Sign in

Blog · Board and management reporting

Ad hoc business intelligence: answer a one-off question

Ad hoc business intelligence answers a specific management question from supplied business exports. Define the comparison, calculate the measures, reconcile the totals, and record a bounded recommendation.

The short answerAd hoc business intelligence is analysis prepared to answer a specific management question outside a recurring report. Begin with the decision, define the period and comparison, and calculate the relevant measures from supplied exports. Reconcile the result to the source totals, test plausible explanations, and state the action together with what the available evidence cannot establish.

Ad hoc business intelligence answers a one-off management question using the business files available for that decision. Define the comparison, calculate the relevant measures, and check the source totals before writing a recommendation.

Write the decision before opening the export

An instruction to analyze sales leaves the scope open. A useful question is narrower: does West's lower contribution justify investigating a regional pricing change?

Specify the decision owner, the closed period, the units included, and the comparison that would support the decision. In this case, the regional sales leader wants contribution by region and product for the same month. No customer-level detail is needed.

The broader business intelligence definition explains the recurring reporting context. An ad hoc analysis can be much smaller: one question, an agreed measure, and a checked answer.

Prepare rows at a consistent level

Use an export with region, product, units, net revenue, and direct costs. Reconcile revenue after credits and returns, and include costs for those same sales. Keep the original export intact so the summary can be traced back to it.

If revenue and cost come from separate files, summarize each to the same period, region, and product before matching. Duplicate matches can multiply amounts. Unmatched products need an exception list rather than being silently excluded.

For this fictional US distributor, direct costs include the agreed costs of supplying the products. They exclude central overhead. Contribution therefore answers a product and regional operating question; it is not company net income.

Calculate the comparison in Excel

Place this illustrative export in columns A through E, with headings in the first row:

Region Product Units Net revenue (USD) Direct costs (USD)
West Standard 200 100,000 68,000
West Premium 100 80,000 56,000
East Standard 250 125,000 85,000
East Premium 150 120,000 84,000

In F2, calculate contribution. In G2, calculate contribution margin, then fill both formulas through the last data row:

=D2-E2
=IF(D2=0,"",F2/D2)

Format column G as a percentage. A zero-revenue row needs investigation; leaving its ratio blank avoids presenting an undefined margin as zero.

In a separate summary area, use these formulas for West's revenue and contribution:

=SUMIFS(D2:D5,A2:A5,"West")
=SUMIFS(F2:F5,A2:A5,"West")

Microsoft documents SUMIFS as adding values that meet specified criteria. Replace the region criterion with East for the other summary. Divide each region's summed contribution by its summed revenue for its margin.

Region Units Net revenue (USD) Direct costs (USD) Contribution (USD) Contribution margin
West 300 180,000 124,000 56,000 31.1%
East 400 245,000 169,000 76,000 31.0%
Total 700 425,000 293,000 132,000 31.1%

Margins are rounded to one decimal place; calculations retain full precision. West's contribution is $20,000 lower, but its revenue is also $65,000 lower. A lower dollar contribution does not by itself establish a regional pricing problem.

Test the explanation at the product level

Compare the same product across regions before attributing a difference to region. From the source rows, calculate the following margins:

Product West contribution calculation (USD) West margin East contribution calculation (USD) East margin
Standard 100,000 − 68,000 = 32,000 32.0% 125,000 − 85,000 = 40,000 32.0%
Premium 80,000 − 56,000 = 24,000 30.0% 120,000 − 84,000 = 36,000 30.0%

Both products have the same margin across regions. West's lower contribution accompanies lower sales, while its aggregate margin is slightly higher. These rows do not support changing West's price list solely because its contribution dollars are lower.

That is a bounded finding. The export does not establish whether discounts, lost orders, or customer mix deserve further investigation. Ask for those fields only if they are needed for the next decision. Business intelligence charts covers how to present a comparison without expanding its scope.

Prove the answer before sharing it

The regional totals must tie back to the source export:

180,000 + 245,000 = 425,000 revenue

124,000 + 169,000 = 293,000 direct costs

425,000 − 293,000 = 132,000 contribution

56,000 + 76,000 = 132,000 contribution

Also prove that the overall margin uses revenue weights. With the unrounded row margins in G2:G5, this must equal total contribution divided by total revenue:

=SUMPRODUCT(D2:D5,G2:G5)/SUM(D2:D5)

Microsoft's SUMPRODUCT documentation explains that it sums products of corresponding ranges. Here it weights each row's margin by revenue. Covirage is not affiliated with Microsoft.

Check that every source row has a valid region and product and that the export matches the trusted monthly revenue control. Agreement inside the spreadsheet will not reveal a source file missing an entire region.

Record the recommendation and its limits

State the question, finding, evidence, decision owner, and next action. For this example: do not change regional prices based only on contribution dollars; ask the sales leader whether the lower volume needs investigation.

If the same question repeatedly returns, define a standard review using automated business intelligence. The final findings can follow the structure in a business analytics report.

Where it goes wrong

  • Changing the question after seeing the answer. Record new questions separately so the original decision remains clear.
  • Comparing different periods. A partial month against a closed month creates an avoidable difference.
  • Averaging margins. Divide summed contribution by summed revenue, retaining full precision.
  • Treating a pattern as a cause. Product margins describe the rows; they do not prove why sales volumes differ.

Answer your own one-off question in Covirage

Bring a sample of your own revenue and cost export in a file; Covirage maps the columns, and its tools compute the comparisons and check the totals. The external AI model explains the findings with every figure traced to its rows. Create a free Covirage account, then set up your business and bring the question and export.

See self-service analytics in Covirage for supporting context.

Questions people ask

What is an example of ad hoc business intelligence?

A regional manager asks whether a lower contribution total warrants a pricing change. An analyst compares revenue, direct costs, and product margins for the same period, checks the totals, and identifies whether the evidence supports further investigation.

How does ad hoc analysis differ from a recurring report?

A recurring report repeats an established set of measures on a schedule. Ad hoc analysis starts with a new decision or exception and uses the comparisons needed to answer it. A repeated question may eventually justify a standard report.

Can an ordinary export support an ad hoc BI question?

Yes, when it contains the fields needed for the comparison and reconciles to a trusted control total. A revenue question may need period, region, product, and net revenue. A margin question also needs costs on a consistent basis.

How do you know when an ad hoc analysis is complete?

It is complete when the defined question has an evidenced answer, the totals reconcile, and the decision owner understands the limitations. Record the action and any missing evidence instead of expanding the analysis indefinitely into unrelated questions.