Key Takeaways
A leak is a change. One period of transaction data has nothing in it to find.
Only five things can move the margin. Asking about each one in turn usually says which of them did.
Convert every lever move into points of margin. Points add up to the fall, so they show whether you have found all of it.
A charge you never raised leaves no line on the invoice. Uncharged surcharges have to be found from the cost side.
Three points of margin on a $40 million book is $1.2 million. It usually sits with fifteen or twenty accounts.
Margin came in at 12% this year, down from 15% the year before. Someone has asked why, and you have a quarter of transaction data open on your screen to answer it.
The answer is in that extract, but the obvious approach will not find it. If you open a quarter of orders and scan for something that looks wrong, you will find nothing, because every line looks like a normal order. Somebody approved each one, at a price the system allowed, on terms a contract supports.
The method that works is narrower. You compare two periods, line up everything that can move the margin, and see which one moved. Then you convert those moves into points of margin until they add up to the fall, and narrow from the lever down to the accounts carrying it.
After that comes one more pass over the money you never billed for, which no analysis of the transactions will ever surface on its own. The rest of this article is that method in order: confirm the leak by comparing two periods, find which of five levers moved, separate real leakage from mix and cost, catch the money that never reached your data, then narrow from the lever to the accounts carrying it.
What Can a Single Period of Data Tell You?
A margin number on its own is a fact, not a finding. Twelve percent only becomes interesting next to what it used to be. The same 12% can mean three different things.
Only the middle row is a leak, and you cannot tell which row you are in from a single extract. That is why the work starts with a comparison rather than a report.
Choosing the Two Periods to Compare
Pull margin by customer and by product for two periods. Look for where it has come down. Three decisions about how you pull it will shape everything you see afterwards.
The Three Comparisons Worth Running
Each choice trades speed against noise.
- The same quarter a year ago. this handles seasonality, and finance already reports it, so nobody argues about the baseline. It is the default for most books.
- Two consecutive quarters. noisier, because seasonal mix works against you, but you find out sooner. Worth this on a book that reprices often.
- Rolling twelve months against the prior twelve. The smoothest view and the slowest to react. It shows a structural decline clearly but hides a leak that started last month.
Why Does the Customer Set Have to Stay Constant?
Check who was buying in both periods before you read anything into the gap. Some of the fall may come from accounts that were not there a year ago. That is a change in who you sell to, not leakage, and it needs a different response.
Pocket Margin, Not Invoice Margin
An invoice stops at the discount, so it never sees the rebate accruing against the order or the freight you absorbed on it. A pricing analytics view built on invoice data flatters every account with a rebate program behind it. Those tend to be your largest accounts.
How Did One Distributor Explain Its Own Decline?
Watsco ran the comparison on its second quarter call in July 2026. Chairman and chief executive Al Nahmad put the two years side by side: "During 2025, OEMs instituted aggressive pricing action in response to inflation and tariffs, benefiting gross margin in 2025. By comparison, 2026 OEM pricing actions were more moderate and consistent with historical levels."
The company reported gross margin down to 27.5% from 29.3% a year earlier, and attributed roughly 130 basis points of that to the prior year comparison rather than to anything that had gone wrong since. That is the whole exercise in one example: a fall in margin, a named cause, and a number attached to that cause.
Lining Up the Levers in Points of Margin
You now know which part of the business lost margin. The next question is which parameter did it.
The Five Questions Worth Asking
There are only five candidates worth checking. Ask each one in turn.
- Did list prices come down?
- Did discounts get deeper?
- Did rebates go up?
- Were surcharges not passed through?
- Are some customers getting services for free under an old agreement?
Take a book that fell from 15% to 12% over two quarters. The answers might come back as list prices down 3%, discounts up 6%, and rebates up 6%. That is useful, but it is also where most analyses stop, one step short of being checkable.
Why Points of Margin and Not Percentages?
A 6% rise in discounts and a 6% rise in rebates do not cost the same, because they touch different shares of the book. Points of margin are the common unit, and points add up.
Those points are computed, not inferred from the percentages. To get each number, you recompute the margin with only one lever changed and everything else left at last year's value, then repeat for the next lever. Always do this in the same order, price first and mix last. When two levers hit the same orders, that fixed order stops you from counting the overlap twice.
What Does a Large Unexplained Remainder Mean?
Currency movement and day count timing leave a small honest remainder, so half a point of unexplained fall is normal.
A large residual usually means a deduction is sitting outside the data you pulled, most often a rebate or a credit note that lives in another system and was never joined back to the order that earned it.
The first line of that table is price variance analysis: the gap between what you actually charged and what your list, target, or contract rate said you should have charged. The rebate line is a different question, because it is about programs, not individual deals. Steady rebate management that re-rates each program against current volume usually catches that one first.
Not Every Point You Lost Is Leakage
Before you act on any lever, one filter has to come first. Margin falls for four reasons, and only two of them are yours to recover. Get this straight before anyone starts tightening discount authority, because attacking a mix shift as though it were a discounting problem costs you volume and recovers nothing.
Cost is the row people get wrong. A cost increase is not leakage, because it happened outside your pricing and no decision of yours caused it. Failing to raise your price to cover that cost is leakage. That gap behaves exactly like a surcharge nobody billed: both are money you had every right to ask for and did not.
See every deduction against the order that caused it
Vistaar Platform runs all workflows on a single transaction schema with no data translation or synchronization lag. The levers line up without stitching four systems together first.
Request a demo
The Money That Never Reached Your Data
he bridge explains the margin you lost inside the data. Some of the loss never made it into the data at all, and there is a faster way to find that part. Look at what you charge for surcharges and services, then compare it against what you should be charging.
This is where the transaction data stops helping. The freight you absorbed never shows up as a freight line, and the expedite you waived over the phone never shows up either. As far as the data is concerned, neither one ever happened.
You have to come at it from the cost side and ask who is carrying the charge.
Two versions turn up on almost every book.
Services That Quietly Became Free
Installation, training, stocking, or delivery stopped being charged because a customer was buying heavily at the time. Then the volume came back down, and the free service never did.
Surcharges Everyone Else Pays
A customer sitting on an older contract is not paying the freight, fuel, or small order charge. The rest of the book pays it without argument.
Pool Corporation went after exactly this on its second quarter call in July 2026. Chief financial officer Melanie Hart described the move: "On the outbound side, we did, earlier in the year, put through some freight surcharges to help us to recoup the extra cost that we're seeing just on the delivery side." She was equally direct about the limit, adding that the company "may not be able to recoup all of it immediately."
Why this is the cheapest money on the list
Passing on a standard surcharge costs you nothing extra to deliver. You already pay the freight. You already provide the service. Every dollar recovered lands straight in the gap you are trying to close. A quarter of a point recovered here beats a full point argued out of a discount structure. Nobody has to give anything up for it.
From a Lever to a List of Accounts
A lever is not yet an answer. If surcharge recovery is the problem, the question becomes which customers are not paying. That is a different query against the same data.
Three Ways to Cut It
Take whichever lever came out worst and break it down three ways. If you already have customer segments built for something else, they usually give you these cuts with no extra work, and customer-specific pricing rules are often where the worst rates are hiding.
- By customer, ranked on points of margin lost multiplied by revenue. A tiny account on a terrible rate should not outrank a large one on a slightly bad rate.
- By product or category. This separates a pricing problem from a portfolio problem.
- By region or business unit. This usually tells you who approved it.
Three points of margin on a $40 million book is $1.2 million. It is rarely spread across four hundred customers. On most books it sits with fifteen or twenty of them, which is a list somebody can work through in a quarter.
Why Fix One Source at a Time?
Fixing everything in the same month costs you the ability to learn anything. Margin comes back and nobody can say which change earned it.
Sequencing protects the account as well. A customer who loses a rebate rate, a discount, and free delivery at once will read that as a repricing. They respond accordingly.
Enriching the Transaction Record
The data usually defeats you long before the analysis does. The rebate sits in one system and the invoice in another, and nothing links the deduction back to the order that earned it, so the bridge cannot be built in the first place.
Every transaction line has to arrive carrying:
- The list price in force on the date of the order, rather than today's list.
- Every on-invoice deduction, separated by type instead of netted into one discount field.
- The rebate accrued against that line, allocated when it happened rather than estimated at period end.
- Freight, surcharges, and services, whether or not they were actually charged.
- Cost as of the transaction, so margin is calculated rather than inferred.
- Customer, product, region, and business unit, so the same data supports every cut above.
Point four is the one teams argue about, and it is also the one that pays off. Record what a charge should have been next to what it actually was. That is the only way an uncharged surcharge shows up in a report, instead of being found by accident two years later.
This is also why price optimization software tends to get bought after a failed margin analysis rather than before one. The analysis is not the hard part. Assembling the data by hand, every quarter, is.
Where Does Price Management Software Change the Job?
An annual review only catches a leak after you have already paid out three or four quarters of it. Run the same comparison every month and you catch it when it is one quarter old. That gap is where most of the value of doing this properly comes from.
What Changes When the Comparison Runs Monthly?
Three things change when the comparison runs on a cadence instead of as a project.
- The cuts are already built. customer, product, region, and business unit become filters you click, not extracts you request and wait for.
- The exceptions come to you. A waived fee, an undercharged surcharge, or a margin below its floor gets flagged rather than discovered.
- The job changes shape. You stop recalculating numbers and start reviewing what has been flagged. Investigating the handful that matter is the part only a person can do.
Which Vistaar Products Do This?
SmartPricing is the Vistaar product that carries margin leakage detection. It detects margin leakage, competitive moves, and pricing anomalies before they become costly problems. It also lets you set minimum margins, price corridors and exposure limits by category or segment. The floor a leak crosses is one the system already knows about.
- SmartOptimizer tests alternative pricing strategies and compares their impact on revenue, volume, and realization before you commit to implementation. That is how you size a correction before making it.
- SmartQuote monitors quote velocity, win rates, price realization, and margin performance. The discount lever gets watched where it actually moves.
Vistaar's own analysis of effective pricing analysis names the same two views. Margin driver analysis shows whether changes were driven by price, volume, costs, or product mix. Price variance reporting highlights gaps between realized prices and list prices, targets, or contract rates. Those are the two reports this article has been building by hand.
Walk one quarter with us
Bring a quarter of transaction data. The Vistaar team will go through it with you, lever by lever, and show you what the bridge says about your own book.
Talk to us
Finish at a Name, Not at a Lever
Most margin investigations stop one step too early. Prices came down, discounts went up, rebates grew: all true, and none of it tells anyone what to actually do on Monday morning.
The work is finished when the fall has a cause, an amount, and a list of accounts attached to it. Anything short of that is a description of the problem, written in more detail than before.
Fixing each lever is well understood and widely written about. Working out which one you need is the part almost everyone skips. It is also the only part that has to come from your own data.
Frequently Asked Questions
How Do I Start Looking for Margin Leakage in Transaction Data?
Compare pocket margin by customer and product across two periods. Then ask which of five things moved: list prices, discounts, rebates, surcharge recovery, or mix. A single period cannot show a leak, because a leak is a change.
Why Convert Lever Moves Into Points of Margin?
Because points add up and percentages of different things do not. A 6% rise in discounts and a 6% rise in rebates touch different shares of the book. Only points show whether the levers explain the whole fall.
Why Will My Data Not Show Uncharged Surcharges?
A charge you never raised leaves no line on the invoice. Absorbed freight and waived expedites are absent from the record rather than wrong in it. You have to start from the cost you are carrying instead.
Is Every Margin Decline Margin Leakage?
No. A mix shift or a rise in input costs lowers margin with no pricing failure behind it. Failing to reprice after a cost increase is a failure, and it deserves the same attention as an unbilled surcharge.
How Many Customers Usually Account for the Fall?
Far fewer than people expect. Three points of margin on a $40 million book is $1.2 million. On most books that sits with fifteen or twenty accounts rather than the whole list.
