Most support arrangements are measured on response time, which is the easiest thing to measure and among the least useful. A four-hour response that closes the ticket by restarting the service, four times a month, is worse than a next-day response that finds the memory leak.
We work the other way round. The question we are trying to answer is why this happened, and the measure we care about is whether it happens again.
What we actually do
Keep business-critical systems running — the ERP, the line-of-business application, the integration that has to complete overnight. Diagnose root causes rather than clearing symptoms. Make small improvements continuously, because the difference between a system people tolerate and one they like is usually a dozen small things. And maintain the documentation, so knowledge survives staff turnover on both sides.
Business-aware, which changes the triage
Knowing that the stock sync feeds the sales team's availability screen changes what you do at nine in the morning. Knowing that payroll runs on the 24th changes what you do on the 23rd. Support that does not know these things treats every ticket as equally urgent, which in practice means the loudest one wins.
So we learn the business calendar and the dependency map, and we say plainly when something you have reported as urgent is not, and when something you have not noticed is.
Paying down debt instead of accruing it
Every quick fix under pressure is a small loan against the future, and support arrangements are where most of that borrowing happens. A workaround becomes permanent, an undocumented change becomes folklore, and three years later nobody will touch the module.
Part of what we do is to keep that from compounding: fixes done properly where it is affordable, recorded honestly where it is not, and a visible list of what has been deferred. Technical debt is not always wrong to take on — but it should be a decision rather than a habit.
What we ask for at the start
Access, obviously — but also the business calendar, the list of what genuinely cannot go down, the names of the people who actually use each system, and an honest account of what has been patched around. The last one saves the most time, and nobody is ever judged for it.
Documentation as part of the service
A support arrangement that leaves all the knowledge in the supplier's head is a commercial arrangement, not a technical one. We write down how things work, what has changed, and why — partly so any competent firm could take over, which we think is the correct standard to hold a supplier to. The reasoning is in documentation that is actually current and in the case for audit trails.
