MELLORCATalk to Mellorca
Patterns

Reporting & Operational Visibility

The Hidden Cost of Waiting for Management Information

When every management question joins a queue at the finance desk, decisions are delayed or taken without evidence. The cost sits in the wait, not the report.

Mellorca Patterns2026-10-03

Observable condition

A managing director asks a straightforward question: which of our customers have reduced their orders since March? Or: what was our gross margin on installation jobs last quarter, compared with supply-only? The answer does not come back that afternoon. It goes onto the financial manager's list, behind month-end, the VAT return and three other requests. It arrives next week as a spreadsheet, by which time the meeting it was needed for has passed.

Nobody treats this as a failure. The question was answered, accurately, by a competent person. But the pattern repeats. Managers learn which questions are "expensive" to ask and stop asking them. Decisions about pricing, staffing, stock and customers are made on what people already know, or on instinct, because waiting for the numbers takes too long.

An illustrative example: a Western Cape packaging manufacturer with around 80 staff produces a monthly management pack that takes the accountant four working days to compile. Between packs, any question that is not already in it goes to her as a request. The sales manager wants to know whether a large customer's price increase is affecting volumes. The production manager wants scrap rates by machine. Both wait. Both, eventually, decide without the data.

Realisation

The hidden cost here is not the time the accountant spends building the answer. It is the delay, and what the business does during the delay. A management question is usually asked because a decision is pending. If the answer takes a week, the decision is either postponed for a week or taken without the answer. Both outcomes have a cost, and neither appears in any report.

There is a second, slower effect. When information is expensive to get, managers adapt by asking fewer questions. The organisation's curiosity is rationed by one person's workload. Over time the business becomes less informed than it could be, without anyone choosing that outcome.

A management information system exists, in its textbook definition, to give managers timely and relevant information for decision-making, monitoring operations and allocating resources. The word that matters in this Pattern is timely. The information may be accurate and complete. It simply arrives after the decision window has closed.

Realisation prompt: what decision did your management team make last month without the numbers, because the numbers would have taken too long to get?

Identification

This is an information-latency and reporting-architecture problem. The visible symptom is a queue of requests at one or two people. The common assumption is that the finance or reporting person is overloaded and needs help. The structural cause is that answering a new question requires a person to extract, combine and clean data by hand each time, because the underlying data has not been organised for questioning.

It is distinct from When the Weekly Report Is Already Out of Date, which concerns the staleness of standard reports, and from Why Management Finds Out Too Late, which concerns warning signals that do not surface. This Pattern is about the time between asking a question and getting an answer.

Diagnosis

  • Data lives in operational systems only. Sales, production and finance data sit in separate applications. Every cross-cutting question requires exports and manual joining.
  • Definitions are unsettled. "Margin", "active customer" and "job cost" mean different things to different people, so each request starts with a conversation about what is being asked. The Pattern When Nobody Owns the Definition of a Business Metric explores this.
  • Cleaning happens per request. Codes, customer names and categories have to be tidied every time, because the clean-up is never saved back into a reusable model.
  • Reporting skill is concentrated. One or two people can build an answer. Everyone else must ask them.
  • Reports are built as documents, not models. The monthly pack is a set of finished tables. It answers the questions someone anticipated and cannot be explored for the ones they did not.

IBM's description of self-service analytics captures the traditional arrangement: data teams prepared and delivered data, and business users waited for reports. The alternative it describes, where staff can create reports and run ad hoc analysis themselves, depends on something that is easy to overlook. IBM stresses that the underlying pipelines and governance must be in place first, so that the data reaching users is reliable.

Commercial impact

The commercial value at stake is speed, margin, revenue, risk and capacity.

Speed is the direct dimension: decisions are delayed by the reporting queue. Margin is affected when pricing, discounting and cost decisions are made without current evidence. Revenue is affected when a declining customer or a profitable segment is noticed weeks later than it could have been. Risk appears when decisions taken on instinct are wrong in ways the data would have shown. Capacity is consumed in finance or reporting staff rebuilding similar answers repeatedly. Further discussion of the consequences sits in the pain of decision-making with outdated information.

Common misidentification

Buying a dashboard tool is the most common response. Modern business intelligence tools are capable: Microsoft's documentation for Power BI, for example, describes connecting to many data sources, modelling and combining data, scheduling refreshes and setting alerts. But a dashboard built on unsettled definitions and uncleaned data reproduces the same disputes in a more attractive format. The Insight Why More Dashboards Do Not Automatically Create Better Decisions addresses this directly.

Hiring an analyst is the other response. It shortens the queue for a while, but if every question still requires manual extraction and cleaning, the queue returns as the business asks more.

Possibility

In a better state, the core data that management questions depend on, such as customers, products, jobs, revenue and cost, is brought together on a regular schedule into one model, with agreed definitions and with the clean-up rules applied automatically. Standard questions are answered by reports that refresh themselves. New questions can often be answered in minutes by filtering or slicing the model rather than by building a new spreadsheet. The finance person moves from producing answers to maintaining the model and interpreting what it shows.

For the packaging manufacturer, that might mean the monthly pack is generated from the model in hours rather than days, and the sales and production managers can look up volumes by customer and scrap by machine themselves.

Intervention

  1. Collect the questions. For a month, log every management information request: who asked, what they asked, what decision it supported and how long the answer took.
  2. Find the common data. Most questions draw on a small set of entities and measures. List them.
  3. Agree definitions. Settle what each measure means and which system is the source for it.
  4. Automate the preparation. Move the repeated extract, join and clean steps into a scheduled process, so the model is ready before anyone asks.
  5. Open access carefully. Give managers read access to the model with appropriate permissions, and keep ownership of definitions with a named person.
  6. Measure latency. Track how long typical questions take to answer before and after. That number is the clearest indicator of progress.

Diagnostic questions

  • How long does it take, on average, to answer a management question that is not in the standard pack?
  • How many people in the business can answer such a question without help?
  • Which questions have managers stopped asking because the answer is too slow?
  • Is the data cleaned once and reused, or cleaned again for every request?
  • Do your managers agree on what your key measures mean?

Bottom line

Waiting for management information is rarely noticed as a cost because the answer eventually arrives. The cost sits in the decisions taken late or taken blind while the business waits, and in the questions that stop being asked. Organising the underlying data once, with settled definitions and automated preparation, turns information from a request queue into something managers can use when the decision is actually being made.

Sources and further reading

Related Mellorca reading

Article summaryWhen every management question joins a queue at the finance desk, decisions are delayed or taken without evidence. The cost sits in the wait, not the report.
Explore our solutionsStart a conversation