At a talk I gave at MarketingProfs B2B Forum, one of my audience members asked a very interesting question:
“We know how to value a macro-conversion like a generated lead or a sale, but how do we value micro-conversions like engagements, video views, etc.?”
Unlike macro-conversions, which have defined formulae, micro-conversions require a few more steps and some grey areas in the math behind them.
Here’s how I approach this problem.
First, which micro-conversions matter? To determine this, I’d take every micro-conversion possible, order them in a linear time-series by date in a giant spreadsheet/data file alongside my macro-conversions, then feed them to IBM Watson Analytics. Mathematically, we have two general methods to determine which variables matter, multiple regression subset analysis or linear analysis of variance (ANOVA). IBM Watson Analytics uses the second most often, so we’ll go with that.
What we seek to understand is which variables are potential drivers of conversion. ANOVA, like all statistical methods, does not prove causality, but does prove correlation. However, data with no correlation is highly unlikely to ever be a causal factor. In this example below, I chose social media shares as my macro-conversion, asking Watson Analytics which variables are potential drivers.
Watson Analytics says that LinkedIn is obviously a driver in this dataset, as is readability grade level.
Testing What Matters
Once we’ve ascertained which micro-conversions correlate with improved macro-conversions, we need to test to prove causality. In the example above, grade level readability is correlated with our macro-conversion. Our next step would be to change the grade level of our top performing content to be more aligned with the average grade level of our audience, or test a variety of grade levels with different types of content. From there we measure to prove or disprove causality, that changing grade level impacts our macro-conversions.
Valuing the Delta
Let’s say we have proven that grade level, as a micro-conversion variable, influences macro-conversions and increases them by 1%. That change, that delta is what our micro-conversion is worth, whatever a 1% increase in conversions is worth. As another example, suppose 100 video views increases conversion by 1%. Suppose 1% conversion increases are worth a total of $100. Every video view’s micro-conversion value would be worth $100 / 100 views, or $1 per view.
This sort of conversion valuation requires us to test and re-assess frequently. Plan to re-evaluate your micro-conversion levels once per quarter in normal business operations, and once per month during whenever your “busy season” is, to better judge how your content marketing is performing.
Want to read more like this from Christopher Penn? Get updates here:
Get your copy of AI For Marketers (2019 Edition)
Also published on Medium.