Data is rarely the first thing a business is missing. The harder problem is knowing what the data is telling you.
Revenue, retention, margin, capacity, implementation time, and delivery performance can all make an outcome visible. They do not always explain what created the change, which conditions were already in motion, or what the organization can realistically do next. A metric can tell you that the business moved. It cannot, by itself, tell you where to put your attention.
That distinction matters because many executive teams respond to an unexplained outcome by adding more information. They ask for another dashboard, another cut of the data, another weekly report, or another meeting to interpret the report they already have. The organization becomes more informed and no more certain.
The useful question is not simply, “What happened?” It is: “What is driving the outcome, and where can we move it?”
The metric is the beginning
A metric is a signal. It is not a diagnosis.
When margin compresses, the margin number is doing its job. It is showing a result that needs attention. It does not tell you whether the cause is customer mix, discounting, service effort, rework, late decisions, poor handoffs, or a cost that has been assigned differently. Those are different drivers, and each points toward a different lever.
When retention declines, the same problem appears in another form. The change could be related to product adoption, implementation friction, executive sponsorship, unresolved support issues, pricing, customer fit, or a shift in the market. A retention chart can show the timing of the decline. It cannot determine which explanation deserves action.
This is why the first step in a useful performance conversation is to define the outcome precisely. Not “customer health.” Not “operational efficiency.” Not “growth.” Name the business result that needs to improve, and define how you will know that it has improved.
The question gets stronger when it has a boundary:
- Which outcome is not behaving as expected?
- Over what period did the change become visible?
- Where in the business is the effect most concentrated?
- What would be different if the outcome improved?
The point is not to create a perfect definition before beginning. The point is to give the investigation something more useful than a cloud of concern.
Ask what changed underneath the number
Once the outcome is clear, work backward.
The visible number is often downstream from several layers of activity. A revenue result may be shaped by pipeline quality, pricing, sales cycle, implementation capacity, expansion motion, and customer fit. A delivery result may be shaped by scope, staffing, dependencies, decision rights, and the way work enters the system. An operating-cost result may be shaped by volume, complexity, automation, approval layers, and work that no longer needs to exist.
These layers are not separate explanations to collect indefinitely. They are a way to distinguish the conditions that are creating the outcome from the activity that merely surrounds it.
That is also why a single-team investigation is often incomplete. A customer outcome can be influenced by Product, Sales, Finance, Operations, Support, and the customer’s own operating model. The team that owns the metric may not own the conditions creating it.
The Lumaire Lens traces the outcome through four questions:
- Outcome: What business result needs to change?
- Drivers: What conditions, behaviors, dependencies, or decisions are producing it?
- Levers: What can the organization realistically change?
- Impact: What economic effect should follow if the lever moves?
The sequence prevents a common mistake: jumping from a visible symptom directly to a solution.
Separate drivers from explanations that only sound plausible
Executives are usually good at generating explanations. The challenge is deciding which ones deserve to guide action.
A plausible explanation is not necessarily a material driver. “The team is too busy” may be true, but it does not tell you whether volume, rework, unclear priorities, or unnecessary approvals are creating the pressure. “Customers are less engaged” may describe a pattern without revealing whether the cause is poor onboarding, a product gap, weak executive sponsorship, or a mismatch between what was sold and what was delivered.
One way to strengthen the conversation is to ask what evidence would support or weaken each explanation. If the proposed driver is implementation friction, where would that friction appear? In time to first value? In handoffs? In rework? In escalation volume? If the proposed driver is customer mix, which segments behave differently after accounting for contract size and service effort?
This is not a request for endless analysis. It is a request to make the reasoning visible. A team can move quickly when it agrees on what it believes, why it believes it, and what would change its mind.
The MIT research on leading indicators describes the value of measures that help predict future system performance before the final result is realized. In a business setting, that means asking which operating conditions tend to move before the outcome changes. A leading signal is not automatically the right signal, but it gives the team a chance to act before the lagging metric confirms the problem.
Identify the lever the organization can actually move
An investigation becomes useful when it reaches a lever.
A lever is not just an idea or a recommendation. It is a change the organization can make, assign, observe, and connect to the outcome. It might be a pricing boundary, a handoff, an operating cadence, a decision right, an incentive, a workflow, a capacity allocation, or work that should stop.
The best lever is not always the most ambitious one. It is the one that addresses a material driver and can be moved with enough control to test whether the logic is right. If margin is under pressure because complex customers consume disproportionate service effort, the lever may be a service boundary or a commercial decision—not an enterprise-wide cost reduction program. If implementation is slow because decisions wait between functions, the lever may be a decision right or a named handoff—not another project-management tool.
This is where data becomes more valuable than more dashboards. The question is no longer “What else can we see?” It is “What do we need to see to decide whether this lever is working?”
Connect the answer to impact
The final step is to define what changes economically.
A process can improve without creating meaningful business value. A response time can fall while cost stays flat. A team can complete more tickets while customers remain stuck. A new automation can remove manual steps while the freed capacity is immediately filled with more work of equal value.
The answer should connect the lever to an impact such as revenue, margin, capacity, risk, cost-to-serve, cash flow, or speed. This does not require false precision. It requires enough clarity to distinguish a meaningful change from activity that merely looks productive.
A clear answer may be narrower than the original question. It may show that one driver matters more than the others, that one lever is within reach, and that the organization should stop pursuing several attractive but unrelated solutions. That is progress. The goal is not to explain every aspect of the business. The goal is to give the next decision a defensible place to stand.
The data was never the answer by itself. It was the doorway. The work is understanding what the number is pointing toward, then choosing the change that can move the business forward.