Writer's Notes · Analytics · Strategy · Decisions
The dashboard is not the decision.
Businesses can now collect more information than most leadership teams can absorb. The real skill is turning that volume into a small number of objectives people understand, own and know how to act on.
By Dr. Rob Urban · Paper Boat Media · Writer's Notes
We have more information and less patience for it.
Analytics platforms record behavior. CRM systems document relationships. Dashboards compare performance. Financial systems track value. Service platforms record friction. Operations tools capture time, capacity and failure. Research adds interviews, surveys and observation. The modern business is not short of data.
It is often short of synthesis.
A leadership team can spend twenty minutes hearing numbers with complex names and still leave without knowing what needs to happen next. One person reports traffic. Another reports qualified leads. Sales discusses pipeline stages. Finance speaks in revenue and margin. Operations sees capacity constraints. Customer service has a list of recurring complaints. Every report may be accurate while the meeting itself produces no useful decision.
The problem is not that numbers are boring. Numbers can be fascinating. The problem is that raw information arrives organized around the systems that collected it rather than the decision the business needs to make.
A useful analysis does not end with “here is what happened.” It reaches “here is what we should do, who should own it and when we will know whether it worked.”
This is the difference between reporting and management. Reporting describes the instrument panel. Management decides whether to slow down, change direction, investigate the warning light or stop pretending the fuel gauge is merely expressing an opinion.
Begin before opening the dashboard
Start with the decision that deserves evidence.
The fastest way to drown in data is to begin by asking what the data says. A better opening question is: what decision are we trying to improve?
A decision gives analysis a boundary. It identifies the people affected, the time available, the alternatives under consideration and the cost of being wrong. Without that boundary, every interesting number becomes eligible for the meeting. That is how a ten-minute review becomes a guided tour of the software.
Name the decision
Are we deciding where to invest, which audience to prioritize, why conversion weakened, whether to expand capacity or what should change in the customer experience?
Name the consequence
What happens if the decision is delayed or wrong? Revenue, margin, trust, retention, safety, employee workload and opportunity do not carry equal risk.
Name the evidence threshold
Some decisions need directional confidence. Others require stronger validation. Analysis should be proportional to consequence, reversibility and uncertainty.
A dashboard question: If removing a chart would not change a decision, prompt an investigation or clarify accountability, why is the chart occupying the screen?
From information to coordinated work
A seven-step path from signal to review.
Synthesis is not the act of shortening a report. It is the act of deciding what the evidence means for the organization now.
Identify the meaningful change, pattern, gap or anomaly.
Frame the business question the signal creates.
Define the outcome that should change.
Assign one accountable decision owner.
Choose the next controllable move.
Select evidence that will show progress or failure.
Set the date when learning becomes the next decision.
The first signal might come from web analytics, a CRM report, an operations queue, customer interviews or a margin review. Its source does not determine its importance. The decision context does.
Google's Analytics developer documentation shows how extensive measurement systems have become. Salesforce describes CRM as a system for managing customer relationships and connected information across functions. Both can supply valuable evidence. Neither can decide the business objective without human judgment.
An objective should change behavior, not decorate a slide.
“Improve performance” is not an objective. “Increase engagement” is not yet an objective. “Use the CRM better” is an expression of disappointment wearing office clothes.
A workable objective gives a team a destination and enough boundaries to make choices. It connects an outcome to a specific audience or process, a time horizon, an accountable owner and a guardrail that prevents improvement in one area from quietly damaging another.
A decision-ready objective contains five parts.
Measures then serve the objective. A lagging indicator tells us whether the outcome arrived. Leading indicators show whether the work is moving in the right direction soon enough to adjust. Diagnostic measures help explain why. Guardrails watch for unintended damage.
That hierarchy matters. When every metric is called a KPI, the organization loses the distinction between the destination, the early signs of movement and the clues used to investigate a problem. My work on KPI and dashboard design begins with the decisions a dashboard must support, not the number of visualizations a screen can hold.
A hypothetical example
Traffic is up. Leads are up. Revenue is not.
Three green arrows can still describe a business problem.
Imagine that website sessions increased 28 percent and form submissions increased 19 percent, yet qualified opportunities and new revenue stayed flat. Marketing can honestly report growth. Sales can honestly report weak lead quality. Finance can honestly ask why spending increased without a comparable return. The disagreement is not necessarily political. Each team is looking through a different window.
Evidence to synthesize
- Traffic sources and landing-page intent
- Form completion and qualification fields
- CRM stage movement and time between stages
- Sales-contact speed and follow-up attempts
- Reasons opportunities were rejected or lost
- Average value, margin and sales-cycle length
- Customer language from calls and service records
The story that produces action
The increase came primarily from broad informational searches. Those visitors converted on a low-commitment form, but many did not match the service, budget or timing required for a qualified opportunity. Sales response was also slower during the same period because volume increased without a routing change.
Objective: Over the next 60 days, increase the percentage of appropriate inquiries reaching a qualified sales conversation within one business day, while maintaining total qualified-opportunity volume and protecting margin.
Actions: Clarify page intent, revise qualification questions, create routing rules, set response ownership and examine rejected inquiries weekly.
That objective cannot be produced by website analytics alone or CRM data alone. It comes from connecting acquisition, intent, qualification, response, opportunity and economics. The result is not a prettier report. It is a coordinated piece of work.
It may also reveal that the apparent conversion problem is only a symptom. The deeper discipline is explored in using analytics to uncover root causes instead of repeatedly fixing symptoms.
Do not narrate the dashboard for twenty minutes.
People rarely need every number in the order it appears on the screen. They need enough context to trust the conclusion, see the tradeoff and understand their part in what happens next.
A decision-ready presentation can often be organized in five moves:
- What changed? State the signal in plain language.
- Why does it matter? Connect it to a customer, risk, opportunity or strategic objective.
- What is our best current explanation? Separate evidence from inference and identify uncertainty.
- What decision is needed? Present the realistic options and recommended action.
- Who owns the next move and review? Name the owner, measure and date.
Complex terms sometimes matter. They should be used when precision requires them, not because a familiar idea sounds more authoritative after being given a twelve-syllable name. If a measure needs explanation, explain it once. Then return to what it means for the business.
The audience should leave remembering the objective, not admiring your ability to pronounce the metric.
This is where the new article and my existing essay on the relationship between data, evidence and creative judgment meet. That essay focuses on responsible interpretation and narrative. This one focuses on the operating handoff from information to objectives and accountable work.
Speed without surrendering judgment
AI can shorten synthesis. It cannot own the objective.
AI can classify thousands of records, summarize call transcripts, find recurring themes, draft alternative explanations and flag unusual movement. That can compress hours of review into a much faster first pass.
It can also make a weak assumption sound polished at scale.
The people using the analysis still need to verify source quality, definitions, permissions, context and consequences. They need to know whether a correlation has been mistaken for a cause, whether missing data systematically excludes part of the customer journey and whether the recommended objective is actually within the organization's control.
The NIST AI Risk Management Framework emphasizes governance, measurement and management across the AI lifecycle. In practical business terms, speed does not remove accountability. It makes disciplined review more important because an unsupported synthesis can now travel farther, faster and with excellent grammar.
Data quality deserves the same attention. The W3C Data Quality Vocabulary provides a framework for describing data quality, measurements and annotations. A confident objective built on unstable definitions is still unstable.
Questions about turning data into objectives.
What is the difference between a metric and an objective?
A metric measures something. An objective defines an outcome the organization intends to change. Several measures may help evaluate one objective, but the metric itself does not supply purpose, ownership or action.
How many objectives should come from a dashboard?
There is no universal number, but fewer is usually more useful. The goal is not to convert every signal into a project. Prioritize objectives by consequence, strategic relevance, evidence strength and the organization's ability to act.
Should every department have its own dashboard?
Departments may need specialized views, but shared objectives require shared definitions. Separate dashboards should not create separate realities for marketing, sales, finance, operations and service.
Can CRM data be trusted?
It can be extremely useful when definitions, required fields, stage rules, ownership and data-entry practices are governed. A CRM records the process people actually follow, including inconsistencies. Treat data quality as part of the operating system, not a cleanup project performed before a presentation.
What makes a business objective actionable?
It names a meaningful outcome, affected population or process, accountable owner, time horizon, measures and guardrails. The team must also possess a realistic lever it can move.
Where should a company begin if it has too many reports?
Choose one consequential recurring decision. Identify the minimum evidence needed to improve it, reconcile the definitions and build a short decision brief. A useful small system is a better beginning than a comprehensive reporting project nobody can explain.
Part of Writer's Notes
Writing is part of the operating system.
An objective is a piece of writing. So is a metric definition, decision brief, CRM stage, dashboard label and explanation of why a plan changed.
Poor language creates operational ambiguity. Clear language allows several teams to act on the same interpretation without requiring the analyst to attend every meeting and translate the work again.
This article belongs to Writer's Notes: Writing, Storytelling & Publishing. Continue with Data Storytelling: Where Evidence Meets Creative Judgment, explore predictive analytics for business growth, or see how marketing analytics and reporting connect measurement to decisions.