Online Engagement Metrics: You Have Their Eyes but What Are They Driving?

Define "online engagement."

Start with "engagement." I've published the dictionary and NextStage definitions of engagement in Attention, Engagement and Trust: The Internet Trinity and Websites. Neither the NextStage nor dictionary definitions deal with online engagement because both the dictionary's and NextStage's engagement definitions are about knowing that an individual is devoting mental resources to some task.

Example: You use a hammer. You're engaged in the act of hammering (say a nail into some wood). You hear your child scream from the next room as you're in mid-strike. Your attention goes from hammering to your child's scream. Your engagement in hammering decreases exponentially (I'm not using "exponentially" as a metaphor, I'm describing the curve that your mental resources demonstrate when this is going on) and your engagement in learning what your child is screaming about increases exponentially (again, not a metaphor).

Because your attention is diverted from hammering and your body is still busy bringing the hammer down on the nail, you miss the nail and smash your thumb. Now you're really engaged in your thumb because almost all of your attention is devoted to the throbbing pain there.

So "online engagement"? Easy enough, "online" is an adjective so "online engagement" is about knowing that an individual is devoting mental resources to some "online" task.

WishCasting 101

A challenge I have with all this talk of engagement is that -- as a reader told me over the phone -- people are recognizing that the market wants some kind of statistic to identify that visitors are "engaged" with a page and they have no means to come up with it so they are wishing it onto the means they have.

That's pretty much a quote. I really love it. It engages me at so many levels.

I also think it's true on so many levels.

I have no problem with people coming up with some concept and calling it "engagement". That's what language is for. But if "online engagement" is going to be a metric applied to the web then both the way it is measured and its definition have to be based on the web's abilities to be both measured and defined. And please let what the market means by "engagement" be close to what people coming up with metrics mean by "engagement".

Metrics, Methods and Madness

I'll share now that I don't put a lot of faith in metrics unless they can repeatedly predict something within a fine line of accuracy. It's the nudge (pronounced "noodge") in me.

Let me go a step further; Not only do I want a fine line of accuracy, I want it to be accurate over a long, predictive time. It's one thing to see someone pick up an item, check their wallet, begin walking to the checkout and say "That person's going to purchase that item", it's a completely different thing to see someone driving down the street and know what they're going to purchase over the next 6-12 months (let alone 6-12 years), when and in what volumes.

Let me give you another example; You have to watch the road to drive a car, correct? I mean, of course you have to watch the road when you drive a car. Well, if it's so obvious, how come people still have accidents?

Well, it's because driving a car isn't a metric (unless it's "How many people are driving a car?"), it's a method. Everybody drives a car the same way differently. You can count how many (metric) cars are going the speed limit (method), how many (metric) are lane diving (method), etc., but are they going the exact speed limit or within a few mph or kph of it? And that person lane diving, is it severe or ... wait a second, what does "severe" mean? Is your definition different from mine? Is your definition or my definition the correct one?

Now I get to talk like a consultant. "It depends..."

What it depends on is whether or not a given definition is answering the question you're asking when you talk about "online engagement" (and that might bring us back to wishcasting).

If everybody drove a car the exact same way (method) then there would be only accidents or no accidents (metrics) because knowing exactly how everyone does something (method) gives you the ability to predict (metric) what they will do and how they will do it (method), hence accidents can either occur or not. Metrics, me thinks, are binary. Methods are not.

(and while I'm thinking of it, how come there's no definition of "vehicular engagement" or "driving engagement"? I mean, do we really need a definition of "online engagement" when the existing definitions of engagement seem to suffice in so many other fields?)

Predictability is a great thing. It allows you to plan. There's a saying in psychotherapy, "Madness is doing the same thing over and over again and expecting a different outcome." Insurance companies, banks, stock analysts, financial institutions, governments, industries, ..., they all like predictability. And they're willing to go mad to get it.

You can't metricize a method. A method is how you practice the application of, for example, some metrics.

Is the method successful? Ah, now we're back to metrics. Define "success" and we can determine if the method is successful according to the defined "success" metric.

So please don't tell me that because someone's navigating a website they're engaged with the website. I completely, totally believe you're measuring something. Your measuring your definition of "engagement" online.

And now the Shocker

I'm actually quite good with that; measuring your definition online and calling it "online engagement". I have no problem that your metric is different from mine and produces different numbers. We're using different definitions, of course it'll be different.

But here's the thing; recognize the difference between a method and a metric, please. It's very difficult to scale a method. It requires training, and pure methods require people to recognize and remove their own biases so they focus (are engaged with) the method rather than their version of the method. Or, if you allow people to use a method according to their own ideas of how to apply it, recognize that any ten practitioners will come up with 12-13 different interpretations of the results because not everybody drives their cars the same way.

Which means there'll be accidents.

Please contact NextStage for information regarding presentations and trainings on this and other topics.

Upcoming Trainings:

• Know How Someone's Thinking in 10 Seconds or Less Half-day training at Toronto Emetrics Marketing Optimzation Summit, 3 April 08
• "Know How Someone Is Thinking in 10 Seconds or Less" Half-day training at the Fashion Institute of Technology in NYC, 13 June 08

• Toronto Emetrics Marketing Optimzation Summit, 31 March - 2 April 08
• New Communications Forum 2008, 22-25 April 08 at The Vineyard Creek Inn & Spa, Sonoma County CA
• San Francisco Emetrics Marketing Optimzation Summit, 4-7 May 08
• SUNY Marketing Professionals Conference at the Fashion Institute of Technology in NYC, 11-13 June 08

Sign up for The NextStage Irregular, our very irregular, definitely frequency-wise and probably topic-wise newsletter.