Writing an operational service promise
5 min read
Lesson objectives
- Write a service promise with a trigger, deadline and controllable outcome.
Please sign in to use all the tools
Once you sign in, the reading tools are yours: larger or smaller text, typeface, line spacing, focus mode, highlighting and bold, and notes saved to your account and gathered in My study.
An organization writes, “We solve every problem immediately.” The line sounds reassuring, but leaves agents and customers asking what counts as a problem and when the promise is fulfilled. An initial message might arrive in a minute while the actual request remains unresolved for a week. The specialist's job is to turn the promise into a commitment the team can deliver, the customer can understand and a supervisor can verify. If the process cannot support it, the process or the wording needs to change.
Start with the class of request covered by the promise. A shipment status question differs from a financial dispute that needs investigation and authorization. Define the event that starts the clock: submission of a form, arrival in the case system or receipt of required information. Each choice changes both measurement and the customer's expectation. Do not use “response” without a definition. An automated acknowledgement, a useful human reply and case closure are different events. Zendesk's product documentation distinguishes first reply, later updates and resolution measures. Those are useful examples of distinct measurements in one product, not a universal rule for every organization.
Write a promise card with seven elements. Specify the audience and channel; the included requests and an accessible route for those excluded; the event that starts the timer; the time limit and whether it uses working or calendar time, including time zone and holidays; the promised result, whether acknowledgement, decision or resolution; the team responsible for the next update; and the recovery path if the expected time cannot be met. Do not bury a decisive exception in fine print while the public line suggests otherwise. The customer should see the boundary that affects their next decision.
Test the card against three cases: a complete request during working hours, one shortly before closing and one that needs more information. For the third, do not hide a silent case under an “incomplete” exception. Ask for the specific missing item, explain when review begins after its receipt and keep the original arrival visible in the record. If the same field is often missing, fix the form instead of making customers bear the cost of its design. Check other channels too. A phone promise does not automatically extend to an inbox monitored at another pace. A customer with an urgent matter needs to see the appropriate route, not a broad statement implying identical service on every channel.
A second worked example shows why measurement matters. An education platform advertises “We reply within a day” without defining reply. It has 40 questions: 24 received an immediate automated acknowledgement, 12 received a human reply within 24 hours and four received one later. It would be wrong to claim 100% compliance from the automated messages if customers would understand “reply” as help with their question. If we provisionally define the measure as a first useful human reply within 24 calendar hours of receipt, 12 are known to meet it. We still cannot publish 12 divided by 40 as the final compliance figure from this description because the timing of the other human replies is unspecified, and the categories may overlap. Build an event table for every request, settle the denominator and only then calculate. Precise wording depends on sound data.
The final check is simple but demanding. Can a new customer tell what will happen, when it will happen and what the statement does not guarantee? Can an agent on duty locate the state, start time and follow-up team? Can a supervisor distinguish real lateness from a clock pause permitted under a clearly published definition? If the answers diverge, revisit scope, trigger and result. A service promise coordinates customer expectation with operations; it cannot be separated from the work behind it.
Lesson summary
Writing an operational service promise with evidence, outcome and authority checks.