‘Mass Media’ – those two words are spat out in distaste by everyone from new media gurus to ‘enlightened’ executives embracing our brave new social media-enabled world. They stand for a bygone Mad Men era when empires were built by bludgeoning through markets with paid media spending and studios and records labels held a tight grip on massive distributions systems. Ideas came from ivory towers and gifted intuition.

And it worked well – back in the day.

In the race to distance ourselves from those cursed words it seems business – and especially the marketing and advertising industries – may be leaping to the other extreme. It seems we are moving from mass media to math media.

Mass media was about ‘big ideas’ pushed down through static distribution networks to passive audiences that were treated like the numbers they were. Conversely math media is algorithm worship plain and simple. It is the belief that human behavior can be cleanly quantified and that marketing is now a matter of measuring impressions, click throughs and keyword optimization. Fundamentally it is the belief that all processes can be broken down and solved algorithmically. (And conveniently by the incredibly powerful computers we have today.)

We’ve been here before.
During the industrial revolution the human body was often described in mechanical terms. This made sense as mechanical innovation was a key defining aspect of the era. Our bodies were seen as machines just like the many incredible devices being invented at the time. Not surprisingly everything from the way we talked about the brain to he way we organized schools and businesses all took on the qualities of mechanical assembly. The vestiges of this thinking are still in place today as a legacy of industrialization.

The defining technologies of the past half century or so have been digital. In keeping with past trends, we tend to describe the brain today in digital terms and envision thought processes the way we think about computer CPUs. The trouble is, the brain doesn’t really work this way. (If you’re a reader, Nicholas Carr’s The Shallows does a nice job substantiating this with neuroscience.) That said, we humans still cling to our analogies.

Back in the day, people talked about thinking with such jargon as “getting your gears grinding”. Now we talk about having “bandwidth” to accomplish tasks. Of course computers are capable of processing far more information than our brains but that hasn’t stopped us from trying to imitate them. While computers can multitask, humans don’t do this well. Yet algorithms, the ‘thinking’ computers do, are increasingly relied upon in many aspects of our lives. A recent article in WIRED about artificial intelligence made note that the financial systems algorithms, in aggregate, are far more complex than humans can comprehend and worse they’re interacting with each other in strange and unpredictable ways. Remember the ‘flash crash’ in the market not too long ago?

Even stepping away from the drama of this Matrix-like extreme, we are nonetheless bumping up against some incompatibilities between our desire to work like computers and the limits and nuances of our humanity.

The trouble with algorithms.
Algorithms are computer friendly. They work one the same boolean logic as computer science. Both live in the black and white world of mathematics. The trouble is, of course, we humans do not adhere to this clean logical way of doing things. We are irrational. We do things illogically, sometimes even if they are not in our best interest.

Take finance for example. The burgeoning field of behavioral economics starts from the premise that traditional finance is flawed because it relies on the clean world of mathematics and the assumption that human investors always behave in a rational manner. Concepts like loss aversion underscore the error in this assumption and figure into behavioral economic models for assessing things like investor risk tolerance.

In social media a similar clash between the clean math of computer programs and the messy world of human relations create novel problems. The Dunbar Number is often cited as a real-world limit on human relationships anyone can sustain meaningfully. In fact, Dunbar himself notes that when you get down to real, ongoing relationships, even 150 is a generous number. Sites like Facebook struggle today to add better ways of handling ‘friends’ online. We are seeing new types of relationships and new categories of ‘friendship’ emerge that call into question the appropriateness of the word ‘friend’ in the first place.

If nothing else, there are new categories of relationships we are now tending to: Consider relationships like “that guy you met at that event who is now connected to you on Facebook even though you only spoke to him for about 5 minutes.” He’s lumped in next to your childhood friend on Facebook. From the computer’s perspective those two are very similar. The same goes for ‘work’ friends vs. ‘going out’ friends. Sometimes they overlap. Sometimes not. Facebook, however, has a hard time telling the difference.

It’s always struck me as ironic that ‘social media’ is largely defined by the most stereotypically anti-social group of people in business (the iconic, buried-in-the-basement-server-bunker computer programmer) and tends to be measured in cold, unemotional, quantitative terms. In media and marketing today a lot of emphasis if put on ‘influencers’ as a key to success in our social media world. Yet influencers are most often measured in quantitative terms – number of friends, number of blog posts, numbers of comments, number of ratings made, etc. As anyone with a little Twitter experience knows, it’s not hard to accumulate quantitative numbers online, especially when a technology becomes trendy as Twitter did in 2009. Similarly the limits of our social connectedness as human beings (see Dunbar’s Number above) calls into question some of the current quantitative indicators of influence online.

This is not to say all megablogs aren’t influential, some of course are. However, when you add to the limits of human nature (both in relationships and attention span) a deluge of connectedness that has our Twitter feed updating several times each minute, it makes one wonder exactly how much of any one person’s Tweeting is actually seen, especially by people who follow several hundred feeds. The same holds true for Facebook status updates, RSS feeds of dozens of blogs, etc. An abundance of information creates a dirth of attention and that dirth of attention means simply accumulating followers or friends in no way indicates influence in terms of share of mind.

Mad Men v. Math Men
A recent Fast Company article covered the popular topic of the Death of Madison Avenue. As is to be expected given the slant of the source, the article points to a number of start-ups that have gone down the Math Men path. From media buying to brand building, numerous start-ups are riding the buzz of crowd sourcing and algorithm worship. BuildABrand.com breaks the creative process into convenient, repeatable steps (you know, the kind computers like). X+1 offers ‘automated decision making and personalization’. Then there’s MediaMath which cuts out the middle man and pesky thinking required to buy media by automating the process. All of these visions sound great right? Click, brand. Click, media plan. Click, decision. It’s the ideal perfect, simple, solution for our world of agile, lean-and-mean, business 2.0 enterprises.

I would say that if the flaw of the mass media ‘creative revolution’ dating back to Bill Bernbach was a lack of measurement and a reliance on creative intuition, today’s Math Men are making their own set of mistakes at the other end of the pendulum’s swing.

Math Men adore measurability. Luckily for them, they work with tools that make (quantitative) measurement easy. But simply making something easy to measure doesn’t make it worth measuring. Remember ‘hits’ on the Web c1999? That was an important measurement back in the day, yet a page with 10 .gif files delivered ten ‘hits’ each time it loaded. This made for big, juicy numbers that weren’t especially useful.

Given the general lack of attention paid to banner advertising today, ‘impressions’ as a measurement is of questionable usefulness. Just because your banner was loaded into a browser doesn’t mean it was seen (it probably wasn’t) and even if it was seen the chances of it being retained are slim to negligible. Unless of course you go back to Mass Media era measurements like ‘recall’ which, in this day and age, are also of limited usefulness. Admittedly, the industry continues to get more sophisticated about measuring online data, developing new metrics and experimenting. All good. Yet the emphasis remains on the quantitative stuff – counts and clicks – even as psychology continues to call into question the validity of the assumptions made based on those quantitative measurements.

So why does math-happy quantitative measurement still dominate the digital space? For one, quantity is easy to measure because computers are designed to count and as noted above, programming is built on mathematical concepts like binary code and boolean logic. Simply put, the Internet is good at logging numbers. Its no wonder then that Web companies offer these measures to their clients and the clients, being handed reams of data, try to work with them.

To a hammer every problem is a nail.

Bringing the qualitative to bear.
What is missing in the brave new world of digital media and start-up businesses with their swarms of numbers being pushed through algorithmic tables and into reports, is any sort of qualitative context for all that information. That’s the messy stuff, the grey area computers and computer scientists don’t like because it refuses to conform to the clean math models.

Businesses have a hard time wrapping their heads around the importance of contextual understanding. If – as stated earlier – we increasingly see our brains as computers and rely on computers for our information; and if computers are good at quantitative measurements, then its not surprising that most people are comfortable with the neat, tidy, precise mathematic approach – even if its usefulness has known limitations. In fact, the shiny object status of all things algorithmic these days sees more and more money and energy being pumped into trying to make human behavior – from purchase intent to social networking – fit into clean mathematical models.

But sociology, psychology, and other studies of the human condition – reminds us that context is king. (Moreso that content actually.) That’s because context can tell you what the content should be if you want someone to respond to it. Content developed without context is guesswork (which just puts us back in the ivory town of creative intuition but this time with more counting tools).

Context comes from understanding the world that the person you’re addressing lives in. It’s a world formed by their own perspectives. When you understand someone’s views on finances, career, education, government, community, family, etc. you can begin to interpret the data from surveys or server logs in a useful manner. Most importantly you can contextualize a product, technology, movie, service, or just about anything else within that persons’ life. From there decisions can be made with a high degree of confidence.

Research has always been a tough sell. It probably always will be. It confronts the ego: “What can you tell me about my own customer that I don’t already know?”

Well… the first point worth making is that any company doing its own research is necessarily biased and those biases sneak their way into the phrasing of questions and interpretation of data. Any statistician will tell you most surveys are designed to confirm theories not refute them. The second and more important point is that contextual research is not about a product or a buying decision. It’s not a ‘would you buy this’ focus group or a ‘tell us what you would improve about this’ survey.

Contextual research doesn’t promise to offer a hidden gem of insight. That’s not it’s reason for being. Instead, it offers a cypher of sorts through which to interpret all the data generated by a business. With a clear understanding of the context within which data is generated, the usefulness of the data can be realized.