With unwavering digital-only focus, Anthony Hitchings, head of advertising technology and digital operations at FT.com, has had a front-row seat to the promises and pitfalls of programmatic media from within a large newspaper publisher’s online unit.
“I see the programmatic space as providing a great opportunity for us. But there are ‘teething troubles,’ if you like, with a large number of supply-side platforms (SSPs) that have been built up,” said Hitchings. “And there are obviously some key demand-side platforms (DSPs) and the two haven’t learnt to communicate effectively together yet.”
“DSPs started out with the basics of trying to find users based on advertiser data and they haven’t realized the full potential yet of what you could do with publisher data and actually using programmatic for brand.”
For example, Hitchings said the ad slots on FT.com have the flexibility to be many sizes, and yet in programmatic, the market can only allow for one ad size at a time. What could be a half-page takeover-type ad quite often ends up being a standard 300 by 250 banner. He added that not all the DSPs and SSPs are providing the opportunity to be as flexible as they could be.
AdExchanger: Are you thinking about yield holistically – in other words, it’s not just about advertising yield? It might be ads and commerce — beyond subscription, for example.
ANTHONY HITCHINGS: Yes. One of the tools we’ve been working with quite significantly in last 12 to 18 months is Metamarkets, which is really for our own internal reporting. We can report all data points such as volume, revenue, CPM figures, as well as an aggregate view for the broader level, demographic, content-type targeting. We’ve made sure that we’ve got robust, internal reporting systems in place.
Then in terms of our holistic approach, if we’re going to work toward a metric such as conversions in CPA deals, which is something we’re playing with now, at that level we do see programmatic eventually being able to compete against our guaranteed, which uses a standard CPM.
Do you worry about cannibalizing your direct sales with some of your digital yield strategies?
Not particularly. We’ve got a structure in place that if somebody is spending money with the FT we would then approach them to spend programmatically. If they’re spending direct with us, we don’t really mind the delivery channel.
What we’re doing is carefully managing the entry point and the floor prices to ensure that our best customers get the best level of service whether it’s programmatic or direct.
Moving on to your data-management platform (DMP) strategy, what were the triggers for you and FT.com to go and find a DMP?
Data has always been at the forefront of our packages that we take to market. When we put the paywall up, which was back in 2002, we started to collect demographic data on users regarding their industry, business position, their responsibility. And, for many years, we’ve also done behavioral targeting. What we’re now seeing is the data-management platforms have stepped up a level in terms of the sophistication that you have available to you.
In terms of reporting capabilities, you can now report back on the performance of segments that you didn’t specifically target which helps you to optimize. It also helps you to report back a broader view to the advertisers on which segments might perform for them in addition to what demographics are performing for them.
What has this meant for your discussions with clients? If it hasn’t impacted yet, what are your expectations?
I would suspect that with a data-management platform and other tools we’re developing, we’ll be able to give a better view of which targeting aspects are going to perform and work best with the given metrics that an advertiser tries to work with.
Whether that’s brand uplift, social engagement or conversions, we’ll be able to do more pre-campaign and then during the actual campaign itself we’ll be able to provide feedback as to what’s working, what isn’t working and be able to optimize against all of those goals. At the end of the campaign, when tan advertiser is ready to repeat their business, they’ve got a strong and compelling argument for doing so.
What about cross-digital-channel data management?
Data is at the heart of everything we do. There’s two reasons for this.
First, analytics and insights for advertisers. What that really means is if somebody is buying a campaign on FT.com we give them insight not just into how many impressions we’ve delivered or the standard advert metrics. But we can also say to them, “These types of users saw your ad, these types of users engaged with your ad.” That comes from both our data as well as other sources.
We’re building in benchmarking capabilities, which give an advertiser a benchmark either against the whole of FT.com or against their competitive set. That’s something that’s fairly unique to FT.com among big publishers.
The other aspect of data is how we use it to optimize campaign performance, which is through multichannels because you need to address users no matter which platform they are on.
Can you talk a little bit about why you recently chose Krux as your DMP?
We like the management team at Krux and thought that they understood the FT’s business. The tools within Krux for audience lookalikes or understanding the performance of the top segments – or, even if they weren’t the top segments — were best in class in our opinion. They’re also fully integrated with key platforms that we use, which are DoubleClick For Publishers, DoubleClick AdExchange and Rubicon Project.
Interestingly, have their own integrated bidding platform. So, if we want to start looking at audience extension-type deals, it’s all integrated into one system rather than being in many different parts.
What do you think the biggest hurdle is for a company like FT.com or your competitive set in order to take advantage of this audience-driven programmatic marketplace?
The challenge here is that you have a massive oversupply [of inventory] and some very big players in the marketplace that can change marketing perceptions in terms of the yield and also the overall planning. What you have to provide is the quality of service, the quality of the analytics and the quality of the content– and then the premium brand’s proposition starts to come to the fore.
Considering what you are putting into place today, what are some milestones that you’d like to hit in the near future?
The big thing we are now using is machine learning across our data. One of the key goals here is to use machine learning to work toward advertiser metrics. If somebody comes to us and says they need to improve their brand perception, we will use machine learning against the campaign data to optimize against brand perception. We’re not going to simply report back at the end of a campaign how it’s performed.
We’ve got a whole string of marketing metrics that we will be looking at and trying to optimize to and report back to. That’s really the short-term milestone in the next six to 12 months with a number of partner advertisers