Marketing automation and Jurassic Park

I recently saw the movie Jurassic Park again. It’s about an island of cloned dinosaurs, intended as a kind of zoo and amusement park, that spirals out of control during a pre-opening inspection. Spiraling out of control is a euphemism for having ravenous pre-historic creatures devouring everyone in sight. (Remind anyone of their last budget meeting?)

It made me think of marketing automation.

One of the characters is a mathematician who specializes in chaos theory, Dr. Ian Malcolm (played brilliantly by Jeff Goldblum). Dr. Malcolm had been a critic of the park from beginning, predicting its collapse because chaos theory says that complex systems, with certain conditions, cannot be tightly controlled — such as the weather, global financial systems, and an ecology of dinosaurs on a 20th century island.

Dr. Malcolm’s warning: trying to force an overly simple management structure upon such a complex system is a recipe for disaster.

Complexity in marketing

I am fascinated by how complex marketing has become in the past decade. If you stop and think about the explosion of search marketing, social media marketing, email marketing, e-commerce web sites, mobile marketing, behavioral targeting, and budding semantic web marketing, it’s breathtaking.

But to fully appreciate the complexity generated by all this, consider how these new vehicles build upon each other:

• all the original marketing vehicles still exist and require attention;
• there are now all these new marketing vehicles (e.g., web site, search, email, social media, etc.);
• each of these new vehicles has near unlimited granularity (e.g., not unusual for a company to manage thousands of keywords in a search marketing campaign);
• the management of these different vehicles is often highly distributed across many people (e.g., the folks handling search may be far removed from the people running email campaigns);
• there are all sorts of interaction effects, within and between vehicles (e.g., a particular email campaign can trigger new searches — are the resulting messages consistent, or at least compatible?);
• particularly in social media — but also in search — end-users and third parties also drive the outcome, preventing marketing from predictably controlling those vehicles;
• virtual proximity multiplies the number of competitors, across more niches and more regions, and each of those competitors bring all the complexity above from their own marketing into the mix;

It’s not just that each new vehicle brings its own complexity to the table; each one simultaneously increases the complexity of all other vehicles too. This causes exponential growth in marketing complexity.

What’s interesting about this is that the actual atomic activity associated with these efforts can be quite simple — at least on the surface. What could be more straightforward than bidding on a keyword in Google AdWords, or posting a link to an article on Twitter?

The complexity comes as a result of lots and lots of these simple atoms bubbling up together into larger structures, catalyzing — and being catalyzed by — reactions in other vehicles. Ironically, the very simplicity that makes it so easy to create a single AdWords ad, causes hundreds or thousands of them to sprout up in unpredictable ways, generating a much more advanced kind of complexity.

This generates often surprising emergent behavior in the overall system.

The irony of marketing automation

The great irony of marketing automation is that it is often pitched as a way to simplify this exploding marketing universe, when in fact, it actually introduces a whole new layer of complexity itself.

Marketing automation, as it sounds, attempts to automate routine marketing processes. Marketers put in place software configurations and rules to handle various trigger events in a systematic fashion. For example, a prospect who fills out a lead generation form on a web site might be automatically queued to receive a series of follow-up “lead nurturing” emails over the next several months.

Sounds great, right?

There are two challenges with this, however, neither of which has been given due respect:

First, the model of interactions that are addressed by a marketing automation system are only an approximation of the real-world dynamics at play. This is by necessity — like trying to predict the weather, there are simply far too many variables all interacting together in the real world to accurately capture them all in a computer algorithm. The difference between the approximated model and reality creates a window for error. Over time, that window tends to grow larger. This is a classic phenomenon of the butterfly effect.

For instance, in our simple lead nurturing email example above, what happens if someone sends out a special offer email to all prospects — as a one-off campaign — that accidentally conflicts with the next email in the automated sequence? Ideally, you would want that scenario to be addressed by the automation system, but that might be beyond its model.

Second, marketing automation configurations and rules themselves introduce a new kind of variable into the environment. These modeled activities and the resulting actions they trigger are a whole new set of elements in the mix that have interaction effects with each other — and all the rest of your marketing initiatives.

Again referring to our lead nurturing example, before when there wasn’t an automated re-marketing program, you didn’t have to worry about one-off email campaigns conflicting with it. Now you do.

For a further dose of irony, the easier marketing automation software makes it to build up more and more automated programs and rules, the more complex the overall environment becomes.

And the example we’ve been looking at here is a very simple incarnation of marketing automation. In contrast, when you consider what multi-channel management and enterprise marketing management (EMM) software platforms are trying to do at a far grander scale, it’s hard not be skeptical — do these folks really think they can keep the dinosaurs penned in nicely bounded cages?

Taming the Tyrannosaurus

Despite my concerns, I’m not against using software to help with this brave new world of marketing. To the contrary, I think there’s tremendous opportunity at the intersection of marketing and computer science.

But as we push into this new era, I believe it’s critically important that marketers understand more deeply the complexity of their new environment. (What you don’t know can kill you!) That way, their strategies and tactics can be crafted with these dynamics in mind. In particular, I expect there will be an rising set of leadership principles in the marketing department around explicitly managing this complexity.

As this complexity begins to be appreciated for what it is, I think it will inspire a qualitatively different kind of software and management architecture to deal with it. It probably won’t embody the command-and-control heuristics of old-school marketing management so much as it will leverage the distributed and emergent nature of this new environment.

As Dr. Malcolm would say, “Life will find a way.”