Another dashboard, tool, meeting, hire, or initiative can feel like progress because it creates movement immediately.
The organization is doing something. A project has a name. A vendor is selected. A team is assigned. Leaders can report that action is underway. When the business does not yet understand what is driving an outcome, that sense of movement can be comforting.
It can also make the underlying problem harder to see.
Starting with the solution turns uncertainty into activity before the organization has defined what improvement means. Different teams respond to different symptoms. Work accumulates. People defend the intervention they sponsored. Months later, the business has more systems, more reporting, and more commitments—and still cannot explain the outcome that started the conversation.
The best intervention is easier to choose after the outcome, drivers, and leverage are clear.
Activity can hide uncertainty
Suppose a company believes customer retention is weakening. One team proposes a health-score dashboard. Another proposes a new customer-success playbook. Product proposes a feature campaign. Sales proposes an executive outreach program. Finance asks for a better forecast.
Every idea may be reasonable. None answers the first question: what is actually driving the retention outcome?
The new dashboard might make risk more visible without changing the customer experience. The playbook might add activity to a team already overloaded with work. The feature campaign might increase adoption among customers who were never at risk. The executive outreach might help a handful of accounts while the broader implementation problem remains.
The issue is not that the organization has too many ideas. It is that the ideas are being evaluated before the problem has been framed.
The Harvard Business Review article on reframing tough problems describes the value of spending more effort examining a problem before selecting a solution. In complex businesses, problem framing is not a delay before the real work. It is part of the real work.
Start with one outcome
Name the result that needs to change.
Avoid starting with a broad category such as “customer experience,” “efficiency,” “alignment,” or “AI adoption.” These may describe an area of interest, but they do not establish a finish line.
Ask:
- What business outcome is not behaving as expected?
- How is the outcome measured today?
- When did it begin to change?
- Where is the effect most concentrated?
- What would improve if the issue were addressed?
The outcome could be retention, margin, implementation speed, cost-to-serve, forecast accuracy, capacity, revenue quality, or a strategic initiative that is not producing the expected result. It might also be a decision that keeps waiting because ownership is unclear.
The narrower the first outcome, the easier it is to see the organization’s leverage. That does not mean the problem is small. It means the investigation has a place to begin.
Trace the drivers
After defining the outcome, ask what is creating it.
Drivers are the conditions, behaviors, decisions, handoffs, incentives, and dependencies that shape the result. They may sit upstream from the team that owns the metric. A delivery team may own implementation time while Sales owns the scope that enters the system. A customer team may own retention while Product owns the missing capability and Finance owns the pricing structure.
This is why solutions selected inside one function often have limited effect. They address the visible part of the problem while leaving the cross-functional condition untouched.
Tracing the drivers does not require a perfect causal model. It requires enough shared understanding to distinguish a material condition from a plausible story. What evidence supports the explanation? Where does the condition appear in the work? Which customers, teams, or moments are affected? What would we expect to see if the explanation were true?
The goal is not to prove a theory beyond doubt. It is to make the working theory strong enough to choose a useful test.
Identify the lever the organization can move
Some drivers are real but not directly controllable. Market conditions, customer budgets, competitor actions, and economic changes can influence performance without being levers the organization can move.
The work is to identify the controllable point inside the system. The lever may be a process, metric, handoff, incentive, ownership structure, operating cadence, responsibility, or work that no longer needs to exist.
If implementations are slow because scope is changing after kickoff, the lever may be a commercial boundary and a decision right. If service costs are rising because every customer receives a custom workflow, the lever may be segmentation and a clearer service model. If an executive team cannot agree on performance, the lever may be a shared definition and a review cadence rather than a new reporting tool.
The lever is the bridge between understanding and action. It is where the organization can make a change, observe what happens, and learn whether the explanation was useful.
Choose the intervention last
Once the outcome, drivers, and lever are visible, the intervention becomes a decision.
Sometimes the answer is a tool. Sometimes it is a clearer handoff, an adjusted incentive, a new operating cadence, a pricing boundary, or a change in ownership. Sometimes the answer is to stop doing work that has accumulated without creating value.
The same solution can be right in one context and wrong in another. Automation may help when the process is stable and the rule is clear. It may make a broken process faster and harder to inspect when the process is still changing. A dashboard may help when the outcome and definitions are shared. It may add noise when the organization is still debating what the metric means.
Choosing the intervention after the diagnosis allows the business to evaluate it against the problem it needs to solve. The question becomes: “Will this change the driver that matters, through the lever we can move, and produce the impact we need?”
Make the decision testable
A solution should come with a theory of change.
State the outcome. Name the driver. Define the lever. Describe the expected impact. Decide what evidence will show that the change is working and when the organization will review it.
For example:
Implementation is slow because decisions wait between Sales, Product, and Delivery after scope changes. We will create a named decision owner and a scope-change threshold at kickoff. We expect fewer late-stage escalations, shorter decision time, and a reduction in total implementation duration.
That statement does more than describe a project. It gives the project a reason to exist and measures that can reveal whether the reason was correct.
If the expected effect does not appear, the organization can revisit the driver or the lever. It does not have to defend the solution indefinitely because activity has already been invested.
Know when the solution is the solution
Problem-first thinking does not mean distrusting every tool or delaying action forever.
If the outcome is clear, the driver is understood, and the lever is within reach, move. If customer response time is too slow because work is being routed manually through a stable set of rules, automation may be the obvious intervention. If the business needs a shared view of a well-defined performance outcome, a dashboard may be exactly right.
The discipline is not to reject solutions. It is to choose them for a reason that can be explained.
This also changes how teams communicate. Instead of selling a tool, a department can explain the business outcome it is trying to improve, the driver it believes matters, and the evidence it will use to judge the change. That is a stronger case to executives and a clearer brief for the team doing the work.
Stop adding activity to an unanswered question
The organization does not need to know the entire root cause before it begins. It needs to stop pretending that movement is the same as progress.
Start with one outcome. Trace the drivers. Identify the levers the organization can actually move. Connect the change to the impact that matters. Then choose the intervention that fits.
Sometimes the answer will be a new tool. Sometimes it will be advice, operating leadership, or one focused project. Sometimes the most valuable decision will be to remove work that never belonged in the system.
The diagnosis determines the intervention. That is not slower. It is how the work becomes useful.