Feedback: Asking For It, Receiving It and Using It

A narrow request produces something usable: it names the one thing to be looked at. Feedback is received by writing it down and asking only clarifying questions, never defending. Applying feedback means changing the work and then telling the person what changed. Feedback on finance work is about the work, and it functions only when a person actually made the choices being questioned.

The mechanism sitting underneath that is not the one most people expect. Feedback fails at the request rather than at the delivery. A broad question about nine hours of work hands the reviewer a small research project: they have to read the whole thing, work out for themselves which part the author is unsure about, decide what is worth saying, and then decide how to say it to a person who might take it badly. Twenty minutes of work go by before the reviewer can open their mouth. A broad question has no cheap honest answer. A busy reviewer gives the safest general comment available instead, and both people quietly conclude that feedback is not very useful. Put a narrow question and the same person, with the same amount of goodwill, gives something specific. The narrow question has made specificity the easy option rather than the generous one.

One assignment runs through this guide from end to end, and it is the same one the surrounding material has been following. The Kavery research desk, four people inside Kavery Capital Services Private Limited, produced a piece of work about Meenakshi Tubes Private Limited, a business Kavery Capital Services Private Limited has money out to. Sharada Iyer wrote it, seven months into the role, working alone on a piece for the third time. Prakash Nadar reviewed it and returned three findings. Latha Menon commissioned it. The product Meenakshi Tubes Private Limited manufactures does not matter. The fifteen minutes between Sharada Iyer and Prakash Nadar is where everything happened.

Why is most feedback useless, and whose fault is that?

Consider the last time somebody handed over a long document and asked what the reader thought. The document got skimmed, one sentence turned up that something could be said about, and something was said about that sentence. Skimming was not laziness. The skim was the only affordable response. The question contained no information about what the person actually wanted, and working that out from scratch would have cost more time than the person asking had any right to expect.

The other side of that is worth standing on. When Sharada Iyer sends a note to somebody with the words let me know what you think attached, the reader has to solve three problems before they can help: which part of this is the author unsure about, how much of this is already settled and not up for discussion, and how blunt the author is expecting them to be. None of those is answerable from the document. The reason most feedback arrives vague is that most requests arrive vague, and the vagueness is the requester's to fix rather than the reviewer's.

There is a household version of this that lands immediately. A friend asked how the food was will say it was lovely. The same friend asked whether there was too much salt in the dal will tell the truth, because the question has shown exactly where honesty is welcome and made it cost almost nothing to give. Nobody in that exchange became a more honest person. The question changed, and the answer changed with it.

How to Request and Apply Feedback on Finance Work: What Makes a Request Answerable?

A narrow requestA request for feedback that names one specific thing to look at, rather than asking for a view on the whole piece of work. does three things at once. The narrow request names the one thing to be looked at, it says what kind of wrong the author is worried about, and it gives the reviewer somewhere to be critical without having to invent the licence themselves. All three matter, and the third is the one people leave out. A reviewer who has not been told that criticism is wanted will assume it is not, because that assumption is free and the other one is expensive.

Because it is short and does all three jobs at once, the request Sharada Iyer actually used is worth reading twice. Is there anything in the computation that a reader could trace differently. Not what do you think. Not any thoughts welcome. Not I would really value your input. The single sentence pointed the review at the inputs rather than at the prose, and the transcription was found for that reason alone. A reviewer reading for style would have read the sentence carrying the wrong number and found nothing wrong with it. As a sentence there was nothing wrong with it. The sentence was a perfectly good one containing 12.5 per cent where 10.8 per cent belonged.

Who should be asked, and is the most senior person the right one?

Two things make somebody useful to ask, and neither of them is rank. The first is whether they know enough about this particular piece of work to be specific. The second is whether they will say the uncomfortable thing out loud. A person who has both is worth a great deal; a person who has one of them is worth something narrow and known; a person who has neither is a pleasant twenty minutes.

Seniority correlates loosely with the second and often negatively with the first. The most senior person available will usually tell the truth, because they have less to lose by doing so, but they have not read the workbook and their comment will be about shape rather than about cell C14. The person who is hardest on a piece of work is frequently the person who has read the most of it. On this assignment that was Prakash Nadar, who had already returned Sharada Iyer's first two pieces for reworkSending a piece of work back to be redone rather than corrected, usually because something structural about it is wrong.. The two returns are exactly why he was worth asking. Somebody who has returned an author's work twice has demonstrated they will do it a third time. Two returns are a form of reliability, however little it feels like one at the time.

How is feedback received without arguing?

Receiving runs in a fixed order, and the order is fixed precisely because the moment is emotionally loud and no decision should be made inside it. Every finding is written down as it is said, in the words the reviewer used rather than in the words the author would have chosen. Only what is needed in order to understand them is asked. Thanks are said, and meant, because forty minutes of somebody's working day went into this. Then the author leaves, and takes any disagreement away.

The writing down is not administrative. The note-taking does three jobs at once. Stopping the author's mouth is the main one; slowing the conversation to the speed of a pen gives the reviewer room to keep going; and the record it produces is what tomorrow's work runs from, once the second finding has been forgotten entirely. A reviewer watching somebody write down what they say is also getting an unmistakable signal that the effort is landing, and that signal is what buys the third finding.

Exactly one kind of interruption is permitted, and it is the clarifying questionA question asked only in order to understand what somebody means. Its whole content is the reviewer's own words handed back for confirmation.: a question whose only purpose is to understand what they mean. Which chart do you mean, the one on the second sheet or the one in the annexure. Do you mean the axis or the scale. Is that finding about the cell or about the sentence quoting it. Such questions are free. Any sentence explaining why the work was done that way is not a clarifying question, however it is dressed. A question that contains the author's reasoning is not a question.

What does one sentence of defence actually cost?

Everybody pictures the failure here as somebody getting angry. It is not. The habit that ends the supply of feedback is quiet, brief, polite and usually correct. The habit is a single sentence, offered while the reviewer is still speaking, explaining why the work was done the way it was. I started the axis at nine so the movement would be visible. The sentence is a defenceExplaining why something was done the way it was, offered while feedback is still arriving rather than after it has finished., and the fact that it is true is exactly what makes it hard to spot.

The defence does not change the current finding. The finding has already been said and cannot be unsaid. The defence changes the price of the next finding. The reviewer now knows that raising a point costs an explanation, and the third finding, not yet reached, is quietly downgraded from worth mentioning to not worth the friction. Reviewers do not decide this. Nobody sits there weighing it up. The downgrade happens in the half second before speaking, in the same part of the mind that decides whether to mention that somebody has something in their teeth.

What happens to feedback after the conversation ends?

Applying feedback is two moves, and almost everybody does the first and skips the second. The first is to change the work. The second is to tell the person what changed, and telling them is closing the loopTelling the person who gave the feedback what changed because of what they said, usually in two or three lines.. The second move is the one that decides what the next round looks like, and it costs under a minute.

Changing the work has one rule attached to it, and the rule is that the check comes before the change. A finding is an observation by somebody who read the work quickly, not a verdict. Two of the three findings on this assignment were plainly right the moment Sharada Iyer looked. One needed checking, which took four minutes, and it was also right. Checking is not defending. Checking happens after the conversation, in the workbook, where being wrong costs nothing and nobody is watching. Timing is the entire difference between the two, not attitude.

Why does closing the loop make the second round better than the first?

The reviewer's side of this is worth standing on again. Somebody has been given forty minutes and three findings. A week passes. The reviewer has no idea whether any of it was used, whether it was even read, or whether the person quietly decided the findings were wrong on all three counts. The next time that author asks, the reviewer is not deciding whether to be helpful. The reviewer is deciding whether to spend forty minutes on something with no evidence of a return, and a rational person spends twenty instead.

Now give that reviewer two lines. Finding one produced a corrected margin of 10.8 per cent. Finding two produced the chart redrawn from zero. Finding three produced a source log item number on the sentence that had none. Two lines are the only evidence the reviewer will ever receive that forty minutes of their working day turned into anything, and the second review is better because it starts from what the first one established rather than from nothing. Three weeks later, the second review opened at the computation rather than at the prose, and it was harder than the first one had been. Politeness is not what is being rewarded. A reviewer now knows exactly where their attention converts into change and is spending it there.

How to Give and Receive Feedback on Finance Work: What Changes on the Giving Side?

Everything above assumed the receiving side. Sooner than most people expect, and often in the first year, somebody hands over a workbook and asks for a look at it. The rule that carries almost all of the weight on the giving side is to write findings by kindGrouping what was found by the class of problem it belongs to, rather than listing each place the problem appears. and to attach the fix.

The reason is a claim about the person receiving rather than about the document. Suppose a check turns up eleven cells with a number typed inside the formula, so the input cannot be traced back to anything. Written as eleven separate comments, that is eleven small judgements for the person receiving it: is this one worth fixing, is this one different from the last one, does this one matter given where it sits. Eleven judgements made under time pressure will not all come out the same way, and the one that comes out differently is the one that survives into the delivered work. Written as a single finding naming the kind, listing all eleven locations and attaching the fix, it is one decision instead of eleven, and one decision is the whole of what a reviewer is trying to give somebody.

Attaching the fix is the part people skip because it feels like doing somebody's job for them. The opposite is true. A finding without a fix hands the receiver a problem and keeps the solution. The receiver now has to reconstruct what the reviewer had in mind, and may reconstruct it wrongly. A finding with a fix attached can be agreed with or disagreed with in one reading, and disagreement becomes cheap, which is exactly the state a reviewer and an author should be in.

What is the difference between feedback on the work and feedback on the person?

Two sentences can describe exactly the same thing and only one of them can be acted on. This chart starts at nine, which turns a 1.2 point fall into 30 per cent of the frame. You are careless with charts. Both are about the same chart. Both may even be fair. The first contains an artefact, an effect and, implicitly, an instruction. The second contains a person.

A statement about the work names something that can be changed in the next ten minutes; a statement about the person names something that can only be denied or accepted, and denying it is what most people will do. The next thirty seconds looks different in each case. After the first sentence, the two people are looking at a chart. After the second, they are discussing whether one of them is careless, and the chart is still starting at nine while that conversation happens.

None of this is a rule about being nice. Being nice would be softening the first sentence until it stopped saying anything. The rule is about where the sentence points. Point it at the artefact and the person can act; point it at the person and the artefact goes unfixed. The sentence that names the axis start and the 30 per cent is by some distance the more direct of the two, and it is also the one nobody has to defend themselves against.

Human Authorship: Who Wrote This, and Why Does the Question Decide Whether Feedback Works?

Everything above has quietly assumed one thing, and it is worth saying out loud because it is the condition that makes the rest possible. Human authorshipA person having actually made the choices in a piece of work, so that the choices can be explained and defended by somebody. here means only this: a person made the choices, so a person can say why. The definition goes no further than that, and it is enough to carry the whole of feedback.

Look at what a finding actually is. Why does the axis start at nine. Why is this cost held flat when revenue moves. Why does this sentence carry no item number. Every one of those is a question about a decision, and a question about a decision needs somebody who made it. The useful part of the exchange is not the observation but the reason that comes back. When Prakash Nadar asked about the axis, the answer was that Sharada Iyer had started it at nine so the movement would be visible. The answer turned a finding into a shared problem. Both people now knew what the chart was trying to do and could ask whether starting from zero would still do it.

Take the author out of that chain and what remains is this. The observation is still there and the question can still be typed, and nothing comes back. Nobody decided anything, so nobody can answer. All that is left is somebody checking a document against rules. Checking against rules finds a spelling error and cannot find a wrong axis. A wrong axis is only wrong once its purpose is known. The first question on any piece of work put up for review is therefore who made these choices, and the honest answer to it decides which conversation the two people are about to have.

What did the whole of this look like on one piece of work?

Assembled, the assignment sets every move in this guide against the same fifteen minutes. Sharada Iyer had written a note on the operating margin of Meenakshi Tubes Private Limited and needed it looked at before it went anywhere. Prakash Nadar ran the review pass on 11 March. The pass took forty minutes against about nine hours of production. The feedback conversation that followed took fifteen minutes, which is about five minutes per finding. Fifty five minutes of Prakash Nadar's working day went into somebody else's work, and that is around a tenth of the time the work itself took to produce.

Read the table downwards and notice what is doing the work. Not talent, not seniority, not being right. One sentence at the top decided where a reviewer looked. One habit in the middle, writing rather than talking, kept the third finding alive. One message near the bottom, costing under a minute, decided what the next review would be worth. All three are available to somebody seven months into a role with no authority and nobody who owes them a favour. Availability to that person is what makes this a method rather than a personality.

Where does any of this show up outside a research desk?

Take a household getting a second opinion on a home loan offer. The broad version is to send the whole offer to a cousin who works in a bank and ask whether it looks all right, which returns the rate looks fine, take it. The narrow version names one thing: is there anything in this offer that changes what I pay if the rate moves. The narrow question sends the cousin to the reset clause and the prepayment charge, where the answer lives, and it takes them four minutes instead of forty. The household did not become more financially literate between the two versions. The question got smaller.

A credit committee runs the same loop on a much larger scale. A proposal goes up, the committee returns findings, and the analyst who comes back next month with a two line note saying which condition was tightened because of which finding gets a different quality of scrutiny from the analyst who returns with a fresh proposal and no reference to the last one. On the giving side, a supervising analyst who writes findings by kind gives a junior three decisions to make; the same supervisor writing forty tracked changes gives them forty, and forty rushed decisions is how a corrected note acquires a new error. The person on the other side of feedback is always making decisions with the material handed to them, so the shape of what is handed over decides how many chances there are to get one of them wrong.

There is one more place this shows up, and it is the one people notice last. A person working entirely alone, with no colleague and no reviewer, can still use the request. The narrow question is written down before the work is looked at again, a gap is left between finishing and checking, and the reading is done against that one question rather than generally. Self-review will not catch as much as another person would. Self-review catches considerably more than a general reread does, and the reason is the same reason the method works between two people: a general question has no cheap honest answer, even when the person being asked is the author.

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.