This is the first of a two part blog series…the next blog will be…why RESEARCHERS should care.
Ted McConnell is an amazing agent of change, having led digital advertising innovation for Procter and Gamble before becoming a consultant and the lead subject matter expert on digital for the ARF. He recently authored the book “The Programmatic Primer” for WARC, the Worldwide Advertising Research Council.
The ARF has rightfully placed emphasis on programmatic buying this year at their Audience Measurement conference.
Ted’s ideas on programmatic buying are exciting to me as they are the mechanism by which marketers will upend the purchase funnel and potentially lead to a reinvention of brand research. Ted sees all the data driving programmatic as fodder for research … we just have to learn how to use it.
Joel: In a nutshell what is programmatic ad buying?
Ted: For online, an end to end buying and trafficking system that is completely computer and data driven. By “trafficking” I mean the process by which an impression gets the right creative. It’s how the creative gets from point A to point B. Before programmatic buying, digital was a very labor intensive process but now computers can do the work instead of humans. Forrester measured that it costs 7 times as much per dollar spent on working media to plan and traffic digital media vs. TV. Programmatic, in one aspect, is just machines doing work that humans used to do.
Joel: Beyond automating processes that create efficiency, how is programmatic different from traditional media buying approaches?
Ted: In traditional media buying, the advertiser is buying an aggregate portion of the audience that the media property accumulates, such as users coming to a particular page on cnn.com or those watching a certain episode of Mad Men. Usually, audiences are bought based on weak targeting variables like age and gender along with some sense or the fit between the content and the brand. For example, traditional media placement does not allow the advertiser to selectively deliver the impression to a viewer who is into upscale beauty products but not their neighbor watching the same show who does not care about such products.
Programmatic turns the process on its ear, allowing you to selectively buy impressions that actually exist, right now. The advertiser or their agent can selectively bid for an impression based on the value for the bidder … and that value depends on what that user’s profile infers about their interests. Say you want 1 million beauty interested females…now you can find them wherever they are on the internet based on behavioral evidence they care about such products from content they consumed in the past.
Joel: What is real time bidding? Is that the same thing as programmatic buying?
Ted: Programmatic for display defines the complete end to end system. RTB is a subset of that. The real time bidding process allows any number of demand sources to be presented with the impression opportunity. It puts demand sources in competition with one another billions of times per day, using computers to decide on how much they want to bid for an impression, all happening in 1-2 tenths of a second! Where it gets really interesting is that advertisers are starting to build their own proprietary cookie pools so they know more about a given impression opportunity than their friendly competitor. It is the essence of competing on data!
Joel: I see that marketers are beginning to commit to Programmatic buying in a big way. For example, Procter recently announced that they are moving 70% of their display advertising buying to programmatic. Why should marketers get so excited about programmatic buying?
Ted: The promise, and reality is that you can improve relevance and ROI … often by 5x. Beyond that, programmatic buying is a blueprint for holistic media. You can now bid for behaviors and surrogate behaviors that deliver an audience that is very much like the audience I want from my research rather than watering it down by demographic buys. Furthermore, marketers can now measure impact more precisely with attribution analysis and actually do in-flight optimization. Programmatic give options to publishers too. They can actually buy their own inventory on the exchanges and resell it based by adding value using their first party data. For example, a Forbes reader is always a Forbes reader even if they are on a different site at that moment.
Joel: Are there words of caution regarding programmatic buying?
Ted: A lot of the value has to do with how you execute. If you aren’t careful, you will buy absolute crap. You want to work with a leading agency, network or DSP.
Joel: What do you see over the next 3 years in terms of shifting money to programmatic buying and in terms of how programmatic buying works?
Ted: Over the next 3 years— Programmatic will overtake traditional buying. It will improve ROIs and be more measurable and reliable. Your data will become your strategic advantage—translated into media advantage.
Joel: Thanks so much for your time and great insights!!
How does programmatic media buying reaches the light category buyers? Taking into account that most of their purchases are not planned, it is plausible they might not leave a digital trail of interests an algorithm can pick up and add him/her to the pool of possible interested people on skin care products.
The importance of this question is based on the fact that brand growth is light-buyer driven, therefore a significant part of such growth is delivered by what media buyers might call waste.
Targeting is good when applied in a broad sense of the concept, I think programmatic might actually generate impressions and impact on people who would purchase from you anyways and leave out those people we actually need reaching but we have no means to clearly know who they are.
This might be why TV is still the most effective medium even when we have tools like these.
Hi Miguel:
I’d love to have you share data that support the following claims:
Light buyers don’t research their purchases online
The source of brand growth is light buyers.
I could easily make the counter-argument on both points but would certainly be open to evidence that proves me wrong.
Hi Joel,
*Disclaimer: I’m not an academic, therefore I’m likely to have overstated my views and/or misunderstood what little research I’ve done.*
Double Jeopardy Law says brands differ mainly on the size of their customer base rather than their loyalty rates, since loyalty rates between brands are somewhat similar. Therefore I find save to assume that brand growth should be dictated by adquisition of new customers, which are mainly category’s light buyers because more experienced category buyers are likely to have already bought your brand in the period of study.
From this assumption, and Peter Fader’s research about the similarities between online and offline shopping (Wharton) I draw the conclusion that people with low involvement in the category are not likely to expend time rationally researching the choices offered to them but rely on heuristics and brand asset recognition to make their purchases.
Sources:
Where does sales revenue growth come from ? http://marketingscience.info/assets/documents/70/7723.pdf
Inside the Mind of the Shopper: The Science of Retailing – Herb Sorensen
How Brands Grow – Byron Hsarp
Hi Joel
great piece on programmatic. Ted is super knowledgeable. I agree that programmatic has the potential to be the way all media is bought and sold. As more TV programming becomes distributed digitally through IP (internet) rather than coaxial cable or over the air, even regular old TV ads can be purchased programmatically. Once that transition happens then every single view can be accounted for and therefore every single ad impression can be bought and sold programmatically.
And once that happens we will be able to measure and research the success and ROI impact of every ad impression, not just measure the size of the audience who MIGHT have seen the ad, as we used to do with traditional TV, radio, or print ads.
Finally, on the point that light buyers don’t research their purchases online, I might make the following enhancement — buyers of small ticket and less complex items (soup and soda) tend to do less research before the purchase; while buyers of big ticket and more complex items (cars and computers) will do a lot more research online. This should lead to the selection of the right media to use. Soup and soda manufacturers tend to use more reach and frequency type media (like TV and display ads) while cars and computer manufacturers tend to use more performance media like search ads (and only pay when they get the click).
kids clothes online in usa
Live from the ARF…The ABCs of Programmatic Ad buying…why MARKETERS should care | Joel Rubinson on Marketing Research