In the past, media-mix decisions have been a fairly static affair. Marketers would take into account existing market research, performance of past campaigns, goals of the current campaign, and other data before allocating budget across the various media types in the plan. In the digital/programmatic world, this would often be split between mobile, video, display, native, social, and sometimes others. Then those budgets were often split again amongst various vendors/networks/platforms for each of the media types.
After the campaign started running, the agency or advertiser would wait two or three weeks for initial results and then reallocate amongst media types and/or vendors. Given the way media has been purchased in the past, this process made complete sense. And under this process point solutions could thrive…just take a glance at the LumaScape for confirmation.
Omni-channel programmatic promises to blow up this model.
As omni-channel programmatic becomes fully developed, this static/manual approach to media mix decisioning just won’t work any more. The promise of omni-channel is to automatically allocate budgets to the most effective impressions across all media types and devices. Instead, Omni-channel done right will require a centralized decisioning engine to determine which impressions, on what types of devices, on what media type, with what creatives, in what locations are the most effective at various times of the day/week/month.
For example, consider an omni-channel campaign for an auto dealer. It could be that for creatives showing sports cars, the most effective impressions will skew towards desktop display during the business day, mobile video during weeknights, and native on weekends. Creatives showing mini-vans may have a different optimal mix. And the optimal mix will likely change during the duration of the campaign.
Many advertisers have taken great strides in centralizing their data in a DMP. However, they will not optimize their ROI to the potential of omni-channel programmatic if they adopt the legacy approach to media mix, in which they parcel out budgets and data from their DMP to various point solutions each with their own decisioning engine. By doing so, they may get the best impressions from each point solution, but they will not get the best performing impressions overall.
To get the best impressions overall, across all media types, advertisers will need to centralize their decisioning within one optimization engine. Here are some reasons why:
Media-mix decisions need to be made minute-by-minute, not week-by-week. A marketer might find their financial services campaign is performing better on mobile devices during the day while users are checking their investments through mobile apps. Yet, their native tactic might perform better at night when those users are being retargeted as they check their social media accounts. Just like the auto dealer example above, media optimizations need to be nimble, adapting based on how users are responding to ads on various media types and devices at various times.
If a media budget has been distributed to various point solutions, then it is very difficult to determine in real time which targeting tactics/devices/media types are most effective and optimize accordingly to the most effective impressions.
If media budgets are distributed to multiple point solutions, frequency capping is very difficult to optimize. For example, a typical campaign may have budgets that are split between a mobile, a video, and a display vendor, each with instructions to deliver no more than four impressions to any user within any 24 hour period.
This is not optimal, as ideally a global frequency cap across all media/device types would be enforced, with the bulk of the spend going to the most effective impressions. The global frequency cap can be implemented much more easily with a centralized engine making decisions on impression bids.
Cross-channel learning / optimization
Optimizations depend on learning from data. The more quickly data is collected, the more quickly optimization models be developed to target budget to the best impressions. When budgets are split between various point solutions, data is collected in separate systems and it lengthens the time required for enough data to be collected to drive meaningful optimizations.
One of the benefits of Omni-channel programmatic through one decisioning engine is that data for optimization is collected more quickly so that models can start influencing campaigns sooner. Also, learnings can be shared across tactics. For example, if a campaign is performing well on a mobile site, the desktop and/or video campaigns can leverage that data and direct more impressions to that site to see if the high performance holds true with those tactics.
Consolidated Insights / Path to Purchase
Many independent attribution solutions can provide views across media and device types on the “path to purchase” that a consumer takes on the journey to a conversion. However, it is difficult for those solutions to collect all the data necessary to provide truly deep insights on the user. For example, if the user was targeted with a search retargeting ad, the attribution system may show the site upon which the impression was targeted, but may not be able to collect the keyword upon which the user searched that drove the targeting. Or on a mobile geo-fence campaign, an attribution system relying on point solutions may not collect specifically which geo fence (e.g. which auto dealer location) the user visited.
Reporting from a single platform can provide advertisers with a complete, more holistic approach to campaign assessment, including more details around the specific data that was used to target each impression.
As advertisers continue to work toward better ROI and deeper insights into what is driving the success of their campaigns, point solutions are in danger of being left behind. Omni-channel strategies that take advantage of a centralized decisioning engine, on the other hand, provide the agility and robust data collection/analysis advertisers need to create truly effective programmatic programs.