For four years our service desk has reported a satisfaction score every month. When a ticket is resolved, the person who raised it gets an email with a link and five stars to choose from. The score sat at around ninety four percent, the best figure on the operations scorecard.
In March the staff survey asked every employee how they rated IT support. Head office gave us eighty eight percent. The depots gave us forty one.
Both numbers were honest. The question was how our own score had never noticed the second one. About two thousand three hundred of our three thousand four hundred staff work in depots, and most of them have a frontline licence with no mailbox. They sign in to handsets and the rota app, and when something breaks they ring us. The analyst logs the call against the shift supervisor, or against a contact called Depot General that each site shares. The resolution email went there. Supervisors rarely rated a fix for someone else's problem, and nobody ever opened Depot General. Depots raised fifty eight percent of our tickets and returned three percent of the survey responses.
So the score was accurate. It was an accurate measure of head office, and it had been presented as a measure of everyone.
We rang a hundred and fifty depot callers in the week after their ticket closed. Sixty three percent were satisfied. The commonest complaint was being told something was fixed when it was not, because the analyst had tested it from a laptop in head office, on a different network from the handset on the loading bay.
Depot calls are now logged against the person, by employee number. The survey goes by text message to the mobile number held in HR, one question and a comment box. The score is reported separately for head office and depots, with the response rate beside each, and any group that answers less than one time in ten is shown as not measured rather than folded into an average. A fix for a depot device is confirmed with the caller before the ticket closes.
Our headline figure fell to seventy eight percent, and the meetings about it became far more useful.
A survey can only measure the people it reaches. Before trusting a score, I now ask who it could never have asked.
– Serguey Shinder
Top comments (0)