The Machines of Marketing
In my daily information diet, I recently came across an interesting infographic on the evolution of marketing automation, which was created by a company called Ifbyphone. They took a look at spending, and it seems that business to business marketers spent $325 million on marketing automation in 2011. This was a 50 percent increase from the previous year, as folks are really starting to think about the whole notion of marketing automation as a way to handle incoming leads. Irv Shapiro, CEO of Ifbyphone defines term “marketing automation” to include: lead management, Web analytics, email marketing, voice-based measurement, inbound marketing, and social media.
I find it really interesting to think about marketing in terms of any sort of automated approach. They’re very clear about saying this doesn’t mean robocalls and the kind of automation that leaps to mind in an old 1950’s science fiction kind of way; it’s just a name to put to a way to track and measure sales leads and sales lead generation. I tend to be somewhat old school, and I worry about many of the things I’m seeing with various marketing automation ideas I am hearing about.
New concepts of marketing, as we enter the century’s second decade, are forming based on the vast storehouses of data being created that can be accessed, analyzed, and in theory understood sufficiently to generate actions that profit the merchant. Consequently, many people do truly believe that you can somehow create an algorithm, or set of them, that will predict behavior or will be able to offer an individual human the proper marketing methods at the right times in the proper manner.
I continue to believe that humans are far more complex than any algorithm can predict. I fully appreciate the incredible work being done by IBM, SalesForce, SAP(Hana), Oracle and myriad others to harness and make available tools to mine these vast data repositories. These companies and many others are taking massive databases of information, sorting through them, and coming out with some basic business rules that allow far more intelligent use of differential marketing messages to each individual, depending upon the system’s assessment of what their needs, wants and desires might be. Yet, still I find it somehow troubling as well.
I understand that it is one of the only ways we have to take this massive data and cut it down into digestible chunks and then manageable marketing activities that can be generated by “greater understanding” of the individuals.
But somehow, wouldn’t this always lead us to kind of collapsing toward the mean?
Consistently trying to deliver to the most “appropriate marketing message” with a caveat of almost by definition one that also doesn’t risk upsetting the consumer might be a flawed notion at its heart. By having an algorithm that describes an individual as “Okay, you’re in this “segment”; therefore you get message 43B,” aren’t we then, in a different seemingly tailored communication actually further depersonalizing the whole notion of marketing? Today, we have this explosion of social activities, which gives marketers increasingly in-depth and well formed understanding of what people, as individuals really want. Yet, one of the ways that we’re reacting to all this new data is by trying to develop systems to control the amount of information, discern patterns we think have value so that we can “efficiently” target and arithmetically increase response rates. Paraphrasing a line from every press release of a publicly traded firm “past results are not predictive of future behaviors”. These are humans we are attempting to categorize – like the real REASON I bought this or that on that particular day is going to be the same later in my life. The latest IBM smarter planet ad spoke of a bakery realizing that people eat more of something in the rain, so when it rains bake more of it (recollections of advertising being what they are who knows if that’s right or not). That kind of analysis makes far more sense to me – patterns that appears in sales among collections of humans reacting to a specific situation with the store as unit of analysis. But, trying to predict what I may personally do in any given situation seems far more tenuous – yes an email campaign comprised of millions of emails that shows even a few basis points increase is in fact worth a lot of sales… and yet I am troubled.
I think the magic formula is to come up with genuinely customized marketing based on what people are really interested in. This may take a little more time and effort, but I think companies like shopittome.com may have taken steps in the direction of an emerging “better” approach. People opt in to get communications on their own set frequency that contain categories, brands and sizes that match their expressed preferences for each of those attributes.
This feels like it’s heading toward the mythical “intelligent agents” spoken about with such reverence in the late 90’s of our last tech bubble. I still long to have exactly that available to me such that I can choose in relatively explicit terms what I am interested in at this very moment. Then I’d send my agent merrily off to return in mere moments with what I want to know, buy, see or hear. Someday…
Meantime marketers should always remember a deep seated feeling we all share, in varying degrees, of wanting to be treated as unique. We all probably do have that one (or more) monthly or weekly emails from a merchant we enjoy, or a service we find interesting (OpenSky perhaps?). You unsubscribe from or ignore all the other solicitations, but faithfully open the Macy’s, Newegg.com, ModCloth, TraderJoes, or even Lowe’s email because that is what you are really interested in and you are therefore entertained by what you get. Because we all wish to have some measure of control over what we direct our most scarce resource, time, towards we pick and choose our inbox opens carefully.
The idea of “personalizing at scale” is a challenge that will not be solved overnight, but I do not think simply greater automation is at the heart of the ultimate solution. Better tools for analysis, more input into why that pattern might be happening, sure marketing automation can help us immensely with that, but a sense of one massive ever tuning machine that knows what we want – not so much. Unless I am able to control my own personal machine moment by moment will I ever get what I truly want? We shall see…