Blog · Board and management reporting
Descriptive analytics turns transaction records into a checked account of past performance. Work through revenue, costs and orders by period, then see how totals, averages and reconciliations support a business decision.
Descriptive analytics summarizes recorded business activity so a team can see what happened in a defined period. It turns invoice lines, order records and cost exports into totals, averages, distributions and comparisons. Its first job is to establish a set of facts that agrees with the underlying records.
A useful summary tells a reader how much activity occurred, where it occurred and how the result changed. Revenue by month answers the scale question. Revenue by customer segment shows the composition. Orders and revenue per order help distinguish more activity from a change in transaction size.
These are observations about the selected population. A revenue increase can coexist with a lower margin or fewer customers. The report should keep those measures visible together and explain their different denominators.
It should also state its limits. A monthly summary does not prove that a promotion caused higher sales or that the next month will repeat the result. Those questions belong to investigation and forecasting after the descriptive figures are checked.
The core methods are straightforward, but they answer different questions:
| Method | Business question | Measurement choice to document |
|---|---|---|
| Total | How much revenue or cost was recorded? | Included accounts, signs and currency |
| Count | How many orders occurred? | Unique order identifier and canceled-order policy |
| Mean | What was revenue per order? | Total net revenue divided by included orders |
| Median | What did the middle order look like? | Order-level values and the included population |
| Distribution | Were results concentrated in a few large orders? | Value bands and treatment of unusual records |
| Period comparison | How did the same measure change? | Matching definitions and comparable periods |
The NIST explanation of measures of location distinguishes the mean from the median and explains how extreme values can move the mean. That matters when a small number of large orders dominate sales: an average can describe the total accurately without resembling an ordinary order.
For distributions, a histogram groups observations into value bands. Use consistent bands when comparing periods. A chart whose bands change each month can hide a change in the underlying transaction sizes.
The separate guide to spreadsheet descriptive statistics covers formulas and statistical outputs. The spreadsheet add-in setup guide owns the setup instructions; choosing an add-in is a separate task from deciding what the business summary should measure.
Consider an illustrative US commercial supplies distributor. Its analyst selects completed orders posted in January and February 2026. Revenue is after discounts and credits; costs are the cost of goods sold associated with those orders. All amounts are USD, and the segment labels remain the same across both months.
The detailed file needs order ID, posting date, segment, net revenue and cost of goods sold. If it contains multiple lines per order, count distinct order IDs after applying the period and status rules. Summing a line-level order counter would overstate activity.
The table below is the checked aggregation of that export. Each order belongs to one segment and one reporting month, so its count appears only once.
| Month | Segment | Orders | Net revenue (USD) | Cost of goods sold (USD) |
|---|---|---|---|---|
| January | Commercial | 80 | 80,000 | 52,000 |
| January | Public sector | 20 | 40,000 | 28,000 |
| February | Commercial | 100 | 90,000 | 63,000 |
| February | Public sector | 25 | 50,000 | 35,000 |
| Both months | All included segments | 225 | 260,000 | 178,000 |
This preserves the relationship between financial results and activity. It also gives the reader a small set of rows from which to rebuild the summary.
Gross profit = Net revenue − Cost of goods sold
Revenue per order = Net revenue / Distinct included orders
Period change = Current result − Prior result
Percentage change = Period change / Prior result
| Measure | January calculation | January result | February calculation | February result |
|---|---|---|---|---|
| Orders | 80 + 20 | 100 | 100 + 25 | 125 |
| Net revenue (USD) | 80,000 + 40,000 | 120,000 | 90,000 + 50,000 | 140,000 |
| Cost of goods sold (USD) | 52,000 + 28,000 | 80,000 | 63,000 + 35,000 | 98,000 |
| Gross profit (USD) | 120,000 − 80,000 | 40,000 | 140,000 − 98,000 | 42,000 |
| Revenue per order (USD) | 120,000 / 100 | 1,200 | 140,000 / 125 | 1,120 |
| Gross profit margin | 40,000 / 120,000 | 33.3% | 42,000 / 140,000 | 30.0% |
Percentage results are rounded to one decimal place; calculations retain their full precision. The gross profit margin guide explains the financial definition in more depth.
Revenue rose by $20,000: 140,000 − 120,000. Dividing that change by January revenue gives 20,000 / 120,000 = 16.7%. Orders rose by 25 / 100 = 25.0%, while revenue per order fell by $80, or 80 / 1,200 = 6.7%.
The facts support a precise statement: the company recorded more orders and more revenue, but less revenue per order. Gross profit increased by $2,000, while its share of revenue declined. The summary alone does not establish whether discounting, customer mix or another factor caused those movements.
The monthly revenue totals must add to the selected export total: 120,000 + 140,000 = $260,000. Costs must also add: 80,000 + 98,000 = $178,000. Subtracting those totals gives $82,000 of gross profit, equal to 40,000 + 42,000.
Check financial totals against the corresponding ledger accounts on the same posting basis. If the report deliberately excludes an account or status, show that exclusion as a reconciling item. An unexplained difference is unfinished work.
Order counts require a separate check against distinct IDs. Financial reconciliation can pass even when an order was duplicated across segments, particularly if an allocation divided its dollars correctly but repeated its count.
The sales manager can ask why orders increased faster than revenue. Finance can investigate the lower margin. Operations can assess whether a higher order count is changing handling work. Each decision begins with the same checked summary rather than a different version of revenue.
The report preparer owns the aggregation and completeness checks. Finance owns revenue and cost definitions. The commercial owner validates segment labels and decides which movement needs investigation. Record those responsibilities with the reporting period so the next review can reproduce the result.
Bring an order or ledger export, including dates, segment labels, revenue and matching costs. Covirage maps the columns; its tools compute and check the figures, and the external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring a sample of your own export.
See self-service analytics for business teams for the supporting approach.
A monthly sales summary is descriptive analytics when it totals net revenue and orders by segment, compares them with the previous month and checks the result against the source records. It reports observed activity; an explanation of customer behavior requires further evidence.
Common methods include sums, counts, distinct counts, means, medians, percentiles, frequency tables and comparisons across periods or groups. The right method depends on the question: totals describe business scale, while distributions show whether a typical transaction resembles the average.
It can identify where a change occurred, such as a segment with lower order values. It cannot establish the underlying cause from a summary alone. Confirming a cause may require price records, customer information, operational evidence or a controlled test.
The business owner defines the question and population, finance approves financial definitions, and the person preparing the report checks completeness and calculations. The final reader should know who can resolve a missing record or explain a change in measurement policy.