If you liked reading this, please click the ❤️ button on this post so more people can discover it on Substack. Thanks!
One of the most common phrases in modern software marketing is “actionable insight”. Business intelligence tools promise to deliver them, but so do AI automation tools, consulting services and every other possible product you might consider. It’s become so common that you might wonder what is wrong with a product that doesn’t promise “actionable insights”.
But what is an “actionable insight”? The phrase is a clever turn of phrase that sounds important but means very little. In theory it means an insight that drives a specific action, but what action?
In business, we don’t care about insights. We care about outcomes.
It’s true that insights might lead to actions, and those actions might lead to outcomes! That is often the path that our decisions take, and insights are the raw material that we use to make them. However, it’s impossible to know if an insight will lead to an outcome until it’s all over. That’s something you decide in retrospect.
As a result, “actionable insights” can be better thought of as “potential outcomes”. The insight itself has the potential to become an action, and that action has the potential to deliver an outcome we want.
Unfortunately, a lot of things can happen between an insight and an outcome…
Can we make sense of the insight? Do we really understand what it means?
Can we identify the right action to take? Do we have the resources to take that action?
Is the outcome one that we need? Is it more important than other outcomes?
Let’s use the example of a direct-to-consumer shoe brand. They realize that shoe sales to men in Oakland, CA, about 2% of overall sales, are down 15% since the same time last year. Is that an actionable insight? No, the actionable insight would be why sales were down in that demographic.
Let’s assume they find out the problem is style, as different kinds of shows are becoming popular in Oakland. Is that an actionable insight? Yes, because actions can be taken such as offering new kinds of shoes in that market. But is that an action worth taking? Right now it would only increase overall sales by 0.3% by recapturing those lost sales.
But what if that style shift is a larger trend, what if sales would continue to slide if no action were taken? Or, better, what if sales could increase with new shoes that fit the new style? Those are further actionable insights we might need to make the decision.
How likely an insight is to lead to an outcome is where experience and expertise come into play. In our example, one insight led to an actionable insight which led to more questions. It might be worth answering those questions, or it might be worth simply taking an action if we believe the outcome is likely. Almost all of our decisions are with imperfect information, as perfect information is either too expensive or impossible to attain.
None of this means that insights are not valuable! As I said, insights are the raw materials for all decisions. Not all insights may be actionable, and not all actions might be worth taking but we should gather them just the same. It is better to make a decision to ignore an insight or not take an action than never to know about them in the first place.
What it means is that you need to be careful about promises of “actionable insights” because it’s never that simple. When you replace that phrase with “potential outcomes” you can really start evaluating them. Are those outcomes you need, and how likely are those outcomes?
When you start thinking that way, then you can start to see through the marketing and find real value.
For more on Metrics and Decisions, see:
I love pushing outcomes and have even thought about pricing my product based on outcomes. The challenge with many products is that “outcomes” can be tricky to evaluate and the causality difficult to attribute accurately. In our particular case, the feedback loop is measured in years or decades. Instead, insights are available almost right away ¯\_(ツ)_/¯
This is a helpful post! Thanks Sean.