Most commercial data requests are too long and ask for the wrong things. They are assembled by adding to last deal’s list, they arrive as a hundred-line spreadsheet, and they produce a data room full of decks. What follows is the request we actually use, organised by what each item lets you conclude — because an item that cannot change your view does not belong on the list.
The most informative part of a data request is not what comes back. It is what does not, how long it takes, and who has to be asked. Design the list so those signals are legible.
Ask for raw, not reported
One principle does more work than any individual line item: wherever possible ask for the underlying export rather than the summary. A CRM extract at opportunity level is worth more than twenty pipeline slides, because the slides tell you what management concluded and the extract lets you conclude something yourself.
This has a second benefit. Reports are prepared; raw data is merely retrieved. The gap between how quickly a company can produce a board pack and how quickly it can produce the export behind it is one of the more reliable indicators of whether the commercial function is genuinely run on data.
Three practical rules follow:
- Specify the fields, not the report. “All opportunities created since 1 Jan 2024 with these eighteen fields” is answerable. “Pipeline analysis” invites a deck.
- Ask for the whole period, not a sample. Sampling is your job, not theirs. A filtered extract has already had a judgement applied to it.
- Include the closed-lost records. Almost every request omits them, and they are where the diagnostic value is.
The list
Grouped by the five areas of the commercial engine. The right-hand column is the point: it is what the item lets you conclude, which is also the test of whether it belongs on your list at all.
| Area | What to request | What it lets you conclude |
|---|---|---|
| Proposition & pricing | Price list history for 3 years. Realised price per transaction with list price alongside. Discount approval policy and actual approvals granted. | Whether pricing authority sits where management thinks it does. Dispersion between list and realised, and whether it is widening, is the single fastest read on commercial discipline. |
| Proposition & pricing | Revenue by product or module by period. Attach rates for add-ons. | Whether the growth story is the core product or a long tail that is quietly carrying it. |
| Marketing & demand | Lead volume by source by month, with the conversion rate at each stage from lead through to closed won. | Whether demand is earned or bought, and what happens to the plan when the paid channel is throttled. |
| Marketing & demand | Marketing spend by channel for 3 years, alongside new customer counts. | The direction of acquisition cost. Rising CAC against flat conversion is a proposition problem presented as a marketing one. |
| Sales execution | Full CRM opportunity export, won and lost, 24 to 36 months: created date, close date, stage history, value at each stage, owner, source, competitor, loss reason. | Almost everything. Velocity, win rate by segment and by rep, forecast reliability, whether stages mean anything, and how much of the number depends on individuals. |
| Sales execution | Quota, attainment and tenure by seller for 3 years, including leavers. | Whether the model is repeatable or rests on a small number of people. Distribution of attainment matters far more than the average. |
| Sales execution | The last 8 forecast submissions against actuals. | Whether the business can predict itself. Persistent optimism and persistent sandbagging are different problems with different fixes. |
| Retention & expansion | Customer-level revenue by period for 3 years, with start date, segment, product and churn date where applicable. | Cohort behaviour, concentration, real net and gross retention rather than the version in the deck. |
| Retention & expansion | Churn reasons as recorded, plus the renewal calendar for the next 18 months with owner and status. | Whether retention is managed or merely occurring, and how much of next year is already at risk. |
| Retention & expansion | Product usage or consumption data by account, where it exists. | Which contracted revenue is genuinely embedded and which is dormant. The most under-requested item on this list. |
| People, data & cadence | Commercial org chart with tenure, plus the comp plan and its history. | What behaviour the business is actually paying for, which is frequently not the behaviour in the strategy. |
| People, data & cadence | The standing meeting calendar for the commercial function, and the last 3 pipeline review packs. | Whether there is an operating rhythm. How a team reviews its pipeline reveals more than what the pipeline contains. |
| People, data & cadence | System landscape: CRM, marketing automation, CPQ, billing, and what is integrated with what. | Whether one version of commercial truth exists. Where it does not, every number you receive has been assembled by hand. |
What the gaps tell you
Some of the most useful findings in commercial diligence are things that do not arrive. These are worth logging explicitly rather than chasing quietly.
- No stage history in the CRM. Opportunities are updated at close rather than worked through stages. The pipeline is a list of hopes and the forecast is a conversation, not a calculation.
- No closed-lost detail. The business does not systematically learn why it loses. Win rate can be improved only by accident.
- Realised price only available in aggregate. Nobody is monitoring discount at the transaction level, which means nobody is controlling it.
- Churn reasons blank or uniformly “price”. Price is what customers say. It is rarely why they left, and a business that records it uncritically has not asked.
- Usage data does not exist. Common and not disqualifying, but it means contracted revenue cannot be tested for durability, and that uncertainty should be priced.
- It arrives fast but hand-built. A single analyst producing everything overnight in Excel is a finding in itself: the commercial function does not have systems, it has that person.
The five questions no data room answers
Documents establish what happened. They rarely establish why, and the why is what determines whether it will happen again. Reserve time for these regardless of how complete the data room is.
Walk me through your last big loss.
Asked of the seller who lost it. The quality of the answer tells you whether the business has a methodology or a set of personal habits.
Who decides when a discount is granted?
Ask the CRO, then ask a seller. The two answers frequently differ, and the difference is where margin leaks.
What changed after your worst quarter?
A specific answer indicates a functioning engine. Market conditions indicates the numbers are weather.
Which customer would hurt most to lose, and what are you doing about it?
Tests whether concentration risk is understood and managed, or merely present.
If I gave you two more sellers tomorrow, what happens?
A team with a working motion can answer in ramp time and expected contribution. A team without one will say it depends.
Send it in two waves
A single hundred-line request produces a slow, defensive response. Two waves work better and cost nothing.
Wave one, day one: the CRM export, customer-level revenue history, and the price list with realised prices. Three items. Everything on this list that genuinely changes a view can be derived from those three, and they are the ones with the longest lead time. Ask for them before anything else and chase only these.
Wave two, once wave one is in hand: everything else, informed by what the first wave showed. Roughly half of a standard request list becomes unnecessary once you have read the CRM properly, and the other half becomes sharper because you know what you are looking for.
Management teams under deal pressure respond considerably better to three urgent items than to a hundred equally-weighted ones, and the sequencing itself signals that you know which evidence matters.
Frequently asked questions
What should be in a commercial due diligence data request?
At minimum: a full CRM opportunity export including losses, customer-level revenue history by period, and price list against realised price. Most other items can be derived from those three or become unnecessary once they have been analysed.
Why request closed-lost opportunities?
Because they show why the business does not win, which is usually more diagnostic than why it does. Loss reasons, competitor presence and the stage at which deals die reveal whether the problem sits in the proposition, the pricing or the execution.
What if the company cannot provide CRM data?
That is a finding rather than an obstacle. It means the commercial function is not run on data, forecasting is judgement-based, and any improvement plan has to start by building the basic instrumentation. It should affect both the plan and the price.
How long should management need to fulfil the request?
Core exports should take days, not weeks, if the systems are in reasonable order. Persistent delay on raw data while polished summaries arrive quickly usually indicates the underlying data does not support the summaries.
Ready to aim higher?
If you would rather not run this yourself under deal pressure, we assess the whole commercial engine across 23 topics and deliver a management-endorsed plan alongside the verdict.