Budget the exceptions before they budget you.
The SLA slide shows on-time rates for the clean cases. Split shipments, payment retries and seller corrections are "handled by the team," and nobody has written down how many of those the team can absorb before something gives.
Exception load is still load. Count it with your own numbers: process volume, the exception capacity you actually staff, how you count an exception-path hit, a review trigger as hits approach capacity, and a named owner. The point isn't a benchmark. It's making your absorption limit visible before burnout forces the redesign.
Happy-path SLAs hide exception load
The SLA slide shows on-time rates for clean cases. Exception paths—split shipments, disclosure holds, payment retries, seller corrections—are “handled by the team.” Nobody wrote how many of those hits the process can absorb before quality, speed, or people break.
Exception load is still load. If you do not budget it, it budgets you: overtime, silent queue growth, and redesign forced by burnout rather than by a review trigger.
This note gives you the TWIMCO SLA Exception Budget Card. Inputs only. No invented industry benchmarks. Distinct from a coordination-cost calculator (handoff dollars), a marketplace exception library (catalog of types), and an exception-path canvas (path design). The card answers: how many exception-path hits can this process absorb before we must redesign or change staffing?
Marketplace and post-transaction operations—work like Paddle8—make exception volume impossible to ignore if you want the customer promise to survive after checkout. Fulfillment-shaped scoping often lives under fulfillment operations or business process automation.
A forthcoming worksheet tool is planned to assemble the formula from your inputs. Until it ships, use the fields below with numbers from your own systems. If you cannot pull the numbers, that is itself a finding: you cannot manage a budget you cannot count.
Do not use this card to invent a vanity KPI for a board deck. Use it to force a staffing or redesign conversation while you still have choices.
If your SLA is contractual, align legal language with the exception budget before renewal. A contract that assumes zero exceptions while your card shows steady hits is a future dispute written in advance.
Fields of the SLA Exception Budget Card
Process & window. Name the process and the time window you will measure (week, month, peak season). Write timezone. If peak season differs, say so explicitly rather than burying it in an average.
Volume input (your number). Count of cases that enter the process in the window—from your system of record. Do not estimate from a blog or a vendor slide.
Exception-hit definition. What counts as one hit: entering a named exception path, manual rework ticket, retry beyond first attempt, or customer-visible delay flag—pick one definition and stick to it. Ambiguous counting makes the budget theater. Write examples of what does and does not count.
Staffed exception capacity (your number). How many hits the staffed roster can handle in the window without planned overtime or dropping other work. This is an operating judgment you own—not a benchmark from elsewhere. Ask the people who do the work; do not take only the manager’s optimistic number.
Budget formula (planning default). Remaining budget ≈ staffed capacity − counted hits in the window (or a simple ratio hits/capacity you prefer). Write the formula you will use. Do not claim it is an industry standard. If you prefer a utilization percentage, define it the same way every period.
Review trigger. When remaining budget crosses a threshold you set, who is notified, and what decision is on the table: redesign path, add staffing, narrow intake, or accept degraded SLA explicitly. Write the options before you need them.
Owner. Named person who maintains the card and calls the review. Not “ops.”
Optional: Peak adjustment—a second capacity line for known peak windows, still from your staffing plan.
The unreproducible asset is your filled card with real volume, capacity, and review trigger—not a generic “calculate your SLA budget” prompt.
How the work moves
Build the exception budget from your inputs
- Volume
Cases in the window from your system of record.
- Hit definition
What counts as one exception-path hit—one definition only.
- Capacity
Staffed hits the roster can absorb without planned overtime.
- Review trigger
Threshold, notify who, decision options, named owner.
An SLA that ignores exception load is a promise written on the happy path alone.
Keep the math honest
Refuse borrowed benchmarks. If someone pastes “industry exception rate,” ask for the source URL and whether the definition matches yours. If it does not, leave it out. Invented precision is worse than an honest range from your own week.
Separate severity when it matters. A card can track all hits, or only hits that breach customer promise. Mixing vanity retries with disclosure holds muddies the review trigger. If you need two lines, use two lines—do not average unlike things.
Do not average away peaks. A monthly average that hides a brutal week is how teams “make the SLA” on paper while operators drown on Tuesday. Prefer the worst week in the window as a stress check even if your official period is monthly.
Tie the card to the customer promise. If marketing claims a timeline the exception budget cannot support at your real hit rate, the conflict belongs on a promise diff—not in silent heroics. The budget card makes the conflict discussable.
Revisit after process changes. New automation can raise or lower hits; update capacity when roster changes. A card that is never re-opened becomes folklore.
When automation absorbs some hits, do not delete them from the definition until you trust the automation’s recovery path. Shadow the count for a period with a human review trigger still in place.
What the card forces into the open
Unowned exception labor. Hits land in a queue with no capacity number. The capacity field refuses that blank.
SLA written only on clean cases. Review trigger forces an explicit conversation when exception load approaches the limit.
Redesign deferred until burnout. A dated review is cheaper than an unplanned staffing crisis.
Partner SLAs without internal capacity. Marketplace and fulfillment partners inherit your exception load. If your card shows you cannot absorb the promised timeline, renegotiate or redesign before the next peak—not after the complaint spike.
These are delivery-shaped patterns, not client metrics. Use them to structure conversation; do not invent percentages to persuade leadership. Your filled evidence lines are the moat.
Bring finance into the review when the option is staffing change. Bring product when the option is redesign. The card’s job is to name the decision—not to pretend overtime is free forever.
Decision guide
Borrowed benchmark vs owned budget
| Consideration | Looks authoritative | Operable |
|---|---|---|
| Numbers | Industry rates without matching definitions. | Your volume, hits, and capacity only. |
| SLA story | Happy-path on-time alone. | Exception budget beside the SLA. |
| When load rises | Silent overtime. | Dated review trigger with options. |
| Ownership | “Ops will handle it.” | Named card owner. |
Put this into practice
- Pick one process with a visible customer or partner SLA.
- Pull volume and exception hits from your own record for one recent window; write the definitions on the card.
- Estimate staffed capacity with the people who do the work—then write the review trigger and owner.
- Bring the filled card to the next ops review; decide redesign vs staffing vs explicit SLA change if the trigger is near.
- For marketplace and fulfillment-shaped work, ground the discussion in delivery patterns like Paddle8 and scope under fulfillment operations.
An SLA that budgets exceptions is still a promise—but it is a promise you can staff.
We build alongside the people doing the work.
Bring us the problem you’re thinking about.