Building a Spend Approval Matrix That People Actually Follow

Before setting dollar thresholds, work backward from what your company can absorb without real pain. Say a $500 mistake barely registers, a $5,000 mistake stings but recovers within a month, and a $50,000 mistake genuinely threatens a quarter's plan; those are the kinds of real pain points worth setting approval tiers around, not round numbers that just feel official. A company with thin margins needs tighter thresholds than one with a large cash cushion, even at the same revenue size, because the real constraint is risk tolerance, not company size.

This exercise also surfaces disagreement worth having early: a founder and a finance lead often have different instincts about what counts as painful, and it's better to reconcile that before the matrix goes live than to discover it during a disputed approval.

A manager approving a purchase they don't understand isn't a control, it's a rubber stamp with extra steps. Route each spend category to someone who can genuinely evaluate whether the purchase makes sense, not just whoever sits at the right level on the org chart. Say a software purchase and an equipment purchase land at the identical dollar threshold; they still often need different approvers, because the judgment required to evaluate each one is different.

If the person you'd naturally route a category to doesn't actually have the context to evaluate it, that's worth fixing directly rather than working around. Either give them the context, a brief on what good and bad purchases in that category look like, or route it to someone else, rather than leaving a rubber-stamp approver in place indefinitely.

The most common workaround is splitting one purchase into several smaller ones that each fall under the approval threshold. This is nearly always a sign the threshold is set too low for how the business actually operates, not a sign the requester is being dishonest. When you see this pattern repeatedly in one category, raise that category's threshold rather than tightening enforcement, since a tighter threshold on an already-avoided rule just pushes the workaround further underground.

A second, quieter workaround is requesting the same recurring purchase under a slightly different description each time, so it never accumulates into a pattern anyone notices. Watching for repeated near-identical requests from the same requester catches this faster than watching individual transaction amounts.

Total cost of ownership for a purchase is rarely just the sticker price. A worksheet with three columns, the upfront cost, the ongoing cost over twelve months, and the headcount hours needed to implement or maintain it, gives approvers a much more honest basis for a decision than the invoice amount alone. Say a $3,000 tool needs twenty hours of setup time from an operations manager earning in the $72,320 to $167,280 range¹; that implementation cost alone can rival the software's price and belongs in the approval conversation, not discovered after the purchase.

Build the worksheet for each purchase with three columns:

• Upfront cost: the sticker price or invoice amount on the request, which is often the only number an approver sees.
• Ongoing cost over twelve months: renewals, seats, and any recurring fees that keep the purchase running after the first invoice.
• Headcount hours: the time an operations manager or other staff member needs to implement or maintain the purchase, counted as a real cost.

Once a quarter, look at every purchase that got routed around the matrix, denied, or approved with hesitation, and ask whether the thresholds themselves need adjusting. A matrix that hasn't changed in two years despite real growth in the business is probably either too loose for current spend levels or too tight for the trust the team has since earned.

MeetMyCOO's AI COO, Olivia, can surface which categories saw the most split-purchase or resubmitted requests last quarter, which gives you a starting list for the review instead of scanning every transaction by hand.