Stakeholder Analysis: Who Cares, and About What

Stakeholder analysis is working out, before anything is written, who will read the work and what each of them is trying to protect. Three questions build the list: who decides, who is affected, and who gets asked about it afterwards. The entry point is then tailored for each reader, and the finding never is. Same number, same conclusion, different first sentence.

The hard part comes first. Everything else in stakeholder analysis depends on it being settled. One version of this subject quietly teaches the presentation of the same facts more favourably to whoever happens to be listening. Shading of that kind is not a weaker form of stakeholder analysis but the opposite of it. Tailoring changes where the message begins; the moment a description of the finding changes to suit the reader, tailoring has stopped and shading has started, and the two are separated by a test that takes thirty seconds. The test appears in full below, and every technique above it sits on the safe side of the line.

The mechanism underneath the answer runs like this. People argue about findings far less often than it looks from the outside. When a meeting goes badly the surface reads as a disagreement about a number, but the number is almost never what is in dispute. The dispute is about which question was worth answering, and a disagreement of that kind is about what each person in the room is trying to keep safe. A reader protecting a four year customer relationship and a reader protecting consistency across other borrowers are not disagreeing with the arithmetic. Each of them is waiting for the sentence that tells them whether the work threatens the thing they hold. Given that sentence, they engage with the finding. Denied it, they spend the meeting looking for it.

One assignment carries this guide from end to end. Sharada Iyer, seven months into a role at the Kavery research desk inside Kavery Capital Services Private Limited, produced an assessment of a credit line to Meenakshi Tubes Private Limited. Prakash Nadar reviewed it. Latha Menon commissioned it and chaired the twenty minute item on 12 March in front of six people, four of whom had read the pre-read of three printed sides. Lending, credit and margins are separate subjects, so what Meenakshi Tubes Private Limited manufactures never matters to the stakeholder list. The analyst in the example is assumed to have no authority whatsoever: no ability to call a meeting, overrule anybody, or be believed on their say-so.

Who counts as a stakeholder in a finance assignment?

A stakeholderAnybody who decides on the work, is affected by it, or will be asked about it later. Not a job title and not a rank. is anybody who decides on this work, is affected by it, or will be asked about it afterwards. The definition contains no rank. The definition does not say senior, does not say client, and does not say the people invited to the meeting. Rank is a poor proxy, and the invitation list is worse. The invitation list was written by somebody solving a different problem.

Three questions build the whole list, and the exercise takes about five minutes on the back of an envelope. Who decides, who is affected, and who gets asked about this afterwards: three questions, five minutes, and a list nobody else ever sees. Shortness is the point. A stakeholder analysis that grows into a project is one that will not be done at all on a busy day, and a busy day is exactly when it is needed.

Question one: who decides?

The decider is the person the decision belongs to, and that is a question of fact rather than of seniority. Very often the two coincide and the question can be dropped. Sometimes they do not. A household choosing a school has a grandparent with the strongest view in the house and a parent who signs the form; the form-signer decides. A wedding caterer quoting for two hundred plates is read by four relatives and answered by one. Assuming the most senior person in the room is the decider costs nothing when the assumption is right and costs the whole meeting when it is wrong. The opening will have been written for somebody who is not going to make the call. On this assignment the decider was Latha Menon, who chaired, and the check took one sentence in the corridor rather than an assumption.

Question two: who is affected?

Affected means the decision lands on them: they have to do something differently, explain something, or absorb something. On this assignment the credit line falls from Rs 40,00,000 to Rs 25,00,000, a reduction of Rs 15,00,000. The person who has to say that out loud to a customer they have held for four years is affected in a way that the person deciding it is not. So is the risk colleague, who now has to explain why the same internal floor was applied here and answer whether it is being applied the same way elsewhere. So is Prakash Nadar, whose review pass produced three findings and who will notice whether they were used.

Question three: who gets asked about this afterwards?

The third question earns its place, and it is the one people skip. The question finds the person nobody invited: the colleague who takes a call about it in three weeks, the person covering during a period of leave, the one who has to answer a question that arrives long after the meeting has closed. On this assignment that person was real and arrived four days later. Sharada Iyer took five days of leave on 24 March, and the colleague picking the file up had never been in the room on 12 March. The colleague was never on the attendee list and is a stakeholder by every test that matters. Being the one who gets asked is the whole of the test.

Look carefully at the bottom line of the drawing above. The bottom line carries something people miss. The list has six entries and the meeting had six people, and they are not the same six. Two of the attendees share the same relationship to the work, so they fold into one entry. One entry never attended at all. A stakeholder list is not an attendee list wearing a different heading, and one that comes out identical to the invitation has almost certainly been copied rather than built.

What is each of these people actually protecting?

Having the list is a quarter of the work. The useful part is the second column, and the second column asks what each person is protectingThe thing a person is actually trying to keep safe, which is often quieter and more specific than the position they argue for out loud.. Write four words. Not a paragraph, not a profile, four words, and then stop. The value is in having the column at all rather than in its depth. Every attempt to make it thorough has the same result: the column does not get written.

Protecting is not the same as wanting, and the gap between the two is where most of the useful information sits. The relationship manager on this assignment wanted the credit line kept. Keeping the line is the position, and the position is stated out loud. The relationship manager was protecting something narrower and quieter: not being surprised in front of a customer they had held for four years. The position and the protected interest are close enough to look identical and far enough apart to need different sentences. Nobody can give the relationship manager the line back. The available move is making sure they are never the last person in the building to hear something about their own account.

Nobody is interviewed to fill this column, and asking directly would produce a worse answer than watching would. Three things get read instead. The first is what this person asked about the last time something like this came round. The second is what they argued for when they had a choice. The third is what they had to go and explain to somebody else afterwards. All three leave a trail in emails, in meeting notes and in the questions people ask, and none of them requires standing, seniority or anybody's permission to look at.

Why are interest and influence read together?

Two readings turn the list into something that can be acted on. InterestHow much somebody cares about the outcome of the work. High interest does not imply any power to change it. is how much somebody cares how this comes out. InfluenceHow much somebody can change the outcome. High influence does not imply that they care about it at all. is how much they can change it. Interest and influence are independent, so either one read alone gives a defensible-looking answer that is wrong. The interesting people are the ones where the two disagree.

Take it out of the office first. A street vendor outside one office building cares enormously about whether the building changes its lunch hours and can influence that decision not at all. Ten shops in one shopping centre care intensely about the centre's parking charge and decide none of it. In both cases the caring is real, the power is zero, and the outcome is the same. Saying something has no route to the decision, so they say nothing at the time and raise it later, loudly, when it is expensive to reopen.

High interest with low influence is the group most often skipped and the single most reliable source of the objection that arrives three weeks after a decision. The logic that skips them is superficially sound. The people in that corner cannot stop the decision, so why spend the time. The answer is that they cannot stop it and they can reopen it, and reopening a decision is far more expensive than including somebody in it. On this assignment that quadrant held the relationship manager, who cared most of anybody in the room about the outcome and decided none of it.

Notice the bottom right corner as well. The bottom right corner explains something odd about the meeting. The two others in the room contributed nothing on this item. Low interest, and yet a seat at a committee. A seat is influence whether they use it or not. The bottom right corner is the entire reason a plain language line existed in the pack at all. A reader who is not following cannot object, cannot agree usefully, and cannot be relied on for anything except silence, and silence in a committee is not consent to anything.

One warning about the grid matters more than the grid does. Nothing on it is measured. A stakeholder grid has no score, no weighting and no total, and one with numbers attached to people carries those numbers as decoration. The positions are judgements about one assignment made by one person in five minutes, and their whole value is that they were made at all.

How to Tailor a Finance Message to Its Audience, and What Must Never Change?

Here is the method in full, in one rule. TailoringChoosing where the message begins for a particular reader. The finding that follows is word for word the same one everybody else receives. changes the first sentence and nothing else. The finding is the finding. The numbers are the numbers. The conclusion is the conclusion, and the words describing all three are chosen once, before the reader is known, and then never touched again. The choice made per reader is the entry pointThe first sentence of a message, chosen so that the reader recognises the subject as one that touches the thing they are protecting.: the sentence that tells them, in their own terms, that this work touches the thing they are protecting.

Why does the first sentence carry so much? Because a reader spends it deciding whether this is about them. The decision is made fast and it is made once. An opening on a margin, put to somebody who is protecting a customer relationship, files the whole thing under work somebody else has to worry about. Their attention then has to be won back later, in front of other people, and no worse place for winning it back exists. An opening where they already are brings the same reader to the margin sixty seconds later of their own accord, having brought themselves.

The mechanics are unglamorous. The four words from the protecting column become one sentence that names that thing and states the fact that touches it, with the identical finding pasted underneath. Three readers means three of these; it does not mean three documents, three versions or three sets of numbers. On this assignment the same pre-read of three printed sides went to all six people, and the tailoring lived in how the work was opened when each of the three was actually spoken to.

Where is the line between tailoring and shading?

ShadingChanging the description of a finding to suit the person listening. Every individual statement can be accurate and the practice is still shading. is changing the description of the finding to fit the listener, and it does not require saying anything false. Saying nothing false is precisely what makes shading dangerous to teach around. Suppose the fall of 1.2 points becomes a modest easing for the relationship manager and a material deterioration for the committee. Both sentences sit on the same 10.8 per cent. Neither is a lie. And modest and material are judgements rather than measurements. On those two occasions the judgement was made by the audience rather than by the evidence.

The test is mechanical and takes thirty seconds: the two versions go side by side and are read line by line, and if anything other than the opening has changed, that is shading. Not a feeling, not a matter of intent. Two documents, one reading, one question. Has a claim changed, or only where I began. The reason a mechanical test is worth having is that intent is a very poor guide here. Almost nobody sets out to shade. Shading happens by drift, one accommodating word at a time, each word chosen to make a conversation go more smoothly.

There is a second reason the line matters, and it has nothing to do with ethics. Shading is a bad plan on its own terms. Shading requires keeping track of which version went to whom, forever, in a place where people talk to each other constantly. Tailoring requires remembering nothing at all. There is only ever one finding, and the openings are disposable.

What does a three minute conversation before the meeting buy?

The highest return move available to somebody with no authority costs three minutes and needs nobody's permission. The person in the high interest corner is approached before the meeting and told what they are going to see. Not to persuade them. Not to soften anything. To make sure that the first time they hear a finding about their own account, they are not hearing it in front of five other people.

The whole conversation fits in three sentences. The first is what the work found. The second is the part that touches whatever that person has been carrying. The third is the question they will most likely want to ask, put to them with an invitation to ask it now rather than be surprised later. The third sentence does the heavy lifting. Inviting the objection instead of hiding from it turns it into an objection somebody has already thought about.

Three minutes beforehand converts a surprise into an objection the room has already half heard, and there is nothing else available at that price. On this assignment the relationship manager did object, out loud, in front of six people. The relationship manager asked whether a fall of 1.2 points really justified touching a four year relationship. The question was not really a question, and the honest answer underneath it was a reluctance to be the one telling the customer. Because of three minutes, the room had heard the outline of it already, the objection landed as a professional disagreement rather than as an ambush, and nobody in the room was learning about it for the first time while deciding.

One thing this conversation is not. The three minutes are not a negotiation, and not a chance to find out what the other person would prefer the finding to be. The conversation tells them what the work found. If the three minutes ever turn into a discussion about whether the finding could be put differently, the conversation has gone somewhere it should not, and the correct move is to say what the analysis covered, say what it did not, and stop.

What happened when six people needed three first sentences?

Six people sat in the twenty minute item on 12 March. Four of them had read the pre-read of three printed sides. Two had not, and one of the two was the relationship manager. Nine slides went in and twelve of the twenty minutes went on one of them. Stakeholders group by what they are protecting rather than by headcount, so the list built beforehand had six entries and grouped into three. The openings that actually went out are set out below.

The bottom row rewards a second and a third reading. The bottom row is the same for all three readers, word for word, and it is where every number lives. Three openings went out and one finding went out, and not a single figure differed between the three: the same 10.8 per cent, the same 12.0 per cent, the same 11 per cent floor and the same 14 per cent. The openings are not claims. The openings are doors into the identical room.

Now the fourth group, the one that is easy to write off. Two of the six contributed nothing at all on this item and one of them had not read the pre-read. The two silent readers are the entire reason a plain language line existed in the pack: Rs 10.80 left from every Rs 100/- of sales, against Rs 12.00 the year before. The plain language line contains no new number. The line is 10.8 per cent and 12.0 per cent restated in units a person can picture without doing arithmetic in their head while somebody talks. A reader who is quietly lost is not neutral in a committee, and the plain language line is what stops quiet readers being manufactured.

The three minutes went to the relationship manager. The objection that came back in the room was about whether a fall of 1.2 points justified touching a four year relationship, and it came back as a question the room had already half heard. Five questions were asked in the twelve minutes altogether. Three were answered straight from the source log by item number. One could only be answered with I do not know, and I will confirm by Thursday, and the confirmation went by email on Thursday with the item number attached. The fifth was the objection.

How does a lender, an analyst or a household use this?

The move generalises well beyond a research desk, and it is worth seeing it in three or four settings so it stops looking like a technique for meetings. In every one of them the pattern is the same: one finding, several readers protecting different things, and a first sentence chosen per reader.

In every row the finding is unchanged and only the door is different. The unchanged finding is the whole of the method and the reason the method survives being used in public. Notice too that in none of the rows does the reader have to be senior, junior, difficult or easy. The sentence is driven by what they are protecting, and that is knowable without knowing anything about them personally.

One closing observation from the same assignment shows the analysis paying off in a place nobody planned. Four action items came out of the meeting, each with one named owner and one date, and two of them were Sharada Iyer's. Both closed on their dates. Both closings were possible partly because the stakeholder list had already named who would be asking about each item. A named asker turns an action item from a task into a specific person expecting a specific thing on a specific day. The colleague picking up the handover on 24 March was on that list too, and had never been in the room.

The Kavery research desk, Kavery Capital Services Private Limited, Meenakshi Tubes Private Limited, Sharada Iyer, Prakash Nadar and Latha Menon are invented.
Educational material. Not advice on any investment, tax, budget or market position.