Clearly, we have an industry awash with talk about programmatic marketing and big data. But what is real, and what is BS? What practical executions exist today? And what does a future world look like if big data really is everywhere?
I am on somewhat of a personal mission to provide a simple way for marketers to think about what programmatic marketing and big data mean. To do that we must think about what a marketer’s goal is – to reach the right person, at the right time and with the right message. But in a RTB (real-time bidding) media world, we must also include the right price, given RTB is by nature an auction.
We make these decisions using data, and always have. Take the humble billboard – a marketer chooses what message to place on each and every billboard, and which billboards to choose in which towns, all by studying the population data of each point in the country. That is data in action. The more data marketers can access, the more accurate they can be with their decisions; and therefore, the more effective their marketing programs will be.
Simply put, the promise of big data is to generate better results.
Unfortunately though, there comes a point of diminishing returns when it comes to using data. The more data one uses, the harder it often is to work with.
To place our billboard, we generate a customer profile and match that against existing databases of population demographics. Now imagine layering over school data, commuter routes, average annual income, and percentage of people who recycle etc., and it is clear that is much harder. And, would the offline marketer really get a much more accurate buy by going to the trouble of marrying those disparate sources together? Probably not.
Programmatic marketing is that solution – it is the glue that binds data sources together, and makes it worthwhile to do so.
Now, a programmatic tactic comes in many forms; but, if you are using multiple data sets and creating rules around that data to better decide right message, right time, right price, then you are already doing “programmatic marketing.”
Real Executions Being Done Today
I meet many CMOs and each has their own view of what programmatic really means; but in general, the feeling is that it’s an evolution from what they were doing a short while ago, and that evolution is using more data.
Individuals are visiting your pages, you are retargeting them, and even if you throw in the dynamic creative factor, you could never be accused of rocket science.
Well, the programmatic equivalent is (funnily enough) “programmatic site retargeting” – the difference being that more data allows us to better determine who should and should not be retargeted, at what bid price and with what message.
Programmatic & The Wizard’s Curtain
Programmatic almost appears invisible to most marketers because many companies are bending over backward to keep it hidden from you – where there is mystery and smoke, there is perceived value. But like Dorothy and friends found, there is a world behind the curtain that is very different (my kids would love that I am making Wizard of Oz references at work).
The next big push in this industry is to transparency, to be honest about what’s behind the curtain and charge a reasonable price for that technology and service. Did you know, for instance, that the average cost of RTB inventory is about $1.00 (not including FBX), yet most site retargeting companies are charging $5 CPM and up? That’s the power of a shiny curtain. And to maintain those margins the companies must keep hiding things, and when they are hidden, the marketer has no idea if anything particularly clever or “programmatic” is really occurring.
You can’t blame a marketer for thinking all this big data talk is just hype!
Now You See Me, Now You Don’t
We see a similar problem with the subject of viewability, although in this case, the problem is not companies hiding the truth, it’s that marketers cannot agree on what they should be asking for.
I see RFPs coming in asking for a % of media above the fold (ATF), and the % below the fold (BTF), yet that isn’t the metric that matters at all given that most Internet users are capable of using their scroll bar! The real question is whether the marketer’s ad was in view and for how long. The problem though is that the tools don’t really exist to measure that properly, and so the industry must find some way to express this to be able to respond to an RFP.
They Need To See It
Making it real is often done with a great visual, and there are some incredible cases where big data has been brought to life. One of my favorites was done by an intern at Facebook back in 2010 and is a map of the world.
The person responsible, Paul Butler, didn’t draw this map by outlining the continents as it first looked, he took a really big chunk of Facebook data and drew lines based on where people were connected to other people, and, from that, emerged the world.
This visualization not only demonstrates the scale of data that Facebook has, but it also shows how interconnected we are with each other on a global scale, and very interestingly, which countries have not embraced Facebook and are missing from this view of the world entirely.
The marketer’s trick is to do the same. It is great if you can tie together a few of your data sets and make them actionable in a DSP (Demand Side Platform), or better, a PMP (Programmatic Marketing Platform), but then being able to report on that data in a way that makes others take notice is one of the next challenges our industry must face, (and why Paul Butler now works for Chango!).
So, programmatic marketing and big data are real. The digital marketing industry didn’t invent big data. It’s existed for a long, long time; but, a programmatic approach that glues it together to make it actionable is brand new.
Marketers are actually executing on programmatic campaigns today, but surprisingly infrequently. Most are still running regular site retargeting, for instance, and haven’t made the shift to a programmatic version to eliminate waste. That shift will come, though, as long as the industry accepts transparency and the marketer can really understand what value each point of data has, so they can learn and adapt from it.
It may be an annoyingly long time before that will come, but when we do solve it, there are some very cool things waiting for us. Billboards and radio have already become digitized, and when we can tie the viewer or listener to the device, we can start to use data to decide what we should be showing at that time. Tie that back to the other possible touch points that person represents, and we have something really programmatic.
Opinions expressed in the article are those of the guest author and not necessarily Marketing Land.