The Dashboard Delusion

Why the CEO's analytics addiction is a distraction from the job

I have been doing this CEO gig for a long time. And for as long as I can remember, CEOs have been looking for what I call the “God system.” This system would gather all the data about everything happening in the company, then present it in a way that lets the CEO make informed, near-perfect decisions.

First, it was ERP systems. They seemed like the God system for CEOs back in the 1990s. The next couple of decades saw the rise of the dashboard. The new God system was a single view of all company analytics—revenue, pipeline, margin, hiring, churn, engagement. The ultimate dashboard would give the CEO something close to omniscience. Those promises never panned out. Today, AI has put us on a new quest for El Dorado. What if a model could understand all the data inside my company, then help me crunch it, analyze it, make decisions based on it? Maybe it could even make the decisions for me!

I understand the pull. I really do. After enough years in the CEO chair, the sheer volume of things you can’t see inside your own company feels like the main obstacle. But if you are still chasing the fantasy of a God system for the CEO, you fundamentally view running the company as an information problem. And if you hope AI will help you solve that information problem, it might be time to update your resume. Because in the future, AI will perform this exercise far better than any human ever could.

Below, I try to state plainly why the God-system concept doesn’t work. The belief is newly spreading with the rise of AI, but I think it leads capable leaders to spend their scarcest resource—time—on the wrong thing.

Data describes the company you already have

Every number in your data systems is a record of a decision that already happened in a functional area of your business (sales, marketing, customer success, HR, etc.). Even your “live” dashboards are merely reporting the recent past. That is useful. It tells you, with precision, how fast each area of the company is running. What it doesn’t tell you is: Are they running in the right direction? Are they making decisions aligned with your vision and strategy for the company?

Current AI implementations will help the various departments in your organization run faster, but running faster in the wrong direction is not helpful. The fundamental challenge for CEOs is aligning everyone’s efforts in the same direction. This is not easy when you have hundreds or thousands of employees. And to do so, you must be continuously examining the following questions:

1. What is our vision? There is no data about a future you have not yet built. By definition, the destination you have chosen to move toward does not exist anywhere in your systems. You are proposing to bring it into being.

2. Do we have the right strategy? Strategy is a hypothesis about how this company will win in a market that has not finished revealing itself. You can test pieces of it against evidence, but the bet itself is an act of judgment about conditions that have not occurred.

3. Do we have the right people? Performance metrics tell you how someone did against last quarter’s job. They are nearly silent on whether that person can hold a role two sizes larger, whether they will rise to a crisis you have not had yet, or whether they fit the culture you are trying to build rather than the one you have.

4. Is the culture high-performing, and are people getting better at their work? Culture and improvement manifest in behaviors, norms, and choices people make day to day, usually out of your view. Your data captures faint shadows of these at best.

These are among the most important questions a CEO faces. The answers are not in the data, because the answers are about the future that has to be chosen and created. Obsessive measuring of the past, and even the making of predictions based primarily on analysis of historical data, will not help you steer ahead.

The decisions you should not make

There is a second problem with the CEO’s God system and it has to do with the fantasy of centralization. Even a midsize company makes thousands of decisions every single day: a pricing exception here, a hiring call there, a customer conversation initiated, a tradeoff made in a code review, a vendor pushed. On and on. The CEO knows about a vanishingly small fraction of these decisions.

That is uncomfortable for a lot of CEOs. The dashboard fantasy is appealing because it gives the illusion of some kind of oversight of all these decisions. It assumes that if enough information flowed to the center of the organization, the center could make all the decisions that matter. But the knowledge required to make those thousands of decisions well is not at the center and never will be. Roger Martin published a great piece on this recently. He writes in part:

Leadership is about making only the choices that you are best positioned to make — and not making the ones for which you aren’t, even if you are more senior than the person who will make the choice. . . .

A surgeon has specific knowledge about the patient that the hospital CEO lacks, even though the CEO is much more senior. So, it is leaderly to not attempt to make decisions for which subordinates have more specific knowledge.

The specialization and knowledge required for good decision making is dispersed across your organization, most of it tacit, much of it perishable, not all of it written down. By the time any of that appears as numbers on your screen, the moment to act on it has usually passed.

This is why running a company from analytics cannot scale. In 99 percent of cases, the CEO does not have a data problem. They have alignment, strategy, and communication problems. The job is to build a company in which thousands of distributed decisions get made well, in alignment, by the people closest to the facts.

What AI actually changes

All this said, I am an optimist about AI in the enterprise. AI is extraordinary at work where the destination and the path are both known, or can be derived with some accuracy. I believe AI will continue to absorb much of this task-oriented work, clearing much of the underbrush of the work done in the organization. I think that wise companies will use this as an opportunity not to slash their workforces but to deploy humans to the kinds of work only we can do (including orchestrating and aligning those AI agents themselves).

But the CEO’s own work sits in an almost entirely different category than task work. The destination is something you choose, and the path to it is unknown. No one has run your company, in your market, at this moment, toward the future you are trying to build. There is no path to copy. There is no “right” answer to retrieve. This is goal-oriented work. AI can amplify a leader who knows where they are going, but it cannot supply the knowing. It can help you play the game well. It can’t tell you if you’re playing the right game.

A dashboard is a fine instrument for the journey. It will tell you how fast you are traveling and whether the engine is sound. It cannot choose where you are going, and it cannot carry you there. Those remain, stubbornly, the work of the person in the driver’s seat.