I have written a lot about what Big Data and Programmatic Marketing really mean (as well as what they do not). As marketers, we spend a lot of our time thinking about how these innovations will impact our world. But what about the person on the street? How will their world change — and is it for the better or for the worse?
(If you are newer to the topic, I suggest you start with “The Real State of Programmatic Marketing” and “Why Do Big Data & Programmatic Marketing Actually Matter?,” and then come back here.)
A Typical Programmatic Day
Let’s start with Bailey, whose day we are going to follow. She lives in NYC and is a runner. On this particular morning, she returns from her usual circuit and looks at the app on her phone to see how far and how fast she ran — likely also noting how her run compared to her previous runs, as well as those of runners who take the same route as she does.
Today is a little different because she has a notification in the app that the soles of her running shoes are wearing thin, as transmitted from the sensor and RFID chip embedded in the rubber.
A Search On Glass
Leaving the house shortly thereafter, she puts on her Google Glass and begins the search for a new pair of running shoes whilst waiting for the subway. She sees PPC ads from New Balance shoes but chooses to ignore them, opting to explore some natural links from brands she is more familiar with.
As can happen with the subway, problems are causing a delay. Knowing she has a meeting she can’t miss, Bailey gets off the subway at the next stop and jumps in a yellow cab.
The RFID chip in her phone or her Glass transits her unique ID to the video ad server tucked inside the screen in front of her, and the taxi computer transmits it to a video exchange. Her previous search has already been collected and stored against this ID, and a video ad from New Balance is shown to her.
As her taxi takes her to work, the same ID is being transmitted to the billboards in the area, and they are deciding what ads to show each moment based on the aggregate data of the people in the area. Bailey sees a New Balance ad, and the recognition of the product and the brand is beginning to register with her.
When she does finally get to work, Bailey does what many people do and checks her Facebook page. With the creation of the Facebook Exchange (FBX) last year, we can target individuals with ads on the right-hand column and also with more content-rich units within the News Feed itself. New Balance chooses to spend on her, and in turn she chooses to Like and Share the Newsfeed ad that she saw. She then goes off to her meeting.
Throughout Bailey’s office building, there is a radio playing in the background. The radio is registering the IDs numbers of all the people in the office and, much like the billboard, is making decisions in aggregate about which ads to play at each moment of the day — and of course which songs to pick, too, for maximum engagement.
What About “Programmatic Product”?
What if we could take this all out of videos and digital ads and use the same data held against Bailey’s ID # to make a physical product that Bailey just couldn’t resist buying? Today, 3D printing is developing, and it will take a while before it becomes a very fast process — but at the current speed of development, it’s not hard to imagine them being prolific in the near term.
And so, as Bailey walks down 5th Avenue during her lunch hour, her ID # is broadcast and the New Balance store queries the data and makes the decision to print a basic mockup of the shoe she has paid most attention to, in her size and in the colors she prefers.
As Bailey walks past the window, she sees that model appear just for her — and she walks inside and makes her purchase.
“Stalking Is When Two People Go For A Long Romantic Walk Together… But Only One Of Them Knows About It”
All that sounds great from the marketer’s perspective: a consumer discovers a need, and a brand can register that and take as many actions as they want on any type of device at any time.
For the consumer, though, it sounds like it sucks! Being online today can be bad enough, with companies guilty of wasting thousands of impressions on you from a visit you made 2 months ago, and for a product you already bought!
The Other Promise Of Programmatic
The real goal of programmatic that most people won’t tell you about is that you should be able to reduce your marketing spend for each consumer and isolate those individuals who are unlikely to buy from you altogether. Companies talk about how amazing programmatic is, insisting that it’s where your budgets should flow — and that’s true if it is more efficient than
everything else you are doing. But, it should not increase for each individual!
When we created Programmatic Site Retargeting, we used more data (or “big data”) to determine the real intent and value of each individual, and we found that most site retargeting campaigns should be reduced by about 40%.
We found campaigns where vendors were showing over 8,000 impressions to one computer based on a 5-second visit to a retailer’s homepage. That is nonsense, and until marketers realize what’s going on, they will keep wasting millions of dollars.
In our story of Bailey, there were many touch points that occurred, and those had to be collected somewhere. Her search first registered for the advertiser New Balance, and from that, many interactions occurred. Each potential interaction had to be judged on its own merits. Does Bailey respond to taxi ads? Is it raining, and will she see the billboard? If we pay for the taxi video ad, can we still afford to invest in her on Facebook? The data we are describing is real-time and has to be treated as such.
The big, brand-name DSPs and DMPs are failing in this area today, and marketers just haven’t realized it yet. They crunch data at set times, add people to segments and use those segments for targeting for a period of time.
In a PMP (or Programmatic Marketing Platform, a term I created whilst at Chango), all the data is queried live, meaning decisions for real-time media buying are occurring using the right data, not old data. The term is being copied now by many — but they key is to crack the code that would make New Balance outperform the likes of Nike and Adidas in winning Bailey’s business at the lowest possible cost.
Opinions expressed in the article are those of the guest author and not necessarily Marketing Land.