Three Decision Points on Programmatic Media Buying
I frequently speak with people seeking guidance on this topic. Like Socrates, I usually avoid advising them one way or another. Neither “you should do it” nor “you should avoid it” seems like adequate advice. Instead, I ask three smaller questions, seeking to clarify the key points and better enable my advisees to make their own decisions.
My essential questions come from the headings in the matrix.
Target a known audience or involve machine learning?
Most marketers are able to describe their audience, some are also able to identify their audience on the chosen platform. Young, affluent couples is a description. Addresses of 1-2 bedroom apartments in 90210 zip identifies an audience that can be mailed to. If you can identify your audience on the chosen advertising medium (mail, TV, online, etc.), you may want to target that known audience. If the best you can do is describe your audience, you may want to involve machine learning to identify them. Even if you can identify them, you may find that machine learning can refine that identification, eliminate some waste, and extract more value from your spend.
Pay a fixed fee or bid in real time?
Fixed fees are paid for direct buys that are delivered through the seller’s ad platform where the seller selects the impressions to deliver. Fixed fees can also be paid for impressions from private exchanges where the buyer can select impressions from a stream delivered by the seller. Another way to buy is by bidding a different amount for each impression in real time. Both methods allow you to select impressions from a restricted pool of impressions, so this decision hinges on which restrictions you prefer to live with. Fixed fee purchasing will give you first pick from the impressions of the publisher or network you negotiate with. Real-time bidding will give you access to almost all the content on the web, but you pick last. In some cases, the impressions you want will sell out so you better step up and pay the fixed fees. In other cases, there are plenty of impressions to go around, or at least there are plenty of equivalent alternatives, so you can take advantage of lower auction pricing.
If you find that you need machine learning to help you identify your audience, you may need to opt for real-time bidded exchanges to get enough variety and reach for the process to work.
Self-service or full-service?
There is one more decision to make. If the previous two decisions have lead you to direct buying, you will need to decide whether to make the buys yourself or involve an agency. Some brand marketers will be very effective making their own buys, but agencies have lots of resources, tools, and expertise that allow them to be both more efficient and more effective in many cases. If your decisions lead you to one of the other three boxes, the decision on which type of service to use will be a bit more complex.
The platforms that enable real-time bidding and machine learning require considerable expertise and resources to operate. Anyone who chooses self-service will need at least one data analyst to control the machine learning and one campaign manager to control ad operations. Usually an ongoing monthly spend of $300k will support these roles. The biggest brands, large agencies, and full-service DSPs have whole departments specializing in each of these functions, which allows them to maintain deep expertise and operating capacity. Even so, these large entities often look to the vendors to operate the more complex platforms. In almost all cases, most brands and small or medium agencies will lean toward a full service provider.
It may turn out that some of your campaigns land in one quadrant while others need different solutions. Hopefully this Socratic framework is useful while making those decisions.
- Pingback from Understanding Paid Media Better | jmcminn
Leave a Comment
- OpenMarket Survey/Infographic Reveal Millennials Prefer Texting But Marketers Miss the Boat
- The Triple Threat of Ad Blocking, Viewability & Fraud
- How Fallout Shelter Schooled App Marketers on Native, In-App Experiences
- Interaction 2016: Online Consumer Behavior Globally
- Up Close & Personalized: How, Why, When, Who