How to Map a Corporate Transaction Process, Step by Step

Mapping a transaction process means writing down, for one specific transaction, the milestones in order, and attaching five things to each: who decides, who must know, what paper it produces, where control leaves the buyer, and what would make the buyer stop. A map that could describe any transaction has described none, and the specificity is what makes it usable.

What is actually being mapped, and where does the map start?

An analyst handed a transaction and asked for a process map is tempted to open a drawing tool and produce seven boxes with arrows between them. Seven boxes with arrows is what a process map looks like in most decks. The harder question comes first: what would somebody do differently tomorrow because this map exists? If the honest answer is nothing, what follows is a picture rather than a tool.

Think about a wedding at household scale. There is a version of the plan that says engagement, invitations, venue, ceremony, reception, and it is correct in every particular and useless to everybody. A second version says which aunt decides the guest list, which cousin has to know the venue before the caterer is booked, and which booking is refundable and which is not. The one condition that would move the whole date is the hall not confirming. The second version is the same sequence carrying attachments. A process mapA written record of the steps of one particular project, in order, with the information somebody needs at each step attached to it. The record is a working document, not a picture. is the sequence plus what is attached to it, and without the attachments it is only a diagram.

The map that follows is built for one transaction. Harivansh Packaging Limited, an invented maker of packaging, is buying 100 per cent of Sundarban Polymers Private Limited, an invented polymer producer. Enterprise value is Rs 1,320 crore, struck as 10.0 times Sundarban Polymers' earnings before interest, tax, depreciation and amortisation (EBITDA) of Rs 132 crore. After deducting Sundarban Polymers' own net debt of Rs 180 crore, the equity value paid to the sellers is Rs 1,140 crore. The bridge from enterprise value to equity value is covered separately and is quoted here only because the map has to carry the number the transaction was struck on.

The record locks seven milestoneA named point in a project that either has happened or has not, with nothing in between. A milestone is a state, not an activity. A state can be ticked, and an activity cannot. points in order: approach and confidentiality, indicative offer and term sheet, confirmatory diligence, documentation, signing, the conditions period, completion. The record also locks two spans and only two. Twenty two weeks ran from the term sheet to completion, and the conditions period was nine of them. Everything else about elapsed time on this transaction is unlocked, and the map says so rather than filling the gap.

What is the procedure, in order?

Seven steps. Steps two to six each attach something to the list that step one produced, so the order matters. A step that attaches to the wrong list attaches to nothing. Step seven is the one almost everybody skips, and it is the one with the highest value per minute spent on it.

1. List the milestones for this transactionNot the milestones a transaction usually has. Strike out any this one does not have, and add any it does.
   Checking: could somebody who has never seen this transaction tell it apart from another one by reading the list alone?
2. Attach who decides at each milestoneNot who is involved. Who decides. One name per milestone, written down.
   Checking: is there exactly one name against each milestone, and does that person know it?
3. Attach who must know at each milestoneThen check the list against the maintained record of the people holding the information.
   Checking: do the two lists agree, name for name, at every milestone?
4. Mark every point at which control leaves the buyerAn approval, a consent, a counterparty's decision, or a state of the world.
   Checking: is every span on the map labelled as one effort can move or one it cannot?
5. Attach the paper each milestone producesThe artefact that will exist afterwards and can be pointed at.
   Checking: is there a milestone with nothing in this column, and if so, how would it be shown to have happened?
6. Record the conditions and who grants eachWith a named person on the buyer's side accountable for every one of them.
   Checking: does every condition have somebody inside the transaction chasing it, including the ones nobody inside can grant?
7. Write down what would make the buyer stopNot what would be disappointing. What would end it.
   Checking: is this written before the term sheet is signed, and is it specific enough to be tested against a finding?

Step one: which milestones does this transaction actually have?

Start with the seven the record locks, and then do the part that turns a template into a map. Go through the list and ask, of each one, whether this transaction has it. Then go through everything the list does not carry and ask whether this transaction has that. A map that matches the general case matches nothing in the transaction at hand, and the striking out is exactly as important as the listing.

On this transaction all seven survive. Surviving is a real answer and not a failure of the exercise. The general shape does not carry the completion adjustments the record locks after completion, and this transaction has all three of them: a working capital peg, a net debt adjustment and an earn-out, all settled in the paper and all arriving after the day everybody thinks the transaction ended. A map that stops at completion tells the team the work is over on the day it is not, so adding that eighth point to the map is the whole point of step one.

Anything the record does not lock comes off. A general diagram might carry a shareholder vote, or a separate financing close, or an integration phase. None of those is locked here. Silence in the record does not mean this transaction has none of them. Silence means only that nobody has written down that it does, and a map that shows a step nobody can describe is worse than a map that shows the gap. Draw the gap.

Step two: who decides at each milestone?

Now attach a name to each point. Not the people involved. At any milestone the people involved are a crowd. Attach the single person whose decision moves the transaction from that milestone to the next one. Step two is the shortest to write and the one that produces the most argument. The argument is a sign it was worth doing.

On this transaction the record locks two people and a clean split of authority. Devyani Kulkarni, chief financial officer of Harivansh Packaging Limited, decides anything that changes the price or the funding. Ashwin Rege, who leads the transaction team, decides how the process is run day to day. Push every milestone through that split and each one lands on exactly one of them. The term sheet sets the price, so it is hers. Confirmatory diligence is a question of how the work is run, so it is his. Documentation ends in an agreement that fixes price and funding, so it is hers again. Signing is hers. The conditions period belongs to nobody inside the transaction at all, and step four is where that finding is put to work.

A milestone with two decision rightThe authority to settle a particular question, held by one named person. A decision right is separate from being consulted, from being informed, and from doing the preparatory work. holders has none, because it will be decided by whoever moves first rather than by whoever should. That sounds like a slogan until it happens. Two names against the same box means that on the afternoon the answer is needed, either both defer to the other and nothing happens, or one acts and the other finds out. Neither is a decision anybody chose. The value of step two is that it surfaces the disagreement about authority in a quiet room in week one, when settling it costs a conversation, rather than in week eleven when it costs the timetable.

Step three: who must know, and how is that list checked?

Every milestone changes what is known and by whom. Attach that. At the approach the transaction team knows and nobody else does. The lenders being sounded have to know before the funding is committed, so they go on the map at documentation and not before. The two counterparties whose contracts change hands are the ones being asked to consent, so they have to know during the conditions period. Everybody else knows from the announcement onwards.

Then do the part that turns this from a courtesy list into a control. The who-must-know column on the map and the maintained record of the people holding the information are the same list, and a difference between the two is itself a finding. Suppose the map shows five people knowing at confirmatory diligence and the record shows eleven. The map is the document being edited, so the instinct is to fix the map. The instinct is wrong. Six people are holding transaction information that nobody planned for them to hold. The map has disagreed with reality in a way somebody has to explain, and that is the most useful thing a map can do.

The household version runs the same way. Somebody tells four people in confidence about a change of job. A month later a neighbour mentions it. The interesting number is not four. The interesting number is the gap between four and however many people actually know now, and the gap is where the information travelled without anybody deciding it should. The Securities and Exchange Board of India (SEBI), at sebi.gov.in, sets what may and may not be done with unpublished information about a listed company's transaction.

Step four: where does control leave the buyer?

Step four changes how the transaction is managed, and it takes about ten minutes. Walking the milestone list means marking every point at which the next move belongs to somebody outside the buyer. There are four kinds: an approval from an authority, a consent from a counterparty, a decision by the other side, and a state of the world that simply is or is not the case.

On this transaction there are three such points, and every one of them sits inside the conditions period. The first is a regulatory approval. The second is consents from the two counterparties whose contracts change hands. The third is the absence of a material adverse change. Nobody decides it. A material adverse change is a condition about the world, and no amount of chasing moves it. The record does not name the authority or either counterparty, so the map carries all three by function and says so rather than inventing names for them.

Marking the control boundaryThe point in a project where the next move belongs to somebody outside it. Before the boundary, more effort produces more progress. After the boundary, it does not. points splits the timeline into the part that responds to effort and the part that does not, and that split is the single most useful thing a map produces. On this transaction the split is stark. Thirteen weeks ran from the term sheet to signing, and everything in them responds to effort: more people on diligence, more hours on drafting, faster turns on comments. Nine weeks ran from signing to completion, and adding people to those nine weeks does not shorten them by a day. A timetable that treats those two spans as the same kind of time will always be wrong in the same direction.

The everyday version is a passport. The applicant controls how fast the form is filled and how quickly the documents are assembled, and after submission controls nothing. Doubling the effort in week one moves the date. Doubling it after submission moves nothing at all, and the only sensible thing to do with the second span is plan around it rather than push at it.

Step five: what paper does each milestone produce?

Attach the artefact. Every milestone on this transaction leaves something behind that can be pointed at afterwards: a confidentiality undertaking at the approach, a term sheet at week zero, a data room and its access log through confirmatory diligence, the agreement in draft through documentation and signed at week 13, and at week 22 the completion deliverables together with the completion accounts. The contents of each of those papers are settled where that paper itself is set out. Step five attaches the name of the artefact without reopening what is inside it.

The paper is the evidence of the milestone, and a milestone that produces no paper cannot later be shown to have happened. That is not a filing-cabinet argument. Six months after completion, somebody will ask when a particular person first had access to the target's customer contracts. The honest answer lives in the access logThe running list of who opened which document and when. The log exists so that a question about who saw what has an answer that does not depend on anybody's memory., not in anybody's memory of the meeting. If the diligence milestone had produced no artefact, the question would have no answer at all, and a question with no answer in a transaction is a question that gets answered by whoever remembers most confidently.

So the useful thing about this column is not the seven filled cells. The useful cell is the empty one, where there is one. A milestone with nothing in the paper column signals either that it is not really a milestone, in which case it belongs to step one and should come off, or that something needs to be produced there and nobody has said what.

Step six: who grants each condition, and who chases it?

Three conditions stand between signing at week 13 and completion at week 22 on this transaction: a regulatory approval, consents from two counterparties whose contracts change hands, and the absence of a material adverse change. Write each one down and put two names beside it. The first is the grantor, and on all three of these the grantor sits outside the transaction entirely. The second is the condition ownerThe person inside a transaction who is accountable for a particular condition being satisfied, whether or not they are the person with the power to satisfy it., who sits inside it.

A condition granted by somebody outside the transaction still needs somebody inside it who is accountable for chasing it, and the conditions without an internal name against them are the ones discovered late. The logic is uncomfortable and worth stating plainly: the reason a condition has no internal name is usually that nobody inside can grant it, and people are reluctant to be accountable for something they cannot deliver. But accountability for chasing is not accountability for granting. Ashwin Rege cannot make an authority approve anything. He can be the person who knows, on any given Tuesday of those nine weeks, exactly where the approval stands and what was last asked for.

The material adverse change condition is a different kind of condition, so it takes a different internal name. Nobody grants it. The condition is about the state of the world, tested against the case the transaction was struck on, and the person who holds the price and funding case is Devyani Kulkarni. The authority granting an approval sets what that approval requires, and the routing for a purchase of this kind sits with SEBI at sebi.gov.in and with company law at the Ministry of Corporate Affairs at mca.gov.in. The record for this transaction does not name the authority or either counterparty, so the map carries all three conditions by function and says openly that it is doing so.

Step seven: what would make the buyer stop?

The last column, and the one almost every map omits. Before anything is signed, the map records the findings that would end the transaction rather than merely disappoint the buyer. The last column is not a risk list. A risk list says what might go wrong. A stopping ruleA statement, written in advance, of the specific finding that would end a transaction rather than merely change it. The statement is written to be tested against a fact, not discussed. says what fact, if found, ends it.

Written as this transaction would have written them at week zero, there are three.

The first: a diligence finding that changes the EBITDA the Rs 1,320 crore enterprise value was struck on. The Rs 1,320 crore is 10.0 times Sundarban Polymers Private Limited's Rs 132 crore of EBITDA, so the multiple and the price move together with that one figure. At week zero nobody knows what a finding will look like, so the rule does not say by how much the EBITDA would have to move. The rule says that the EBITDA is the figure the whole price stands on and that a finding against it is a stopping question rather than a negotiating one.

The second: a consent that cannot be obtained at any acceptable cost. Note the shape of that sentence. The rule is not about a consent that cannot be obtained. A counterparty being asked to consent knows perfectly well that consent can usually be bought for something. The rule is written about the price of the consent rather than its availability, and the price is what makes the rule testable in the conditions period.

The third, and the one worth the most: funding that cannot be arranged at a cost the accretion arithmetic survives. The purchase is already dilutive at the agreed funding, taking earnings per share from Rs 12.50/- to Rs 12.14/-, so the cost of the Rs 1,000 crore of new borrowing bears directly on whether the transaction should proceed at all, and a rule written about it at week zero is worth more than any amount of discussion about it at week eleven.

The arithmetic behind that sentence is settled elsewhere and is quoted here only because the rule points at it. Rs 1,140 crore of equity value is funded with Rs 140 crore of the acquirer's own cash and Rs 1,000 crore of new borrowing at a contracted 9.0 per cent. The 9.0 per cent is the rate this transaction contracted, and half a point either way would move the whole accretion answer with it. The target's contribution and the borrowing's cost have to be struck on the same base to be compared at all. Sundarban Polymers' profit after tax is Rs 61 crore, a rounded figure that computes to Rs 61.35 crore on the exact chain, and Rs 61 crore is 5.35 per cent of the Rs 1,140 crore paid. The Rs 67.5 crore of after-tax interest on the new borrowing is 5.92 per cent of the same Rs 1,140 crore. The spread is minus 0.57 points, or minus Rs 6.50 crore. Spread over 18.00 crore shares, minus Rs 6.50 crore is Rs 0.36/- a share, and Rs 0.36/- takes Rs 12.50/- to Rs 12.14/- exactly. The purchase is dilutive on the rounded figure and on the exact one, so nothing about the stopping rule turns on the rounding.

Why does the timing matter so much? Because a stopping rule written at week eleven is written by somebody who has already spent eleven weeks, sat through the meetings, defended the transaction internally and told people it was happening. Somebody who has spent eleven weeks is not neutral about whether the transaction should stop, and neither is anybody on the other side of the table. Nobody is lying. A looser rule is more comfortable to live with, so the rule simply comes out looser. Write it at week zero and it is a fact about the transaction; write it at week eleven and it is a fact about the people writing it.

The household version is a house purchase. Before any viewing, a buyer can say plainly that a structural finding above a certain cost ends it. After four viewings, two conversations with the bank and an announcement to the relatives, the same finding becomes something to negotiate about. The finding did not change. The amount already spent on the outcome changed.

What does the finished map for this transaction look like?

Set out as rows rather than as a picture, the whole map fits into a single table. Every cell below is filled from the record for this one transaction, and where the record carries nothing the cell says so instead of estimating.

Two things about that table are worth noticing. The first is how many cells say something specific to this transaction and could not be copied on to another one: two named people, three named conditions, two locked spans and a set of artefacts that belong to this deal. The second is where the map admits it does not know. The span before week zero is not locked. The thirteen weeks from term sheet to signing are not split between diligence and documentation. The authority and the two counterparties are not named. An estimate is indistinguishable from a fact once it has been drawn on to the map, so a map that marks its own gaps is more useful than one that estimates across them.

The announcement point deserves a word. The record locks seven milestones and does not fix an announcement date. The map marks announcement against signing, where the who-must-know column stops being a list of names and becomes everybody. Marking announcement at signing is this transaction's own commercial sequencing rather than a rule. SEBI sets, at sebi.gov.in, what a listed company must disclose about a transaction and when, and the resulting filing appears on the National Stock Exchange of India (NSE) at nseindia.com and on BSE Limited (BSE) at bseindia.com.

What has the map not settled, and what does it hand on?

Two more things belong on the finished artefact before it is put to work, and both are about how it is maintained rather than how it is built. The first is that the map is a live document, not a deliverable. Every one of the five attachments can change: a decider changes, the who-must-know list grows, a condition is added. A map that is right at week zero and never touched again is a record of what somebody once expected. The second is that every gap the map marks is a question with somebody's name on it, and gaps close as the transaction proceeds rather than staying gaps forever.

The map does not settle whether the transaction is a good idea. The map carries the price, the funding and the earnings effect because a stopping rule points at them, and it stops there. Whether buying Sundarban Polymers Private Limited at 10.0 times its EBITDA is worth doing depends on what the money would otherwise have done and on what the combined business achieves, and no map settles either. A map makes sure that if the answer turns out to be no, somebody finds out on a week when finding out is still cheap.

A last note on the completion deliverableThe set of documents and items that have to change hands on the day a transaction completes, listed in advance so that nothing is discovered missing on the day itself. column. The completion deliverable column is the one part of the map that people build backwards, from the last day towards the present, and backwards is the right way round for it. Everything else on the map is built forwards from the approach.