A service team sets up QodFlow. They put SLAs on every stage: Intake, In Progress, Review, Done — each with a target. Six weeks in, the dashboard says everything is green. Customer complaints are up 20%.
Where the clock actually starts
The customer's clock starts the moment they hand something off — a device at a repair counter, a request in a shared inbox, a bug report. Yours starts when someone moves the ticket into the first working stage. Those are often 4-8 hours apart, and sometimes a whole day.
Take a repair shop as a concrete example: the customer dropped off a laptop at 9am. The intake person was busy. They put a sticky note on the laptop. At 4pm somebody finally entered the ticket. The SLA clock starts at 4pm.
From the customer's perspective: they handed it over seven hours ago. Nothing has happened. They're going to check in tomorrow morning asking what's going on, and the dashboard will say "18 hours remaining, on track." The dashboard is lying.
The fix: an intake stage with a punishing SLA
Add a stage called Intake before real work starts. Give it a 15-minute or 1-hour SLA. Its only job: get the request entered into the system as a real ticket with a real owner. Nothing more.
When the intake SLA hits the 30-minute mark, somebody does something about it. The backlog of unlogged work evaporates. The customer's mental clock and your system clock start within minutes of each other.
This sounds like overhead. It's actually the cheapest customer-satisfaction win you'll ever ship. Customers who feel seen in the first 30 minutes don't call back to check in. The noise stops.
The other clock mistake: not pausing when you should
The second-most-common mistake: clocks that run while you're waiting on someone else. The job is technically "in review," but actually it's waiting on a customer to approve a quote, or a stakeholder to reply. The clock keeps ticking. The job goes red. Everyone's metrics look bad. Nobody's fault.
Every stage SLA needs a pause condition. When the ball is in someone else's court, the clock pauses. When they respond, it resumes. Your team's metrics reflect what your team actually controls.
What good SLAs look like
- Intake — <1 hour. The ticket exists, has an owner, requester has a tracking link.
- Triage / Diagnostic — 1 business day. Scope or quote ready.
- Awaiting Approval — paused clock, with a 3-day "follow up" reminder.
- Blocked / Awaiting Input — paused if the blocker is documented and external.
- In Progress — 1-3 business days depending on complexity.
- Review / QC — same day. If it's been sitting longer than 4 hours, something is wrong.
- Ready / Delivered — 7-14 days, then escalate to outreach.
The real KPI
Most teams measure "average turnaround time." The number that actually predicts satisfaction is time from hand-off to first ticket update visible to the customer. Get that under 30 minutes and the rest of your SLAs barely matter.
The clock the customer is watching is not your clock. Until you fix that, no SLA setup will save you from the angry-check-in problem.