that an asset manager would pick different funds to create a portfolio that performs to its maximum. But that’s pretty tough if you’re only dealing with one or two funds, which is what the agency trade desks tend to have — one big DSP partner, and maybe another one tacked on. Whereas we’ve got eight or nine, and it keeps everybody honest in some respects.

On the challenges agencies face

I had a conversation just last week with a major bank that was not sure how their first party data was being used. And this discomfort goes back to when media agencies first started in the late 80s — the big clients in a media agency that had the clout were worried they were driving value for the small clients that rode on their coattails. And it was true.

In the same way now, some of the clients that I’ve spoken to are worried that their data – which is absolutely the lifeblood of their business – might not be used solely for their benefit. It’s creating a picture that’s useful for buys across other sectors or other clients. How effective are those supposed Chinese walls?

Then you have conflicts within agencies, especially big agency groups where you’ve got individual agencies that don’t want to be directed to use a trade desk that they’re told to. They want to play a wider market. Some do, and others do what they’re told. It’s a three-line whip. So there’s quite a lot of tension in big agency groups.

I don’t think you can call what’s going to happen in the next five years. If you’re a client, you’ve got to try and remain uncommitted and fleet of foot, so that you can be bending with the flow as things change.