Due Diligence: The Four Workstreams and What Each Tests

Due diligence on a private transaction runs four workstreams. Commercial tests whether the demand is real, financial tests whether the numbers say what the seller says, legal tests what is owed and held and disputed, and operational tests whether the business can be run. Each found one specific thing on the transaction that became holding 4 of Nilgiri Growth Partners Fund II, invented, and not one of the four stopped it.

Start with the thing almost every explanation of this subject gets slightly wrong in its first sentence. Due diligenceChecking what a buyer or an investor has been told, before money moves. is not a hunt for something bad. Diligence is not a test the seller passes or fails, and the people running it are not looking for a reason to walk out of the room. Diligence replaces what a buyer was told with what a buyer has checked, and most of what it finds changes the price, the conditions or the documents rather than the answer. That single reframing decides everything that follows, because once a review stops being read as an accusation it can be seen for what it actually is: four different kinds of person, asking four different kinds of question, about the same business, in the same eight or ten weeks.

Whose business is being checked here, and who is doing the checking?

One phrase has to be nailed down before a single workstream is named. The trap in this whole subject sits in that phrase, and it catches experienced people. The phrase operational due diligence names two completely different exercises. Not two flavours of one exercise. Two exercises, with different subjects, different runners, different questions and different consequences, sharing a name by accident of history.

Take it out of finance for a second. A household buying its first flat does two quite separate kinds of checking, and nobody confuses them because they have different names in ordinary speech. The household checks the flat: the seepage on the north wall, whether the lift actually works, what the society dues really are. And it checks the person handling the paperwork and the money: is this agent registered, does the money sit in an account the household can see, does one person handle the whole thing on their own. Both are checking. Only one of them is about the flat.

In private markets both of those are called operational due diligence, and a sentence that does not say which one it means has said nothing at all. On one side, a fund runs it on a business it is buying, where it is one workstream of four and it asks whether this company can actually be run and whether the fund can sit on its share register without going blind. On the other side, an investor runs it on the manager it is thinking of committing money to, where it is not one workstream of anything: it is close to the whole review, and it asks who can move cash, who strikes the value, and who audits. The phrase carries both meanings, and a sentence using it has to name the side it is on.

What is diligence actually for, if it is not a search for reasons to stop?

Consider buying a second-hand scooter from somebody in the next lane. The buyer listens to the engine, looks at the papers, and checks whether the person selling it is the person named on those papers. Almost none of that ends in walking away. A buyer of that scooter usually ends up paying a bit less because the tyres are gone, or telling the seller that payment follows the signing of the transfer form. The checking did not find a reason to stop. The checking found three or four small things, and those things moved the price and the conditions.

A private transaction works the same way at a much larger scale, with specialists instead of the buyer's own ears, and with the findings written down. The buyer arrives holding a story: the seller's account of what the business sells, to whom, at what margin, with what obligations attached. Every workstream exists to convert one part of that story from something the buyer was told into something the buyer has seen. Where the two match, the workstream reports that they match, which is a real result even though it reads like nothing happened. Where they do not match, the gap goes into a findings register, and the transaction absorbs it somewhere.

The distinction matters for how any completed review is read. A review with no findings is not a good review, and a review with many findings is not a bad transaction. The number of findings shows how much checking was done and how complicated the business is, and it shows nothing else at all. Each of the four workstreams tests a different part of the story the buyer was told, and on one invented transaction each of the four found exactly one thing.

What are the four workstreams, and why does a review need all four?

A workstreamOne strand of a review, run by one kind of specialist, with its own question set. is one strand of the review, run by one kind of specialist, with its own question set and its own report at the end. The four are commercial, financial, legal and operational, and the reason there are four rather than one is unglamorous: those four kinds of checking need four genuinely different kinds of person, and none of them can do another one's job.

The commercial people talk to customers and competitors and look outward. The financial people take the reported numbers apart and look inward. The lawyers read every contract, register and dispute file. The operational people walk the building, sit with whoever closes the books, and ask who signs what. Put a commercial specialist in front of a charge register and they will read it correctly and notice nothing. The workstream that was cut is precisely the one whose questions nobody else was ever going to ask, so a review that runs three of the four has a gap it cannot see.

Here they are as a set, with the question each one is actually trying to answer, and with what each found on the transaction that became holding 4 of Nilgiri Growth Partners Fund II, invented. The fund entered the position in its Year 2 Quarter 4 at a cost of Rs 50,00,00,000, and added a further Rs 10,00,00,000 in its Year 5 Quarter 2, taking the total cost of the holding to Rs 60,00,00,000. The diligence described here happened before the first of those two dates, and what happens once the investment starts is covered separately.

What does the workstream that looks outward at customers test?

Commercial Due Diligence

A tea stall has stood outside one large office building for eleven years. Every morning it sells four hundred cups, and every one of those cups is bought by somebody who works in that building. The stall is profitable, the owner is skilled, the tea is genuinely good. The only question that matters to a buyer of the stall is what happens if that office moves. Nothing about the tea changes. Nothing about the owner changes. The demand changes, and the demand was never really the stall's own.

The tea stall is the commercial workstream in one image. Commercial is the strand that looks outward rather than inward, and it asks whether the demand this business reports is real demand, where it comes from, whether it would still be there under a different owner, and whether the position the business holds is defensible against somebody who wants to take it. The people who run it spend their time on the telephone to the business's customers, its former customers, and the firms it competes with, and their report is mostly other people's sentences rather than the seller's.

On the transaction that became holding 4 of Nilgiri Growth Partners Fund II, invented, the commercial workstream produced one finding, and it is worth reading it exactly as written rather than as a verdict. The workstream found that the largest single customer accounted for 31.0 per cent of the revenue the business had reported for the year before completion, and that the contract with that customer had fourteen months left to run. Both of those are facts, one about a denominator and one about a calendar, and the commercial workstream's job ended when it established them.

The finding is not calling 31.0 per cent a problem, and it is not calling fourteen months short. Whether either of those things matters depends on things this record does not hold, such as how long that customer has been there and what it would cost it to move. The finding takes a sentence the buyer had been told, that the customer base is broad, and replaces it with a measured share and a date. The replacement is the whole of the work.

Do the numbers say what the seller says they say?

The financial workstream is the mirror image of the commercial one. Where commercial looks outward at customers and competitors, financial looks inward at the reported figures and takes them apart. The financial workstream is not an audit and is not trying to be one. An auditor's signature, and what it does not reach, is covered separately.

The financial workstream separates the part of the reported result that will happen again from the part that will not. A buyer paying a multiple of earnings is buying the repeating part. Anything sitting inside the reported figure that arrived once, for a reason that will not recur, is money the business genuinely received and is not money a buyer is buying. The quantity left after that separation is the run rateWhat the numbers say once anything that will not repeat has been taken out..

On the transaction that became holding 4 of Nilgiri Growth Partners Fund II, invented, the financial workstream found that Rs 2,40,00,000 of the reported earnings before interest, tax, depreciation and amortisation was a one-off recovery. Nothing had been misstated: the reported figure was right, and it simply contained something that was not going to happen twice. The run rate was therefore lower than the headline by that amount.

One honest gap has to be flagged here, and most write-ups quietly fill exactly this kind of gap. The invented record holds the Rs 2,40,00,000 but does not hold the headline figure it sat inside, so the share of reported earnings the one-off represented is unknown, and no estimate of it stands in its place. A share with no stated denominator is not a fact, and a plausible invented denominator would read exactly like a real one.

What is owed, what is held, and what is disputed?

Legal Due Diligence

Take the flat again. Before the money moves, somebody has to establish three things that have nothing to do with how nice the flat is: whether the seller can actually sell it, whether anything is charged against it, and whether anybody is arguing about it in a court somewhere. A beautiful flat with a live dispute over its title is not a beautiful purchase. The legal workstream is that, applied to a company.

Its three questions are what is owed, what is held, and what is disputed. Owed covers every borrowing, guarantee, lease and obligation the company has taken on, including the ones that sit off to the side of the balance sheet. Held covers what the company actually has: its shares and who they belong to, its properties, its licences, its intellectual property, its key contracts and whether any of them fall away when control changes. Disputed covers litigation, tax matters, employment claims and anything else where a third party is contesting something.

The lawyers running it read documents rather than talk to people, and their raw material includes public registers. In India, a company's board, its charges, its filings and its constitutional documents ultimately sit with the Ministry of Corporate Affairs at mca.gov.in, and a legal workstream on an Indian target will always compare what the registers show against what the seller has handed over. The requirements of that ministry, and their deadlines, are set out in the current text at mca.gov.in.

On the transaction that became holding 4 of Nilgiri Growth Partners Fund II, invented, the legal workstream found an unregistered chargeSecurity over an asset that has not been recorded where it should be.: security given over an asset that had not been recorded where it should have been. The charge was cleared before completion. The transaction did not close until the position had been put right, and then it closed. That is the ordinary life of a legal finding, and it is why the legal workstream so often looks, from the outside, as though it found nothing.

Can the business be run, and can the fund be a shareholder in it?

Operational Due Diligence

The first of the two exercises the phrase names is the operational workstream a fund runs on a business it is buying, and not the review an investor runs on a manager.

The workstream asks two questions that sound like one. First, can this business be run: does it have the people, the systems and the routines that let somebody actually operate it, or is it being held together by two individuals and a spreadsheet nobody else understands? Second, and this is the part outsiders miss, can the fund be a shareholder in it: will this company be able to produce, month after month, the numbers and the reporting the fund needs in order to know what it holds?

The second question is why the workstream exists at all in a private transaction. A listed shareholder receives whatever the company publishes and has no say in it. A fund buying into a private company is signing up to years of monthly packs and board meetings, and if the company cannot close its books, those packs will arrive late, wrong, or not at all. The operational workstream is the strand that finds this out before the money moves rather than eighteen months afterwards.

On the transaction that became holding 4 of Nilgiri Growth Partners Fund II, invented, it found two things. There was no month-end close discipline, meaning the business did not shut its books to a timetable and produce a consistent monthly result. And there was no second signatory on payments, meaning one person could move money out of the company alone. The fund made both of them a condition to completionSomething that must happen before the money is paid, written into the agreement., so the transaction did not complete until a close timetable and a second signatory were in place.

What does a finding actually do?

Asking what a finding does separates people who have read about diligence from people who have sat through one. A finding does not float. A finding has to land somewhere, and there are only a small number of places it can land. Most write-ups skip the destinations entirely. Readers are then left assuming that finding something means stopping something.

There are five destinations, and only one of them is dramatic. A finding can move the price, where the buyer pays less because what it is buying is not quite what it was told. A finding can become a condition to completion, where the money does not move until the thing is fixed. A finding can be written into the agreement as a protection, where the seller carries the risk of something specific if it turns out badly. A finding can change nothing on the transaction at all and instead become a thing the buyer watches for the next several years. Or, rarely, it can end the transaction.

On the four findings from this invented transaction, the destinations were two into watch-afterwards, one cleared before completion, one made a condition, and none into ending it. Two plus one plus one is four, which is the whole set. Across the same manager's whole investment period, only one transaction ever ended on a diligence finding, and that transaction appears below.

How does a review change when an investor is checking a manager instead of a business?

How Private-Fund Due Diligence Is Structured

Everything so far has been on one side of the split: a fund checking a business it is buying. Cross to the other side now. An investor thinking about committing money to a private fund is not buying a company. The investor is handing money to a manager, for ten years, into a vehicle it cannot get out of, on terms it mostly cannot change. The thing being checked is therefore the manager and the machinery around the manager, and the operational element of that review is close to the whole of it rather than one strand of four.

Investor 5 of Nilgiri Growth Partners Fund II, invented, is a fund of fundsAn investor whose own investors' money is committed to other funds rather than to businesses.. Investor 5 committed Rs 50,00,00,000 to that fund. The commitment is 10.0 per cent of the fund's Rs 5,00,00,00,000 of total commitments. Before it committed anything, it ran a review on Nilgiri Alternatives Advisors Private Limited, invented, the manager. The review tested the administrator's independence, the valuation process, who can move cash and how many people it takes to do it, the business continuity arrangement, and whether the auditor had ever issued anything other than a clean opinion.

Reviews like that are not run as one long conversation. Reviews run in stages, and the reason is purely practical: each stage decides how much of the next stage is worth doing. An investor moves from the record, to the process and the terms, to the operations and the people, to the reporting it will actually receive, and each stage is the reason the next one is shorter. By the time the reviewer reaches the reporting stage, it already knows what the manager says it does, so it is only checking whether the paper that will arrive every quarter matches that.

Stage three is where the sequence quietly becomes the point. The operations and the people stage is the one where the reviewer stops reading and starts asking, and it is where this side of the split reaches back to something covered separately: a manager's own written account of a quarter can be checked in some places and not others, so an operational review of a manager asks who signs that letter and on what basis they are able to sign it.

How is the list that review runs on built?

How to prepare a Private-Fund Due-Diligence Question List

A question list is a structure, not a pile. The difference between a structure and a pile is the thing most people who have never built a list get wrong. The temptation is to write down every question that comes to mind and send it. Gaps are only visible against headings, so a pile produces a document nobody can answer in order, nobody can allocate to a colleague, and nobody can check for gaps.

Investor 5's list, invented and belonging to that one investor rather than to anybody else, ran to 84 questions in six headings. No authority requires that list, sets its length, or treats it as sufficient. One investor built it, and its shape is worth looking at.

The operations heading is the largest at 19 of 84, being 22.6 per cent of the list, and it is the heading most readers expect to be the smallest. The reason is worth sitting with. The record can be answered from documents: the numbers exist, somebody sends them, and the reviewer checks the arithmetic. The operations cannot be answered from documents at all. Who does what when nobody is watching has to be asked person by person and then checked against what a different person says.

Headings 5 and 6 together come to 31 of 84, being 36.9 per cent. More than a third of this list is about machinery rather than about results. The 36.9 per cent is the most informative single thing in the table, and it is the opposite of what somebody who has only read about track records would draw up on a blank sheet.

Where do the two exercises with the same name actually part company?

How Operational Due Diligence Differs From Investment Due Diligence

With both sides now described, the difference can be stated rather than asserted. One axis settles the argument fastest. A full side-by-side of the two exercises against the same set of criteria is covered separately.

Here is the axis: the two exercises differ in what a bad answer does. A finding in investment diligence, meaning the four workstreams a fund runs on a business it is buying, is almost always answerable with money or with paper. The earnings contain something that will not repeat, so the price moves. There is an unregistered charge, so it is cleared before completion. There is no second signatory, so a second signatory becomes a condition. Every one of those has a price or a clause attached to it.

A finding in operational diligence, meaning the review an investor runs on a manager it is committing to, frequently has neither. Suppose the review establishes that one person at the manager can move cash on their own. There is no discount that fixes that. The exposure is not to a number but to the possibility of the money not being where it is said to be, and no clause an investor can insert into a document makes that acceptable. The conversation does not get cheaper; it ends.

The difference is why the two exercises feel so different to sit in even though they share a name. One is a negotiation with facts attached. The other is a set of conditions that either hold or do not. Neither is more important than the other: they are two different exercises, run by different people, on different subjects, that happen to be called by the same phrase.

What did the whole programme look like, and what fell away?

Everything so far has been one transaction. Step back and look at the shape of the work that produced nine of them. Over the five-year investment period of Nilgiri Growth Partners Fund II, invented, the manager reviewed 412 opportunities. The manager signed 31 confidentiality undertakingsAn agreement to keep what is shared during a review private.. Signing one is the point at which a seller starts handing over anything real. The manager issued 14 non-binding offersAn indication of price and structure that neither side is yet held to.. The manager went to exclusivityA period in which the seller deals with one buyer and no other. on 11. Exclusivity is the point at which the seller stops talking to anybody else and the expensive part of diligence begins. And it completed 9.

Eleven exclusivities, nine completions, so two fell away. One fell away on a finding in the operational workstream. One fell away because a second bidder paid more, and losing to a higher bid is not a diligence outcome at all.

9 of 412 is 2.2 per cent, and 2.2 per cent is the shape of the work rather than a virtue or a fault. A higher figure would need explaining and so would a lower one, and this record gives no basis for saying what either would mean. The ratio shows where the money and the hours go. Most of the 412 were read and put down in an afternoon, and all four workstreams ran in full on only 11 of them.

What does a review that runs three workstreams cost?

Who actually uses the four workstreams, and how?

Three different people pick the four workstreams up and do three different things with them, and the third of the three is by far the most common.

A fund's deal team uses the four as a budget and a calendar. Before anybody knows what will be found, somebody has to decide how many weeks each workstream gets, which specialists are hired, and what the whole review will cost. The budget is why a workstream gets cut: not out of carelessness, but because a review has a cost and somebody is trying to keep it proportionate to a transaction that may never happen. On the manager described here, all four ran in full on 11 opportunities out of 412 over the fund's five-year investment period. Rationing on that scale is what a budget under pressure looks like.

An investor's reviewer uses the structure in the other direction. The reviewer is not sizing a review. The reviewer is checking that somebody else's review existed. Heading 2 of the 84-question list, the process, is doing exactly that: it asks how an opportunity is sourced and who approves it, and it asks what the pipeline was. The 412, 31, 14, 11 and 9 came from that question in the first place. A manager who cannot describe its own pipeline has not answered heading 2.

An analyst reading a completed diligence report, the most common of the three, uses the four workstreams as a checklist for what is missing. A review reports what it looked at and never reports what nobody was asked to look at, so the most useful question to bring to any diligence pack is which of the four workstreams is not in the folder. Three reports where four were needed look exactly like a complete set until somebody counts them.