We're bored to death with the classic "descriptive/diagnostic/predictive/prescriptive analytics" framework. It's a relic of data & IT consultancies from decades ago and doesn't match the reality of how business leaders make decisions, accomplish their goals, or solve problems.

If you don't believe us, try talking to a non-data person about diagnostic analytics and watch their eyes immediately glaze over. It's not that these business leaders aren't smart or technical enough to get it; it's just that the entire framework *doesn't matter.* Outside of a few fun theoretical or historical examples ("Two years ago, we noticed that metric XYZ was below threshold!" etc.), people have a hard time remembering when all this money & effort pursuing higher-tier analytics resulted in a tangible business outcome. And even if they did get some kind of outcome, they're probably left with that wriggling doubt in the back of the mind like, "Man, I'm not sure this was all actually worth it."

Real-world problem-solving looks a lot different, and if we took that seriously, maybe our approach to business analytics would also. When solving problems, people go through a four-step process:
‍
🤔 Representation — Bring the right historical context and data into working memory at the right time and place. What is actually going on in the real world right now that affects this problem?
‍
📖 Planning — Figure out how to get from current state to goal state. What actions can I take that would lead to the desired outcome, and what might happen if I do them? Which should I pick?
‍
🛠️ Execution — Execute this task, monitor progress, resolve errors, and handle impasses. When something goes wrong, how do we get unstuck?
‍
🔍 Evaluation—Did we succeed or fail at what we were trying to do? What can we do better, and what should we remember for the next time we face this problem?

Throughout every step of solving a problem, we handle rapidly changing contexts, focused and defocused attentional states, micro-problems that threaten to upset the whole thing, and real-world events that are completely outside the data but still impact the current situation.

Take a moment to remember the last time you solved a big problem and note the tools you used to do it. I would make a bet that very few of them were analytics or data tools, but instead, things like your chat applications, knowledge bases, documents, and good ol' fashioned paper and pencil. If you did use data, it was probably for progress monitoring but not much else. So my challenge for data people is: Why? Why don't our tools actually match the reality of how we solve problems? And what can we do to make more of an impact?