pricingcustomer memorymarginsmall business

Customer Memory: The Invisible Margin Walk Costing Service Businesses 10-20% a Year

Andrew Jacob · May 30, 2026

A machine shop owner I talked to last month described his pricing process like this: "I remember Acme. They get the standard rate minus 10%. Usually."

The "usually" is where the margin walks.

Pricing from memory works fine in one shop, on three customers, for two years. It stops working the moment any of those numbers go up. By the time a service business is doing 30+ quotes a month across 40+ customers, the owner is making roughly 100 pricing decisions a quarter — and getting maybe 80% of them right relative to what they actually charged the customer last time.

The 20% miss costs.

How the leak compounds

Here's the pattern across most service businesses:

Customer asks for a quote on familiar work
Owner re-derives the price from scratch — material, labor, margin
Owner adds a "customer discount" based on rough memory
The number is sometimes 5-15% lower than what the customer last paid for the same thing
Customer signs immediately — because they got a better deal than last time
Owner doesn't notice because the quote felt normal

Multiply across a customer base. A shop with 40 active customers, doing ~3 repeat quotes per customer per year, lands on roughly 120 repeat-customer pricing decisions. If 20% of those are 10% below the prior price, that's 24 quotes a year underpriced by 10% — call it ~$15K of margin handed away if the average deal is $6,000.

Add the cases where the owner over-prices a repeat customer by 10% and the customer walks (because they remember what they paid last time even when you don't), and the total leak from pricing inconsistency lands around 10-20% of margin for most shops. Paid by you. Invisible on every report.

Why "remember everything" doesn't scale

Three structural reasons:

1. Memory degrades. What you charged Acme in October is sharper in November than it is in March. The further the lookup, the more the number drifts. 2. The team grows. The moment a second person is quoting — office manager, sales lead, fabrication lead, anyone — the per-customer pricing patterns walk out of the building. The new person doesn't know that Acme gets net-60 instead of net-30, that Beta gets 12% off setup, that Gamma always pushes for $50 off freight. They quote standard. The customers notice. 3. The patterns change. Material costs go up 7% in March; some customers get the pass-through, some get held to the prior rate per their last quote terms; some get a discount because they're high-volume; some get a markup because they're a pain. A human can hold 4-5 of these patterns. Not 40.

The right shape of the fix

You don't need a CRM. You don't need a salesforce. You need a system that, when you start drafting a quote for an existing customer, surfaces what you charged them the last 3-5 times for similar work, before you start typing.

That's it. Not a workflow tool. Not a marketing automation. A recall system that turns the "what did I charge Acme last time" question from a 5-minute spreadsheet hunt into a 2-second auto-populate.

What good looks like:

New RFQ from Acme arrives
System recognizes Acme, pulls the last 3 quotes you sent them
System pulls the discount/payment-terms patterns Acme typically gets
Draft quote is built from those references with the current material costs layered in
Your job is to review and adjust — not to remember

This is the moat between a shop running on memory and a shop running on a system. The shop running on memory walks 10-20% of margin every year. The shop running on a system doesn't.

What it changes operationally

The first 90 days of running quotes against customer memory looks like this:

Week 1-2. The first time a repeat customer's quote auto-populates the prior pricing, you'll notice it caught you. You were about to quote 12% lower than last time. You correct, the deal stays at the right margin. Week 3-4. You start trusting the system more than your memory. The 5-minute spreadsheet hunts disappear from your workflow. Month 2. A second person on your team starts using the system. Their quotes are suddenly consistent with yours, because the system holds the patterns that used to live in your head. You stop being the sole pricing authority. Month 3. You realize you're consistently quoting at your actual rates, not your remembered rates. Margin tightens 5-10% across the customer base. Revenue doesn't change; profitability does.

The shape of that lift is consistent across shops that adopt customer memory as an operational practice. The number isn't the wow-stat. The wow is realizing you'd been giving away that money every year.

Customer memory plus per-vertical defaults

Customer memory is the per-relationship pattern. There's a second layer underneath it: per-vertical defaults. Even on a brand-new customer, the system should know that A36 carbon steel runs $0.42-$0.55/lb in your region, that machine-shop setup amortization on a 5,000-piece run typically discounts at the 5K break not the 10K, that contractor quotes typically run net-30 not net-60.

That's industry pattern data — the cold-start a shop needs before they've quoted their first customer. The combination of per-customer memory (warm) and per-vertical defaults (cold-start) is what closes the consistency gap end-to-end.

For more on how customer memory plays with the other four RevOps practices that compound to close the cash cycle gap, the SMB Cash Cycle Scorecard is the free playbook.

Setell remembers per-customer pricing automatically — every signed quote and every QuickBooks invoice for the customer feeds the recall layer. Next time you quote them, the references surface before you start typing. Start free — 10 quotes on the house.

Ready to quote faster?

Start free — no credit card. See your first AI-drafted quote in minutes.