Operations • For wholesale distributors and supply houses
Margin Leaks Out One Override at a Time
No single price override sinks a distributor. Thousands of them, plus stale price levels and cost increases that never reached the customer, can quietly eat a year’s profit. How AI can watch every line so your people can decide.
By Justin Beegel · · 6 min read
Here’s a made-up week that will sound familiar. A contractor leans on the counter at 6:50 in the morning with a pickup ticket for 500 feet of wire. “The place across town is cheaper on this.” The counter person has a line of six people behind him, so he knocks the price down to keep the guy happy. It takes three seconds. Nobody will ever look at it.
The same week, a manufacturer’s 5 percent increase takes effect. Your matrix customers pick it up automatically. The 14 customers on fixed contract pricing for that product line don’t, because their prices are typed-in dollar amounts and nobody remembered to touch them.
And a job price you set for a hospital project 18 months ago is still loaded on that contractor’s account. The hospital opened last spring. The contractor is still buying at hospital pricing for every small job since.
None of these shows up as a problem. They show up as a gross margin that’s down four tenths of a point, and a monthly report that can’t say why.
Here’s where I’m coming from. I’ve never set a price matrix or run a branch P&L. I spent almost 18 years running a digital agency and now work with AI to fix real operating problems inside businesses. What I’ve learned is that small leaks are usually less a people problem than a visibility problem. Nobody can watch 40,000 invoice lines a month by eye.
Where the margin actually goes
Margin leakage isn’t one thing. It’s five or six small habits, each with a reasonable excuse.
- Manual overrides. Counter and inside sales cut prices to close a sale. Some are smart. Many are reflex. Without a reason code, nobody can tell which is which.
- Stale customer price levels. A customer was assigned a top price level when they were your biggest account. Five years later they buy a third as much and still pay the best price.
- Cost increases not passed through. Fixed-dollar contract prices, long-validity quotes and special pricing don’t move when your cost moves. The margin on those lines shrinks quietly.
- Job pricing that outlived the job. A project price was meant for one job. It stays on the account and gets applied to everything.
- Charges that get waived. Freight, small-order fees, cut charges, restocking fees. Each waiver is small. Added up over a year, they aren’t.
- Quotes priced off old cost. A big quote goes out at current cost, sits for six weeks, and converts after the cost went up.
Every one of these is easy to fix once you see it. The trouble is seeing it.
What AI does with every line
It checks every quote and invoice against the rules you set
You define the rules: margin floors by category, the customer’s assigned level, expiration dates on job pricing, which charges should apply. The system reads every quote and invoice line and sets it against those rules. That’s work a person can’t do at volume, and software doesn’t get bored doing it.
It flags, and a manager decides
Below-floor lines and large overrides land on a manager’s list, grouped so the patterns are obvious: by rep, by customer, by product line. A manager can see that 22 of last week’s 37 below-floor lines came from one rep and one customer. That’s a conversation, not a crackdown.
It catches cost changes before they cost you
When a supplier cost update loads, the system lists every place that cost doesn’t flow through automatically: contract prices, open quotes, job pricing, special agreements. The pricing manager gets a list instead of a hunch.
It finds pricing that has gone stale
Job prices with no activity on the job in 90 days. Customers whose level doesn’t match what they now buy. Charges waived on the same accounts every time. Each one shows up with the numbers attached.
People still set price. The system just makes sure nothing slips by unseen.
Run the numbers
The formula is simple: annual sales × margin points lost = dollars lost per year.
Say, for the sake of the math, a distributor does $40 million in annual sales and leaks half a point of margin through overrides, stale levels and unpassed costs. Half a point of $40 million is $200,000 a year.
Now look at it against net profit, because that’s where this hurts. Keep the made-up example and say this distributor’s net profit margin is 3 percent. That’s $1.2 million on $40 million in sales.
Recover just a quarter point of the leak, $100,000, and net profit goes from $1.2 million to $1.3 million. That’s about 8 percent more profit without one new customer, because the costs behind those sales are already paid.
Here’s another way to see it. A dollar recovered from the leak lands on the bottom line, because the sale already happened and the costs are already paid. A dollar of new sales brings only what’s left after product cost, freight and the effort to win it. Plugging the leak is usually the cheaper route to the same profit.
To run it on your own business, pull three numbers:
- Annual sales, from your financials.
- The override gap. Export a year of invoice lines with the system price and the sold price. Add up the difference on every line sold below system price. Divide by sales to get the points lost.
- Your net profit margin, from the P&L. This tells you what each recovered point is worth to the bottom line.
If your ERP doesn’t keep the system price next to the sold price, that’s worth fixing first. It’s the one field that makes this measurable.
Picture a Monday, before and after
This is a made-up scene. Picture a sales manager at a four-branch electrical distributor.
Before: the monthly margin report arrives. Gross margin is down four tenths of a point. The manager asks the branch managers, who say it’s competitive pressure. Maybe it is. Nobody can prove it either way, and next month looks the same.
After: Monday morning, one page. Thirty-seven lines last week sold below floor, and 22 came from one rep on one account, all on the same wire product. Three job prices expired with no activity. Fourteen contract customers are still on old pricing for a product line whose cost went up on the first of the month. Freight was waived on 41 orders, and 30 of them went to the same five accounts.
The manager spends an hour on calls instead of a month on guesses.
What has to be in place first
- A price matrix in the system. Customer levels and category pricing recorded where the software can read them.
- System price kept next to sold price. Without both, there’s no override to measure.
- Reason codes on overrides. “Meet competition,” “damaged,” “manager approved.” Simple, and short enough that people actually use them.
- Margin floors the owner agrees to. By category, written down. Floors nobody agreed to get ignored.
- Expiration dates on special pricing. Every job price and contract price should have an end date, even a long one.
- Cost updates loaded promptly. If supplier costs sit in an inbox for two weeks, everything downstream is late.
Guardrails worth keeping
- People set every price. The system flags and suggests. Managers decide.
- A flag is a question, not an accusation. Some overrides are exactly right. Treat the list as a coaching tool, or people will stop entering reason codes honestly.
- Price changes to customers are approved by a person. Don’t let any system send increase notices on its own.
- Pricing data stays private. Business-grade tools, access controls, and nothing used to train public models.
- Keep the history. Every flag, decision and reason, so the next pricing review starts with evidence.
How you’ll know it’s working
- Share of invoice lines sold below system price.
- Dollars below system price per week, by branch.
- Gross margin on overridden lines versus all lines.
- Days from a supplier cost increase to updated contract and job pricing.
- Number of expired job prices still active on accounts.
Where I’d start
Don’t build anything yet. Export one year of invoice lines with system price and sold price, and measure the override gap. That single number tells you whether margin leakage is a rounding error or the biggest line item nobody is managing. Then decide whether the daily flag list is worth building.
If this sounds like your week, send me a note through the contact page.
Questions owners ask
Won’t flagging overrides slow down the counter?
It doesn’t have to. Most flags are reviewed after the sale, in a daily or weekly list. Only lines below a hard floor need to stop for approval, and the owner decides where that floor sits and which lines it applies to.
Can AI set the price matrix?
It can suggest changes, like customers whose level no longer fits what they buy or categories where overrides cluster. The matrix itself, and every price in it, stays a management decision.
Is every override a mistake?
No. Meeting a competitor to keep a large order can be the right call. The point is to see overrides as a pattern, by rep, customer and product, so the good ones are deliberate and the bad ones stop.
Keep reading
Operations · · 6 min read
Money Your Vendors Owe You That Nobody Claimed
Special pricing claims and vendor rebates are earned on the sale and lost in the back office. Where AI can match every invoice line to the right agreement and keep a person in charge of what gets filed.
Operations · · 6 min read
The Contractor Who Quietly Started Buying Elsewhere
Outside reps cover too many accounts to notice when one stops buying water heaters but keeps buying fittings. Where AI can read the invoice history every week and point reps at the slipping accounts while there’s still time.
Operations · · 6 min read
Every Emailed PO Gets Typed Twice
Purchase orders arrive as PDFs, emails and phone photos, and inside sales retypes them line by line. Where the minutes and the mis-shipments come from, and what an AI-assisted order desk changes.
Want this thinking applied to your business?
Tell me what is slowing the business down, and I will tell you where I would start.
