Working context
A report is a set of promises: that the numbers mean what their labels say, that periods are comparable and that someone could reproduce them. Many review meetings go astray because those promises were never made explicit, and the group ends up debating whether a change is real instead of what to do about it. This guide covers the preparation that makes reports trustworthy before they are shown.
The guide offers reading frameworks, not accounting rules or reporting standards, and every example uses invented figures. Bring questions about how your records must be kept to the people responsible for them. What this page can do is help you build a short data dictionary, choose comparisons that are fair and write notes that let the next meeting pick up where this one stopped. A few agreed definitions go a long way.
SECTION 01Write a one-page data dictionary
For every measure in a review, write its name, plain-language definition, source system, owner, refresh time and known limitations. Take a hypothetical measure called new customers. Does it count anyone who registered, anyone who purchased for the first time, or anyone who purchased and was not previously a customer under another email address? Each definition produces a different number. Record also how refunds, cancellations and test accounts are treated. Keep the dictionary in a place everyone can see, with a version date, and log changes. When a definition changes, mark the break on charts so that no one compares figures that were calculated differently without realizing it. Short is better than perfect; a dictionary people read beats one they avoid.
SECTION 02Choose comparison periods that are fair
Week-on-week, month-on-month and year-on-year comparisons answer different questions, and each has pitfalls. Week-on-week comparisons are sensitive to holidays and to how weekends fall; month-on-month ones are affected by the number of days and working days; year-on-year comparisons reduce seasonal effects but can mislead if the business changed substantially. Hypothetically, a store comparing February with January should adjust for the shorter month, perhaps by looking at daily averages. Leap years, moving holidays and the weekday on which a month begins can all shift a comparison, so list them beside the chart. Use the same cutoff time and time zone for every period, and note when data arrives late. Where possible, show the underlying counts beside a percentage change so readers can judge how much a movement means.
SECTION 03Report uncertainty and decisions, not only numbers
A good review page answers three questions: what do we see, how sure are we and what will we do? Add a line that states the confidence in the data, for example that last week is provisional because payments are still being reconciled. Separate observation from explanation, and label explanations as hypotheses until tested. Record the decisions taken, the owner and the date to revisit them. If the answer is to keep watching, say so and name what would change your mind. This turns a report into a working record, and the next discussion can begin from the previous reasoning rather than repeat it from the start. Brevity helps: one page that people read is worth more than ten that they skim.
Checks to carry into the decision
- Write a name, definition, source, owner and limitations for every measure in the review.
- State how refunds, cancellations, duplicates and test records are treated.
- Mark chart breaks wherever a definition or tracking rule changed.
- Use the same cutoff time and time zone for every period.
- Show underlying counts beside every percentage change.
- Label provisional data and say when it will be final.
- End the review with decisions, owners and a date to revisit.
How to get the most from it
Read revenue drivers to understand how a total is built, measurement and reporting to make sure the figures behind it are defined and comparable, and growth decision planning to turn the findings into a considered next step. Each guide uses invented figures to illustrate the reasoning, so replace them with your own definitions before drawing any conclusion.
Common questions
Questions readers ask
Who should own a metric's definition?
Someone who understands both how the measure is produced and how it is used, often a person in operations or analysis, with input from the people who rely on it. Ownership means keeping the definition current, explaining it and flagging changes, not controlling every use of the number. Name the owner in the data dictionary.
What should I do when two reports disagree?
Start with definitions and periods before blaming data quality. Check what each report counts, the cutoff times, the filters and the source. Often both are correct for different questions. Document the reconciliation so the explanation is available next time, and agree which version the group will use for decisions. If a gap remains unexplained, raise it with the system owner.
How much history should a comparison include?
Enough to show normal variation, which for many businesses means at least a full seasonal cycle. A single previous period can exaggerate or hide movement. Where history is short, say so, and treat comparisons as tentative. Keep a note of events such as launches or outages that affect interpretation. A chart with context is more useful than one with decoration.
Bringing it together
Reporting earns trust slowly and loses it quickly. Agree on definitions before a meeting, set comparable periods side by side, show the counts and be candid about what remains uncertain. When your records feed accounts, audits or external statements, rely on your accountants and finance professionals for the rules that apply; this page can only help you ask better questions of the numbers.
These guides explain business measurement and planning concepts in general terms, so consult your accountant, finance team or other qualified professionals before relying on them for accounting, tax, legal or investment decisions.
