Clients ask "how long will diligence take" like it's one number. It isn't. A commercial due diligence (CDD) workstream is a series of distinct workstreams, each with its own pace, and understanding that shape is the difference between a proposal you can defend and one you eat.
Here's what a typical CDD actually looks like, day by day: the honest version, not the "we'll have it in two weeks" version. These are planning numbers, not a promise; every deal distorts them.
Day 1–2: Data collection and the information request
The work doesn't start with analysis. It starts with getting the data, which is always slower than anyone budgets. The information request (RFI) has to be specific enough that the target company can actually answer it: financials, customer lists, pricing, contracts, org charts, and (for technology deals) codebases, infrastructure, and licensing.
This is also where you set the tone. A good RFI signals you know what you're asking for. A bad one gets a polite version of "please clarify," which costs a week. The RFI is the most important document in the first phase, and it's usually written last-minute. It deserves the opposite.
Day 3–5: The baseline model
Before you can stress-test anything you need the as-is financial picture. This is the part analysts actually spend most of their time on: cleaning the data, normalising one-off items, building the base model. The model isn't the deliverable: it's the scaffolding. But it's the part that can't be skipped, because every later chapter reads off it.
Two things make or break this phase. First, a single source of truth for the numbers: one workbook everyone reads from, not five versions drifting in email. Second, written normalisation decisions: what you treated as one-off, why, and what happens if you're wrong. That note is worth more than the model itself when the seller pushes back.
Day 5–8: Market and competition
This is where the two-week engagements usually start to slip. Sizing the market, mapping competitors, and understanding the growth drivers takes real research, and the quality of the answer is directly proportional to the quality of the sources. This is also where AI can help most and hurt most: see why AI fails in consulting for the failure modes. The market chapter is where a fabricated statistic gets written into a deck and never questioned, because everyone assumes someone else checked it.
The discipline that saves you: every number in the market chapter carries either a source you can open or an assumption you can state. No orphan numbers. That rule turns the chapter from the risky one into the defensible one.
Day 8–11: Customer diligence
If the market chapter is where the number gets invented, the customer chapter is where it gets checked. Interviews, callbacks, and reference checks take calendar time: you can't compress other people's schedules. But the synthesis of those calls, the pattern-spotting across interview notes, is exactly the grunt work that AI can scaffold well.
The output of this phase isn't a transcript dump. It's a set of themes with the evidence behind them: "three of the five largest customers say pricing is up for renegotiation" plus the quotes and call notes. Themes without evidence are gossip. Evidence without themes is a filing cabinet.
Day 11–13: The deliverable
The last two days are assembly and review. Chapters get merged, the storyline gets sharpened, and (if the process worked) every claim in the deck points back to a source or an assumption. This is where verification lives. The deck that goes out is only as good as the discipline applied in days 3 through 11.
The review matters more than anyone wants to admit. This is the point where fabricated citations survive or die, and it's the point most likely to be rushed when the deadline lands. The teams that ship clean decks are the ones that protect the review day like it's the most expensive day of the project: because in terms of reputation, it is.
Where the time actually goes
The honest split is roughly:
- ~40% on data, models, and analysis
- ~30% on research and sourcing
- ~20% on customer interviews and reference calls
- ~10% on assembly and review
Most of the automation opportunity is in the first two buckets: the analysis and the sourcing. The customer calls can't be sped up. The review can't be skipped. But the research-and-model chunk can go from days to hours if the tooling is right.
What automation does and doesn't change
Automation compresses the analysis and sourcing days. It does not (and should not) compress the verification. If anything, the opposite: the faster you generate, the more important it is that every number and every citation carries a source you can open.
That's the practical trade-off, and it's the one the AI-enabled firms are getting right. They use the tool to buy back the two days of tabbing between sources, then spend that time on the review and the storyline: the parts that actually win the deal.
The questions that kill or save a diligence
A CDD deck lives or dies on a handful of decisions, and it's worth naming them up front because they shape how the days above get spent:
- Is growth real or bought? Organic growth compounds; growth from price cuts or channel stuffing reverses the moment you stop paying for it. Every growth number in the deck should be traceable to what drove it.
- How concentrated is the revenue? A single customer at 30% of revenue changes the whole risk picture. The customer chapter isn't a courtesy: it's the answer to the question the buyer is actually asking.
- What's recurring, and what's one-off? The baseline model's normalisation decisions are where sellers push back hardest, because one-off revenue being reclassified as recurring is the classic inflation move.
- What happens if the thesis is wrong? The best CDD decks carry a downside scenario the buyer can stress. Teams that skip it are betting the deal survives contact with reality.
Why the RFI deserves an afternoon, not a half-hour
Most RFIs fail because they're written as requests for documents, not as a plan for the analysis. A document list tells the target what to send. A good RFI tells them what you're trying to figure out, so they can send the right things in the first pass.
Structure it around the chapters, not the file cabinets: "to understand revenue quality, we need X, Y, Z." That framing does two things. It stops the back-and-forth that eats a week, and it tells the client you've done this before, which is worth more than any discount in the proposal.
When the answer is "we don't know"
The hardest moment in any diligence is the honest one: after the research, the picture is genuinely ambiguous. The professional move is to say so with a structure: here's what the data says, here's what it can't tell us, here's what would resolve it: rather than to force a false precision into the deck. Buyers respect the first and punish the second, usually at the reference-check stage if not sooner.
Watch the commercial due diligence demo to see the full workstream scaffolded end to end, or start a market sizing on your own to see how fast the sourcing phase can be when the structure is already built.
Get new posts and free tools
Join the list: one email when we publish. No spam.