Late Payment Fees: Structuring Terms That Actually Get Paid On Time

Why the Clause on Paper Rarely Changes Anything

A 14-person design studio I know of had "1.5% per month on overdue balances" sitting in their master services agreement for four years. In that time, they applied it exactly twice, both to a client who'd already fired them. Meanwhile, their second-largest account paid on day 52 of a net-30 term for six straight quarters, penalty-free, because the founder didn't want an awkward email to interrupt a good relationship.

That's the pattern. The fee exists to change behavior, but it only works if the client believes you'll actually charge it. An unenforced clause teaches the opposite lesson: net-30 is really more like net-50, and there's no consequence for treating it that way. Once a client learns that, they'll extend the same courtesy to you that they extend to every other vendor who doesn't push back, which is to say, none.

This is also why days sales outstanding, one of the metrics we cover in our KPI guide for agencies and consulting firms, tends to drift upward slowly rather than spike all at once. Nobody decides to pay you 60 days late. It happens two or three days at a time, every quarter, because nothing pushes back.

What Actually Needs to Be in the Clause

A late fee clause that gets enforced is specific enough that nobody has to interpret it when the invoice goes overdue. Vague language is exactly what lets a fee slide, because someone on your team has to make a judgment call, and judgment calls get postponed. Here's what a working clause actually needs to spell out.

• The exact fee and how it's calculated, a stated percentage per month (or per week) or a flat dollar amount, whichever is greater, so there's no debate about the math when the invoice is 12 days overdue
• A grace period, typically 5 to 10 days past the due date, that gives room for a payment already in transit without making every fee assessment a fight over timing
• Whether it compounds, a simple flat monthly rate on the original balance is easier to defend and explain than a compounding rate that snowballs and starts looking punitive to a court or a client's finance team
• The notice requirement, whether the fee applies automatically on a set date or only after a written notice, and who on your team actually sends that notice
• What happens after 60 or 90 days, a second, higher threshold or a suspension-of-work clause, so the late fee isn't the only lever you have if a client goes truly dark

Write this into the signed agreement, not just the invoice footer. A fee that first appears on an invoice a client never agreed to in writing is much easier to dispute or simply ignore.

How High You Can Legally Go

1% to 1.5% per month is the range most agencies and consulting firms settle on, and there's a reason it clusters there rather than at, say, 5%. Usury laws in most US states cap interest charges somewhere between 10% and 24% annually for commercial contracts, and 1.5% per month works out to 18% annualized, right in the middle of what courts generally accept as reasonable rather than punitive. Go much higher and you risk a judge simply refusing to enforce the clause at all, even on an otherwise legitimate unpaid invoice.

For smaller invoices, a flat fee often works better than a percentage. 1.5% on a $900 invoice is $13.50, not enough to motivate anyone. A flat $35 or $50 fee, or a stated percentage, whichever is greater, keeps the deterrent meaningful even on your smallest retainers.

Enforcing It Without Burning the Relationship

The firms that actually collect late fees don't apply them with more aggression, they apply them with more consistency. That distinction matters. A fee that shows up automatically and predictably on every overdue account, big or small, reads as a policy. A fee that only shows up when a project manager finally gets fed up reads as a punishment, and clients notice the difference.

The practical fix is to take the decision out of anyone's hands day-to-day. If your invoicing runs through a system where overdue status, the grace period, and the fee calculation are already connected to the invoice itself, the fee gets applied the same way every time, whether the client is a $2,000/month retainer or your single largest account. That's a big part of what billing automation is actually for, and it's worth seeing how it works on the features page if you're currently doing this by hand in a spreadsheet.

Keep a human override, though. If a longtime client's payment is three days late because their AP person was out sick, waiving the fee once, and telling them you did, builds more goodwill than the $40 you'd have collected. The point of automation isn't to remove judgment, it's to make sure judgment only gets used for exceptions instead of becoming the whole process. Most teams find the actual cost of setting this up is small next to what they're losing to stretched-out payment cycles, and it's worth comparing against what a connected invoicing plan costs before assuming it's not worth the switch.

One more thing worth saying plainly: a late fee is a deterrent, not a collections strategy. If an account is 90 days out, the fee stopped mattering weeks ago and you need a phone call, not a bigger percentage.

Frequently asked

Related reading