You Ask, I Answer: Content Marketing Measurement Formula?

In today's episode, Sunny asks, how do you create an argument that a piece of content is working using data? Is there a go-to formula you can pull to say yes, kind of sorter, needs work or no? There's a bunch of different ways to prove that uh piece of content is working based on what your outcomes are. So that's the big thing is you have to know what the outcome is first in order to be able to say, yes, this content is working or no, this content is not working. Most of the time for most companies, some sort of online conversion will be the indicator that a piece of content is working, whether it's a shopping cart checkout, uh directions, driving directions, or someone calling a phone number, filling out a form, downloading something.

But there's generally some sort of action someone can take that is a proxy for the outcome that you're looking for. So the first and most important thing is to have good analytics software set up and tracking those those conversions. Um you can use Google Analytics, Adobe Analytics, Matomo, you know, take your pick. Once you know that, then depending on the features that are built into the software and what capabilities you have, uh, you could create a couple of different types of analyses. There is the most basic one, which you'll see in Google Analytics, which is just page value.

And you this is calculated by the software based on uh the conversion values that you pass into Google Analytics. If you say the a form fill is worth you know $125 dollars, then it will amortize out, spread out that value when a conversion occurs across all the different pages of your website that a person visits on the way to conversion. That's probably the easiest type of content valuation. Look at the engagement rates with that content by whatever uh measures you are typically using. It could be likes, comments, impressions, page views, time on site, engaged users procession, whatever the the uh metrics are.

And then based on that, uh do a regression analysis against your outcome, like form fills, uh demo requests, things like that. And you can start to look at are there are there specific channels or specific pieces of content that suggest uh you know, if if you a user consumes them that somebody is likely to convert. The most advanced models use things like uh Markov chain modeling, which is a type of machine learning, to analyze the propensity of somebody to convert based on being exposed to a piece of content. Um, this is something that um I wrote some software for for myself, which essentially looks at whether or not a piece of content was consumed and what the probabilistic outcome is that consuming that piece of content leads to a conversion. That would be the most advanced uh method for doing that, but also one of the most effective because it allows you to also take into consideration all those times that somebody consumes a piece of content and they don't convert, right?

That's the advantage of a more advanced machine learning model, is you can uh account for that non-response bias, right? Which is very hard to account for in simpler forms of content attribution. The other thing that is a general best practice that not a lot of companies do is asking people in a free form way, whether it's a survey, whether it's a web form, whether it's uh customer interviews one-on-one, you know, whatever you choose, but asking people, hey, what made you what made you come in today? What made you buy something today? What made you uh request a demo today?

Asking people that question and seeing what they say, what kind of response they give. If everybody and their cousin's saying, Oh, yeah, I saw this amazing uh webinar, okay. After enough people say that, you know that that webinar, that piece of content worked. If enough people say, I read your newsletter, cool, you know that piece of content worked. You could ask them, okay, well, which issue was the one that pushed you over the edge.

Uh, was it you know the political one? Was it the the behind the scenes one? See if they uh can can tell what individual piece of content really moved the needle for them. Those would be my suggestions for how to create a data-driven argument for the value of your content. Use whatever is best scaled to your own technical capabilities, the more mathematically and statistically rigorous you can be, and the more say the more sophisticated the algorithm, the better you're going to be able to explain to somebody um what the actual value of a piece of content is and uh why you should or should not continue to invest in it.

So do the best that you can to really level up uh your content analytics skills. Uh really good question. It's a very challenging question. Uh so thanks for answering it. If you like this video, go ahead and hit that subscribe button.