Opinionated, un-opinionated analytics tools offer different views for publishers
Smart Data Initiative Blog | 28 March 2022
The head of data at a large American media group was telling me last week: “The journey starts with descriptive analytics. How do you bring that together so you can accurately answer the question, ‘What happened?’ It requires common language, storytelling …”
The words that hit me there were “common language, storytelling.”
Because the question of “what happened” is something you can find in a good un-opinionated analytics tool. Was there a click? Was there an event recorded? Who came from where? You can get all of this from your standard analytic tool.
But the storytelling, that’s often not easy to find in our analytics toolbox. Or rather, it’s built in certain tools but not others.
Context, storytelling — for descriptive analytics tools — are usually trend lines and baselines (more rarely, scores). It is when the tool says, “This information relative to that information is the meaningful thing.” It’s when the user interface of that tool comes strong to put a big number up top that says: “This one is the big one,” and by definition the other ones are less of a focus for the humans.
Even so, context may still be opaque. The way a baseline is constructed, for example, may not be transparent to the user of the analytics tool. To say nothing of edge cases. We’ll get to this later.
But in any event, you don’t really get this out of the box from some of the most common analytics tools we use. Pretty much every company out there uses one large unopinionated analytics suite as the foundation of their analytics — and that’s often Google Analytics. Sometimes, it’s Adobe Analytics. When we do get more of this context from our analytics solutions, it’s usually because we’re using a more opinionated tool.
(Puts on product manager hat.)
So what is an opinionated analytics product?
In software products, there’s a style of product called “un-opinionated.” These are the products that are light on defining specific personas or limiting use cases for the usage for the product. Un-opinionated products try to cover “all the possible contexts in which you may need our class of product.”
A classic example of an un-opinionated product is Microsoft Word. As a product, it tries to serve “any need around word processing.” From printed flyers to macros, Microsoft Word tries to cover all possible use cases.
Meanwhile, at the other end of the spectrum in the same category of products, you have Google Docs, a very opinionated product: Google Docs’ first filter is: “a very simple experience of online Word processing for your everyday digital document writing needs.” Because “a very simple experience” is high in its brief, Google Doc does a lot less than Microsoft Word. While the two products have obvious overlap in users, circumstances of usage, and features, they stand with very different philosophies for what the product is relative to its user market.
Why does this matter? Un-opinionated products tend to come with more transparency in their set-up, and fewer, if at all, assumptions of usage. That’s pretty much what makes them un-opinionated.
More in tomorrow’s blog on which is best.
If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.