The Deal Team: Who Does What on a Live Transaction

A deal team is the named group of people who may know about a live transaction and who do the work on it: a sponsor who decides, a lead who runs it day to day, and a person answering for each of the finance, legal, tax, commercial and people workstreams, with external advisers alongside. Every person added holds information they may not act on, so membership is a control rather than a courtesy.

Start with a picture that has nothing to do with finance. A household is arranging a wedding. Somebody has decided how much will be spent and can call the whole thing off if the money runs out, and that is not the same person who spends the day ringing the caterer. Somebody else knows which cousin has been invited and which has not. Somebody is responsible for the venue. Nobody at all is responsible for the transport. The transport is discovered eleven days before the date, at which point it becomes everybody's problem and therefore nobody's.

The wedding party is a deal teamThe named group of people inside a company who are permitted to know about a live transaction and who carry the work of getting it done., exactly. A transaction is a job of work with a budget, a deadline, several strands that have to finish at the same time, and a secret to keep while it happens. The last of those four is what makes staffing it different from staffing anything else. On an ordinary project, adding a helpful person is a kindness with no cost attached. On a transaction, every person added is a person who now knows something about an unannounced purchase, and knowing it constrains what they may lawfully do for as long as it stays unannounced.

So the team is a control, and the list of its members is the same list as the list of people holding the information, read under a different heading. One sheet of paper, two purposes. The left column is an organisation chart. The right column is a register of everybody whose ordinary freedom of action has just been narrowed. Most descriptions of a transaction team offer the first reading and stop. Both readings are needed. The second reading is what turns a staffing question into a governance question.

Harivansh Packaging Limited, an invented manufacturer listed on both Indian exchanges and making rigid and flexible packaging, is buying the whole of Sundarban Polymers Private Limited, an unlisted maker of flexible packaging films. Devyani Kulkarni is the chief financial officer of Harivansh Packaging Limited and Ashwin Rege leads the transaction team. External advisers appear only as the buyer's advisers and the sellers' advisers.

Who is actually on a deal team, and what does each person do?

Titles differ between one company and the next and functions do not, so answer it by function rather than by title. There are three kinds of seat, and every transaction has all three whether or not anybody has said so out loud.

The first is the sponsorThe senior person who authorised the transaction, who carries the consequences of it, and who has the standing to stop it. Not the person who does the day to day work.. The sponsor is the person senior enough to have authorised the work, who carries the consequences of it, and who can stop it. The power to stop is the test of the seat: if a person cannot say no on a Thursday afternoon and have that no hold, they are not the sponsor, whatever their title says. On this transaction the sponsor is Devyani Kulkarni, chief financial officer of Harivansh Packaging Limited, and the decisions reserved to her are the ones that move the price or the funding.

The second is the lead. Somebody has to run the transaction day to day: chase the outstanding items, hold the calendar, know which of forty open questions actually matters this week, and know where every document is. Ashwin Rege holds this seat. The lead is not a junior version of the sponsor. The lead's job is a different job with a different skill. The sponsor's value is judgement at intervals and the lead's value is attention continuously, so the two seats are rarely held well by one person.

The third kind of seat is repeated: one person answering for each workstreamOne strand of transaction work with a defined thing to finish, such as the funding, the agreement, or the customer position. Each has one person who answers for it.. Finance, legal, tax, commercial and people. Each of those has something specific to close out before completion, and each has one person whose name is against it. Not a committee. One name. When the lead asks where the consents have reached, there is then a person to ask rather than a function to email.

Now the sentence that makes the picture useful. The seats are roles, not people: one person can hold two of them, and the only real defect is a role that nobody holds at all. On this transaction Ashwin Rege runs the work day to day and is also the single named point of contact with the sellers. Two roles sit with one person, and there is nothing wrong with it. A small team where three people hold six roles is perfectly sound so long as each of the six has a name against it.

The transport at the wedding is what is not sound. A role held by nobody is not discovered at the moment it is created. An unheld role is discovered late, usually by the trouble it was supposed to prevent. Nobody was answering for the people workstream, so nobody had thought about what employees are told on the morning of the announcement, and the morning of the announcement arrives on schedule regardless. The moment an unheld role surfaces is the moment its work was already due, so the cost of an unheld role is always paid at the worst moment.

External advisers sit alongside this structure rather than inside it. The mandate fixes what an adviser is engaged to do. Mandates are covered separately. The point to carry forward is only that an adviser does not hold a company's decision right. An adviser can say what is usual, what is defensible and how the counterparty is likely to respond. An adviser cannot say that the price has been agreed. The sentence is not the adviser's to say.

Who decides what, and where does a decision actually get taken?

Most descriptions of a transaction team leave out who may say yes, and the omission determines whether the team works. A list of names shows who is present. The same list does not show who may say yes.

A decision rightA settled statement of which named person or body may take a particular kind of decision, agreed in advance rather than worked out when the decision arrives. is a statement, agreed in advance, that a particular kind of decision belongs to a particular named person or body. Three levels are enough for almost any transaction. The lead settles the running of it inside what has already been agreed: the sequence of the work, who answers a request, what goes into the next call. The sponsor settles anything that moves the price or the funding. The board settles anything that commits the company beyond what the board already authorised.

Here is the household version. Two people are having a bathroom rebuilt. The two agreed at the start that either of them may choose the tiles, that a change of more than a certain size to the total needs both of them, and that borrowing more money to finish it needs a conversation nobody wants to have twice. Nothing about that arrangement is clever. Its entire value is that it was settled on a calm Sunday rather than on the Wednesday the builder rings to say the floor has to come up.

Now the practical test, and it is worth memorising in the form it actually happens in. If a question arrives at eight in the evening and the answer moves the price, the team should already know whose question it is. Not who will find out. Not who will be consulted in the morning. Whose question it is, by name, before the question exists.

Why does the timing matter so much? Because a decision right invented after the question has arrived is not a decision right, it is a rationalisation. In the moment, the pressure is always in one direction. The counterparty is waiting, the call is live, the person on the phone has the most context and the least authority, and every minute of delay feels like weakness. Under that pressure, decisions migrate downwards to whoever is available and upwards to whoever is loudest, and nobody notices until much later that the company agreed to something no authorised person ever agreed to.

The three levels also have to be exhaustive. The requirement is stronger than it sounds. Every question that can arise must fall into one of them. If a category of question exists that fits none of the three, the transaction's next crisis has been found in advance. Advance is the cheapest place to find a crisis. The remedy is to write the fourth level rather than to hope the question does not come.

One more thing the levels do. The levels protect the lead. Ashwin Rege can say to a counterparty, without embarrassment and without stalling, that this one is not his to settle and he will come back within the hour. Coming back within the hour is a stronger position than either guessing or going quiet, and it is available to him only because somebody wrote down in advance which questions were not his.

What are the workstreams, and what does each have to close out?

A list of five words is not a team. Finance, legal, tax, commercial, people is a list of five words. A team is five workstreams that each have something specific to finish on this specific transaction, with a name against it and a date by which it has to be done.

So staff this one properly. Harivansh Packaging Limited is buying the whole of Sundarban Polymers Private Limited. Devyani Kulkarni sponsors the transaction and holds every decision that moves the price or the funding. Ashwin Rege runs it day to day. Underneath them, five workstreams, each with something concrete to close out.

Finance has the funding. The purchase is funded with Rs 140 crore of the buyer's own cash and Rs 1,000 crore of new borrowing at the buyer's own contracted 9.0 per cent. The funding comes to Rs 1,140 crore in all. The total is the equity value the sellers actually receive. The equity value is not the enterprise value of the purchase, and the two are different amounts for a reason set out under equity value and enterprise value. Finance also has to make sure the money is available on the day. Having the money available is a separate problem from having agreed it.

Commercial has the customer position. Sundarban Polymers sells to some of the same customers as Harivansh Packaging. The customer overlap is the reason the transaction exists at all, and it is also the reason commercial cannot be a formality. Somebody has to establish who those shared customers actually are, what each of them buys from each business, and on what terms. A food stall that sells to the workers of one office building has one customer wearing many faces. So does a packaging business, sometimes, and finding that out before completion rather than after is commercial's whole job.

Legal has the agreement and the three conditions that stand between signing and completion: a regulatory approval, the absence of a material adverse change, and consents from the two counterparties whose contracts change hands. Each of the three has to be satisfied before completion, and a condition that is not satisfied stops the transaction. The requirements of any approval regime are set by the bodies named below.

Tax has a pointer. The point for the team is that a tax seat exists and that somebody sits in it. How this purchase is treated for tax is covered separately.

People has the employee communication, and the record of this transaction leaves the content of that message open. The silence is not an oversight to be filled in. The content of the message is simply not part of the record.

Notice what has just happened to the list: it stopped being generic. Any transaction has a finance workstream. Only this one has a finance workstream whose job is Rs 1,000 crore of new borrowing at a contracted 9.0 per cent plus Rs 140 crore of cash. Any transaction has a commercial workstream. Only this one has a commercial workstream whose job is a customer overlap between two packaging businesses. The generic list is a template; the specific list is a team. If a company's transaction plan reads like the template, nobody has done this step yet.

Why do the items that sink transactions sit between two workstreams?

Here is the point the list of five hides, and it hides it so effectively that most weekly transaction meetings are structured in precisely the way that guarantees nobody sees it.

Go back to the worked transaction. Two of its items do not belong to a single workstream, and they are the two most likely to cause trouble.

The first is the working capital peg of Rs 96 crore. On the face of it this is a finance item: it is a number, it goes in the agreement, it adjusts the price. But where does Rs 96 crore come from? The peg comes from a view about how much working capital this business normally runs on, and that view is a commercial judgement about how Sundarban Polymers actually trades: how long its customers take to pay, how much film it keeps in stock, how its year is shaped. A finance person who sets the peg without that understanding has written down a number. A commercial person who understands the trading pattern but never sees the agreement has an understanding that never reaches the paper. The peg is a finance figure resting on a commercial understanding, and it belongs to neither of them alone.

The second is the pair of consents. Two counterparties hold contracts that change hands on completion, and each has to agree. Legal drafts the request, tracks it and knows what happens if it does not arrive. But a consent is not obtained by drafting. A consent is obtained by somebody who knows that counterparty ringing them up and having a conversation about why this is fine. The instrument is legal; the route to it is commercial. Neither workstream can deliver it alone, and each can plausibly believe the other has it in hand.

The consequence changes how a transaction is run, so say it out loud. A join has no workstream to report it, so a weekly meeting that goes round the workstreams in turn will never surface an item that sits on a join. Finance reports on the funding and says nothing about the peg's commercial basis. The commercial basis is not finance's to describe. Commercial reports on the customer work and says nothing about the peg. The peg is a number in the agreement. Everybody has reported honestly. The item has not been discussed.

The fix is not more meetings. The fix is a different question at the same meeting. Instead of going round the workstreams, the lead asks which items need two workstreams to be finished, and asks those two people to answer together, in the same sentence, in the room. Ten shops in one shopping centre share one driver of footfall, and none of the ten shopkeepers reports on it. Somebody has to ask about the thing between them, or it goes unwatched until it moves.

Why does every addition to the team have to be justified?

Because an addition is not free, and the cost is invisible at the moment it is incurred.

Think about how the request actually arrives. The request is never unreasonable. A manager who knows the sector wants to help with the commercial work. A category specialist could answer one question quickly. An analyst could take the model off somebody's hands. A regional manager knows one of the two counterparties personally and could make a call. Each of those is a real contribution, offered in good faith, and refusing it feels petty.

A cost that cannot be named is a cost that will not be weighed, so the cost has to be stated precisely. Each person added is three things at once. Each addition is a person who now holds information about an unannounced transaction and whose ordinary freedom to act on their own account has been narrowed for as long as it stays unannounced. Each one is also another entry on the list of who was told what and when, a list somebody has to keep current. And each one is another person in the room at the moment a hard question needs asking. Another person in the room changes whether the hard question gets asked.

Hence the addition testA one line test applied before anybody joins a live transaction team: name the specific task this person will do that nobody already on the team can do., in one line. Name the thing this person will do that nobody currently on the team can do, and if the honest answer is that they would find it interesting, the answer is no.

The test has a second half that is easier to forget than the first. People must come off the team as well as onto it. When the commercial diligence is finished, the people who did it do not need to stay on the weekly call, and leaving them there is not a courtesy either. Leaving them there keeps them inside a constraint they no longer need to be inside, and it keeps the room larger than the conversation requires.

The addition test is easy to mistake for a judgement about people, so be plain about its limits. The test is not a statement that anybody is untrustworthy. Every single person on that list may be entirely reliable, and the addition test still applies to all of them. The test is about necessity rather than character. A household does not tell four relatives about a hospital appointment because three of them are gossips. The household tells the one who is coming to the hospital.

What does the single point of contact rule actually prevent?

The rule is one line. One named person on each side handles all traffic between the two sides. On this transaction Ashwin Rege is the buyer's named contact, and the sellers of Sundarban Polymers Private Limited name one person of their own. Everything that crosses between the two companies crosses through those two people.

The rule sounds bureaucratic until what happens without it becomes clear. The tangle is worth drawing.

The rule prevents three distinct things, and it is worth separating them because they fail differently.

The rule prevents two workstreams giving different answers to the same question. A seller asks about the treatment of a particular contract. Buyer legal answers one way because they are reading the draft. Buyer commercial answers another way because they are describing what they intend to do afterwards. Neither has lied. The sellers now hold two positions from the same company, and will quite reasonably rely on whichever suits them.

The rule prevents an answer given in a corridor that nobody records. Conversations at the edges of meetings are where transactions actually move, and they are also where a company commits itself without noticing. Single point of contactA rule that all traffic between two sides of a transaction runs through one named person on each side, so that every exchange has one source and one record. discipline does not stop people talking. The discipline stops a conversation counting as an answer unless it went through the named channel and was written down.

And it prevents a counterparty learning something from a person who did not know it was sensitive. The third failure is the quietest of the three and the most damaging. Somebody deep in one workstream mentions, in passing, a fact that is unremarkable to them and highly informative to the other side. The person was never supposed to be in that conversation at all, so nobody thought to tell them that the fact mattered.

The rule is not about control of people, it is about there being exactly one version of what this company has said. A household dealing with a builder learns this fast: if three people in the house can approve a change, the house will be paying for three changes it does not remember approving.

Who does the team hand over to, and what goes with them?

Every transaction team has an end date, and it is completion. Completion is the moment the buyer's side stops being a project and starts being a business, and the people who ran the project are usually not the people who will run the business.

So there is a handoverThe transfer at completion from the people who ran the transaction to the people who will run the business and the integration, along with everything the first group knew., and it goes in two directions at once. Whoever runs the purchased business will live with what was agreed, so part of the handover goes to them. Whoever runs the integration has to deliver what the purchase was supposed to deliver, so the rest of the handover goes to them. Neither of those groups is the transaction team, and neither of them was in the room when the assumptions were made.

Place it on this transaction's own calendar. Twenty two weeks ran from term sheet to completion, of which the conditions period was nine. Twenty two weeks is this transaction's own span, and a purchase of the same shape can run in half the time or twice it. Signing therefore falls at week thirteen, being twenty two less the nine week conditions period. The team runs the whole of that span and hands over at the end of it.

Now the failure the handover exists to prevent, and it is not the failure most people expect. The people who know what was promised, what was assumed and what was deliberately excluded leave the room, and everything they carried leaves with them unless it was written down somewhere a stranger can read it.

Think about what the transaction team actually accumulated over twenty two weeks that is not in the agreement. The team knows which of the sellers' explanations it believed and which it accepted with a shrug. The team knows that a particular customer relationship was described as long standing and that nobody was able to confirm it. The team knows why a figure was set where it was, and knowing why is different from knowing the figure. The team knows which three things were argued about for a fortnight and how the argument was resolved. Signed documents record conclusions and not the reasoning that produced them, so none of that reasoning is in the signed document.

A transaction therefore keeps two written things alongside the paper. A decision log records what was decided, by whom and why, as it happened. And an assumption registerA running list of the things a transaction team took to be true without proving them, so that later on somebody can go back and check whether they held. records what the team took to be true without proving it. An assumption that is never written down cannot be revisited, and nobody afterwards even knows it was an assumption rather than a fact. The register matters more than it sounds.

The household version is a house purchase. Six months after moving in, somebody says the surveyor mentioned the roof. Did they? What exactly did they say? Was it a problem being flagged or a detail being noted? Nobody wrote it down, the surveyor is gone, and the roof is now a question with no answer. A transaction is that, at scale, with money attached.

What changes when a team grows, apart from the number?

Something specific and measurable, and it is not the workload.

Take six people. Six people make fifteen distinct pairs, so the number of pairs who have to stay aligned with each other is fifteen. Now take thirty. Thirty people make four hundred and thirty five pairs. The team has grown five times over. The number of connections that have to be kept aligned has grown twenty nine times over.

Both counts come from the arithmetic of pairs and nothing else. The counting is the reason the experience of a large team feels so different from the experience of a small one.

The gap is what coordination costThe effort a group spends keeping itself aligned rather than doing the work, which rises far faster than the number of people in the group. means. In a team of six everybody hears everything and the gaps are visible in the room, so the team holds what it knows in one conversation a week. A team of thirty cannot do that at any length of meeting. A team of thirty needs process where it used to need conversation: a written status, a defined route for a question, a document that says who settles what, an actual list of open items with names and dates against them.

None of this is an argument for keeping a team artificially small. The argument is for noticing where the coordination cost changes, and for putting the process in before it is needed. A team that adds process at six is bureaucratic. A team that adds it at thirty is already late. The fortnight it takes to build the process is a fortnight the transaction spends uncoordinated.

There is a second loss, and it is harder to see because nothing visibly breaks. In a room of six, somebody will say that the customer concentration argument does not hold up. In a room of thirty, containing two people who built that argument and one person whose manager built it, nobody says it. The meeting is entirely productive. Everybody reports. The difficult sentence goes unsaid, and it goes unsaid for reasons of ordinary human decency rather than any failure of courage. A hard conversation needs a small room, and the sponsor has to keep one.

How does a lender, an analyst or a board member read a deal team?

Three readers, three completely different first questions, and none of them is asking who is talented.

A lender being asked for Rs 1,000 crore of new borrowing at a contracted 9.0 per cent reads the team as evidence about delivery. A lender is lending against a purchase that has to actually complete, and the conditions have to be satisfied for that to happen. So they want to know that somebody is named against the regulatory approval and against each of the two consents, that the funding workstream and the legal workstream talk to each other, and that there is one person who can answer a question about status without convening anybody. A lender is not impressed by a large team. A lender is reassured by a named one.

An analyst, once the transaction is announced, reads the team as evidence about integration. Not the transaction team itself, never visible to an analyst, but its shadow: was there a person answering for the people workstream, and does the company seem to know what it will do with the customer overlap between Harivansh Packaging Limited and Sundarban Polymers? A company that can describe the workstreams of its own purchase has thought about it. One that describes only the price has thought about half of it.

A board member reads the team as a control question, and asks two things. Who may take which decision, in writing, before the decisions arrive. And how many people currently hold information about this transaction, and can somebody produce that list today rather than next week. The two questions are the entire board interest in a transaction team, and both of them can be answered badly in a way that sounds fine.

The household version of all three is the same reflex, and it is worth naming because it generalises. When a household is dealing with anything large and slow, a hospital admission, a house purchase, a court matter, the useful question is never how many relatives are involved. The useful question is whether one person is holding the thread, whether everybody knows who decides what, and whether the things being taken on trust have been said out loud to somebody who can check them later.

What does the same team look like from the governance side?

Turn the sheet over. The same names, read differently.

Earlier in this sequence, materiality settled why the list of people who know about a live transaction is itself a control. The list of people who know and the transaction team list are the same list, plus the small number of people who have to be told without doing any work: a company secretary, perhaps, somebody in a treasury function, whoever has to sign something.

Which means the team has two custodians and they can pull in opposite directions. The lead wants the people who will get the work finished. The governance side wants the smallest number of people consistent with the work getting finished. The two are not the same objective. The addition test is the instrument that reconciles them, and it is the only question both sides accept as legitimate. The lead cannot object to being asked what a person will do. The governance side cannot object to a person who has a named job.

The list has to be current, not roughly current, and that is a maintenance job with a name against it rather than a thing that happens. A list that is a week behind is not a list; it is a memory of a list. Who has to maintain such a record, in what form, and what a listed company must do about it, is settled by the Securities and Exchange Board of India, written as SEBI, at sebi.gov.in.

One last thing about the count. The record of this transaction does not say how many people worked on it. Any number would be read as a norm, and somebody would compare their own transaction against it as though the comparison meant something. The test matters rather than the size: a team of four that passes the addition test on every member is better controlled than a team of forty that never applied it.