Twenty-two clauses
and four honest gaps
ISO 18295 is the standard for contact centres and comes in two parts: requirements for the centre itself and requirements for the client of the service. I assessed my own support function against the first part, to know what the corporate contact centre programme could rest on. What proved useful was not the overall conformance percentage but the list of things we did not have.
Why measure against a standard at all
A support function almost always believes it has "broadly everything in place". Until an external frame appears there is nothing to argue with: each item on its own looks like a detail, while together they are the system. A standard is useful precisely because it is not yours. It knows nothing about your priorities and asks the same questions about the parts you enjoy and the parts you would rather avoid.
The second purpose was practical: the corporate contact centre programme needed a document showing the starting point rather than the desired future. Certification against the standard is one of the programme's goals, so the gap had to be measured before, not after.
Method: four statuses instead of two
- Closed — there is a document or a working process you can point to
- Partial — the practice exists but is not recorded, or does not cover the requirement in full
- Gap — there is no record at all
- Not verified — checked against a document's description rather than its content
The fourth status is the important one, and I added it after the very first pass. The initial map was built from the document catalogue, meaning from descriptions. That is fast, but a description always promises more than the file contains. Until an item has been read in full, its status stays in question, and the report shows that with its own mark.
What turned out to be missing
Four gaps found in a function that looked mature:
- Complaint handling as a formal process. Complaints were dealt with, but there was no register, no categorisation, no deadlines and no clear route for a customer to raise one
- Employee satisfaction. Motivation and feedback existed; a systematic measure of the agent experience did not
- Demand forecasting. Shifts were planned; a traffic forecast and any assessment of forecast accuracy did not exist
- Service continuity during an outage. The order of play was undocumented: notify the customer, offer an alternative, state a timeframe
After the review that last item became an incident plan covering twelve failure directions, with roles, activation thresholds and compensation rules. In other words the analysis did not stay a report: three of the four gaps turned into tasks with an owner and a deadline.
What I took away from the work
- A report where everything is green is of no use to anyone: it yields no decisions
- "Partial" without an explanation of what exactly is missing is as useless as "conformant"
- A standard does not replace judgement: I deliberately marked one metric from the annex as not applicable to my traffic, and recorded why
- The analysis stays alive as long as it keeps being updated. Mine is a living document with statuses, not a one-off presentation