Joel Christie Discusses The BSkyB Exchange Strategy, Attribution Modelling And Managing Bid Inflation While Working With Multiple Exchange Traders

Joel Christie is the Senior Marketing Manager at BSkyB. Here he discusses exchange-buying strategy from the perspective of one of the UK's biggest display advertisers, how attribution is informing these buys, and dealing with the issue of bid inflation while working with multiple trading partners.

Sky as a buyer has been very pro-active on the exchange space. How is data-driven media buying delivering for the Sky brand?

It’s really helping to fuel our future strategy as it gives us far more control around the audiences we are targeting, and ultimately visibility in to the results we see on the back of it. The goal is to move away from one broad generic sales target, and look for patterns of behaviour within our current buys – isolate the mechanics of what helped drive the sale and go out and look for more of it. This hasn’t always been possible with our DR buys as we were reliant on blind network run of network buys which can fluctuate month to month. That fluctuation still occurs due to the marketplace environment but at least we now have greater visibility of what is driving the volume. Data has always been a key part of a DR buy, but the great thing now is that advertisers have the insight to see the data from themselves, rather than each networks interpretation of that data and their labelling of it in segments.

Are you working with multiple exchange buyers – or is all the buying handled by one trading desk?

At the moment we work with a small number of exchange buyers. We are really looking to test and learn within the space, and it seems that at present certain platforms are working better across certain exchanges. It is clear that consolidation is something that is bound to happen and we have taken the view that we’d really like to test a number of suppliers. One of our key decisions when selecting our agency partner for display was their technology teams and how across this market they were. It was clear that Mediacom had great experience and knowledge in this field, so we have forged a strong working relationship with Rob Webster and the team there to understand the space and what the opportunities are.

What’s the real benefit of having multiple partners buying your media across dynamic supply?

As mentioned above it seems that there are some providers that are more suited to certain spaces – what is working well in one exchange doesn’t necessarily translate in to others. The key thing to avoid where possible is to stop the “robbing Peter to pay Paul” mentality whereby the exchanges become a “Wild West” for your partners to fight it out in. The other challenge for us is that Partners who were traditionally in the network space are now evolving to become trading desks – adding to that duplication issue. The space is moving at such a pace, that we are having to re-assess our position monthly to ascertain what is happening with our buys, and where duplication is happening. Ultimately the benefit to us is that each partner is giving us new insight, and allowing us to test their platforms in the various spaces. If we liken it to the network space of say 5-10 years ago then going with one partner meant that you had all your eggs in one basket – at this stage we need to assess the different options and the buyers with ultimately the best technology, and account teams to understand the findings will grow on the plan and drive the results for us.

Given that you might have different DSPs, ATDs and Ad Traders bidding on similar impressions, how are avoiding bid inflation? Is there a process that marketers can use to avoid this outcome?

We identified this fairly early on, and look where possible to assign exchanges and audience targets for partners to operate in, but this isn’t without its problems. Policing this can be difficult, but we have found that by working closely with our buyers we can see when bids are raised from the norm – and this highlights to us that the increased CPMs are being inflated. It is very fluid, but we have been able to identify when partners are bidding in exchanges, and for audiences that we have specifically told them not to. This has meant that we can stop them bidding and reduce our price back down.

What approach does Sky take to attribution? Are you looking at all touch points in the sales funnel, in terms of awarding credit, en route to the conversion?

At present we operate a last click model, and probably will continue in this vein for the immediate future. There is a great appetite internally within the marketing team to get a view on attribution and to see where our ad spend is best used to fill the funnel. Going back to the point earlier, by understanding the impact of ad formats to audiences at various stages of the sales funnel we can look to go out and buy these audiences through the data driven model and drive incremental sales as a result.

Is the industry wedded to the last click/ad, and is it distorting the effectiveness of the post-view window?

I wouldn’t say wedded, but it’s pretty ingrained now. I think this stems from the fact it took a long while for marketing teams to convey the logistics of online marketing within organisations and get buy in from senior stakeholders. It’s a complex market to articulate, especially to those who aren’t from a digital background so to engender large scale change within operations and get buy in and sign off takes a lot of time and effort. I would foresee that change will be driven from smaller, more entrepreneurial companies and once the results are evident and we see the benefits the market will be more likely to switch its focus.

Is the attribution model you employ having a big bearing on the type of vendors you partner with on media buys?

I would say yes, it certainly has an impact on who is successful on the plan.

Does Sky have a data strategy in terms of automated buying? Are you for instance using a DMP to leverage your first party data to power your DR campaigns on exchanges? Or is the strategy more tied into your analytics data?

We are currently in discussions with our analytics provider Omniture to ensure that we are utilising the data we pick up on the site and turn it in to an actionable buy on the exchanges. There have been very few vendours who have been able to sync up with omniture and make these buys for us, so we are very excited about setting this up. It will give us the ability to segment our audience and make relevant buys with tailored creatives.

How important is display advertising in the media mix? How does it compare against search or even Facebook display? Is exchange-buying a move in the right direction for display?

Display is a large part of what we do in paid media and plays a different role from PPC and Affiliates, in that is sits across all areas of the funnel from cold to hot. The Exchanges, and moreover, data-driven buying has really helped display in that we can be far more targeted and less broadstroke in our approach. Previously we have had to run a set of creative across the whole plan, or use behaviourally targeted buys if we wanted to achieve any sort of segmentation. Now we can understand who are audiences are, where they sit in the funnel, and then create a strategy for each incorporating both the media placement and the creative execution they are exposed to.

Will video advertising on exchanges be a game-changer for brands – and could it attract more brand budget into the automated channel?

Definitely, for all the reasons I have outlined for standard display in that it provides far more granularity and control for the marketer. Although in an acquisition role, we haven’t quite seen the benefits of video – I’m sure this will play a role for the brand teams, who have already begun testing in this area.