“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Ari Paparo, CEO at Beeswax.
The growth of walled gardens has quickly brought media buyers through the first four stages of grief:
Denial: Those little text ads don’t really work anyway.
Anger: Why won’t they let me send them a pixel? It’s 2015 and I need my pixels!
Bargaining: We won’t buy you anymore unless you implement our MRC-accredited vendor.
Depression: The duopoly will destroy us all! Doom, I say!
I’m a pragmatist, so I’d like to jump ahead to the acceptance stage and think through the longer-range tactical implications of the fact that walled gardens are here to stay. I don’t have all the answers, so I laid out some of the key challenges I see with some implications and ideas for combatting each.
Techniques Change
For advertisers, controlling the appearance and placement of their display advertising is always a key goal. Walled gardens fragment this control into multiple technology systems, causing, in the best-case scenario, more work to maintain a similar level of control. In a worst-case scenario, certain techniques advertisers believe to be important to their advertising simply no longer function reliably.
Implications and ideas:
Frequency caps no longer work the way you think: Marketers should still control frequency per channel to avoid spending their budgets unevenly, but the idea that they can control and optimize the exposures of their brands to specific users over time is out the window. There is no “optimal frequency” any more.
Whitelist and blacklist maintenance is a chore: Multiple systems = multiple lists, each of which need to be updated frequently to avoid problems. No one wants these kinds of problems … cough … Breitbart … cough.
Geofragmentation: For retail and other businesses where geofencing is a key tactic, they are now relying on multiple different geographic databases depending on the channel. How do they know where they are reaching people and at what level of accuracy?
Sequencing is a fantasy: Ad sequencing never really worked, and now it is a joke, outside of long-form video.
Rich media is a niche technology: Not new news, but clearly the salad days for rich media are over. Fatboy RIP.
From Media Execution To Data Execution
The ultimate nirvana of buy-side ad tech was a world in which a single system could, in near real time, optimize media buys across channels at the user level. For buyers, this meant some combination of a demand-side platform (DSP), a buy-side ad server and an attribution or analytics system.
Unfortunately, this dream is also pretty much dead. These technologies aren’t real-time in big swaths of the media landscape since walled gardens have their own optimization and only expose limited levers to the advertisers. Marketers can use their DSP to optimize in real time on, say, 50% of their digital spend. Less than that if they include search.
Implications and ideas:
Data and transparency become more important as decisions need to be made at a higher level across channels.
Collecting email addresses for hashing into walled gardens is a key goal.
Within the non-walled-garden channels, results are likely to become more chaotic and unpredictable, so advertisers and their agencies need control and active engagement of their DSPs and media buying.
Attribution Becomes Even More Difficult
While the smart set in digital has been pushing for 10 years to get away from last-click attribution, we now find ourselves in a place where clicks are the only common denominator across digital channels. It’s ironic, and not in the rainy-day-wedding kind of way.
Attribution vendors and ad servers have made great strides developing and surfacing multitouch and other methodologies to better measure the correlation between digital spend and results, but when a growing portion of a marketer’s media buys do not include impression-level data, these models are at best partial pictures of the customer journey.
Implications and ideas:
Partner wisely: Try to cobble together solutions that break through the walled gardens, such as those partnered with Facebook and Google.
Correlation works: Focus more on correlation of audiences and results rather than the impossible-to-measure causation.
Control groups: Use control groups extensively to identify real lift with the factors being static.
Blackouts: Consider short-term “blackouts” where media spend stops entirely on certain channels or geos to enable clear measurement of incrementality.
Last-click?: For mobile apps, maybe last-click isn’t so bad? (hides under table) If you’re buying rewarded video or interstitials on a CPI basis, there’s a clear path from call to action to conversion.
Programmatic Duplication Is The Norm
DSPs and their customers have been increasingly concerned about bid duplication causing auction prices to rise. The theory is that since exchanges claim to run second-price auctions, if you are bidding on a high-value user through multiple exchanges you can second-price yourself and not get the discount you expect.
While header bidding contributes greatly to the likelihood you may be bidding against yourself, the growth of walled gardens does even more so. Facebook, Amazon and Google all actively bid on your behalf in the open auctions, and you can bet they have pretty similar data profiles of your customers as you do.
Implications and ideas:
Reduce vendors: Not a new suggestion, but still relevant – don’t use multiple DSPs on the same account. Consolidate to a single DSP and invest deeply.
First price: The clear implication here is to stop assuming you will have a second-price auction at all, something I’ve written about previously.
Limit retargeting: It is harder than it seems to limit the number of vendors doing retargeting, but at least you can stop ad networks and black box DSPs from getting your conversion data.
Look, folks, the walled gardens are here to stay. Let’s move from complaining to deeply understanding the impact on our businesses and planning accordingly.
Follow Ari Paparo (@aripap), Beeswax (@BeeswaxIO) and AdExchanger (@adexchanger) on Twitter.
