Measure, Fast And Slow
Have you read “Thinking, Fast And Slow” by Daniel Kahneman? Kahneman won the Nobel Prize in Economics, and he also may have written the greatest non-fiction book of the past decade. In case you haven’t read it, let us summarize the thesis for you: People are stupid.
Or, to put it more gently: People are lazy. What Kahneman shows in his book (and we’re oversimplifying quite a bit, of course), is the lazy choices people tend to make. If you give people a choice between an easy decision that’s wrong and a hard decision that’s right, people tend to choose the easier option to conserve mental energy. This phenomenon is nicely captured in this charming quote:
Now this may come as a surprise to you, but it turns out that tech marketers are also people, and subject to the same cognitive biases as all other humans. As Bob Hoffman, the Diogenes of Advertising, puts it: “Marketers prefer precise answers that are wrong to imprecise answers that are right.” And nowhere is this more obvious than in the case of digital analytics.
Whether it’s click-through rate, last-click attribution, or cost-per-lead, pretty much every metric that tech marketers use to measure the effectiveness of their digital ads makes absolutely no sense. Most would prefer to measure marketing influenced pipeline, total opportunity value or lifetime customer value. Most marketers are well aware of this but use these metrics anyway.
Take click-through rate, aka CTR, the dominant metric in online media, the KPI that dictates the ebbs and flows of billions of dollars. Many marketers have claimed that CTR is a worthless metric because it in no way correlates to brand awareness or even sales.
I actually think it’s much worse than that – CTR is a negative indicator. The things that get you a lot of clicks are disastrous for your brand. Things like click-baity headlines, low-frequency campaigns, and targeting robots, who are much more likely to click than real-life humans.
In other words, CTR is at best worthless, and at worst it’s a negative KPI (High CTR = Low Effectiveness). And yet, all across the world right now, marketers are re-allocating their budgets to whatever channel gives them the best CTR.
What explains this? “Thinking, Fast and Slow.” CTR is easy to track and measure, so tech marketers continue to measure it, despite being well-aware of its enormous limitations.
“Smart marketers,” meanwhile, ignore CTR and measure “last-click attribution” or “cost-per-lead.” But it turns out these metrics are often just as meaningless as CTR.
Last-click attribution is what it sounds like – whatever channel drove the last action before the purchase gets credit for that purchase. Now if you tell a room full of tech marketers that last-click attribution makes no sense because multiple touchpoints could have contributed to the sale, you will see a lot of nodding heads. Of course, it makes no sense. And yet, we continue to use it.
That’s “Thinking, Fast and Slow.”
B2B tech marketers marketers generally prefer Cost-Per-Lead (CPL) as their metric of choice, since business products and services are generally not bought online directly by a consumer. But CPL is as goofy a metric as CTR because it ignores a rather obvious question: How many of the leads actually became customers, and how much did those customers spend?
If you give tech marketers marketers a choice between a $5 Cost-Per-Lead from Channel A and a $100 Cost-Per-Lead from Channel B, guess what choice these marketers will make? They will take every dollar out of Channel B and re-allocate it to Channel A.
And what if it turns out that $100 leads became customers 30% more often, and spent 10% more over time? Well then, by optimizing towards the cheapest leads, this marketer may have lost their company millions of dollars in revenue over a multi-year period.
The good news is that things are starting to change. B2B technology marketers are starting to give more consideration to metrics that matter. One of those is “Revenue Per Lead,” which could have been used to avoid the sad outcome described in the above scenario.
Measuring by “Revenue Per Lead” isn’t easy – it’s difficult to track. But it is the right decision, and it’s getting easier.
We recently met with a company called Bizible, for instance, which now has an API integration with LinkedIn. Bizible ties together ad exposure data and CRM data to determine revenue per channel, using an algorithmic multi-touch attribution model. It can give you the sales pitch better than we can, but it struck us as the future of measurement.
When we tell clients about Bizible or other solutions, we always get the same pushback – it’s too hard to implement. They say, “Our salespeople would need to constantly update Salesforce. All our media would need to be tagged differently.” In other words, it’s a hard decision, and marketers prefer easy decisions, even if they’re wrong.
Look, we don’t mean to be hard on tech marketers. There are a million metrics to track, we need to report on something, and humans are hard-wired to prefer simple answers. All we’re really asking is for either a little more science in the form of revenue-per-lead. In other words, do what makes sense, even if it will be impossible to measure with quantitative metrics.