Support teams live on small requests. A customer asks which version a fix landed in. A colleague wants a log line explained. A ticket needs a status change. Handled well, these keep a service desk calm; handled badly, they become a queue nobody wants to open. Below are straight answers to what teams ask when they first apply applying the two-minute rule to support work, one question at a time. In short: when a request needs only a couple of minutes and you hold the information, deal with it immediately, and save the ticket system for the work that needs tracking.
What exactly counts as a two-minute support task?
Anything you can finish without waiting on another person or system and without touching live configuration. Typical examples are answering a factual question from existing documentation, correcting an out-of-date line in a runbook, setting a ticket’s status after you have actually done the work, and adding a short note explaining what you tried. If it needs an escalation, a change approval or a lookup you do not already have, it does not qualify.
Why not just log everything as a ticket?
Because logging has a cost, and for small items the cost is bigger than the work. Creating a ticket means a title, a category, a priority and an owner. Reading it later means finding it, reopening context and deciding again what to do. A queue padded with trivial items hides the ones that matter, and response metrics get skewed by tickets that were open for days only because nobody prioritised them.
Won’t we lose our audit trail?
Not if you leave a trace that fits the size of the task. The reply you send is itself a record. A one-line note on the existing ticket, a commit message or a changelog entry covers most small fixes. If your service agreement requires every change to be recorded, follow it. The two-minute rule does not override contractual or compliance requirements; it just argues for light-weight records where they are allowed.
A quick example helps. A customer writes asking whether last week’s patch is in the current release. The answer sits in the release notes, which are already open. Replying takes ninety seconds; creating a ticket, tagging it and returning to it would take longer, and the customer would wait for no good reason.
What about requests that look small but are not?
This is the most common failure. A customer asks for “just a quick restart”, or someone wants “a tiny config tweak” on a shared service. Appearance of size is not size. Ask three questions before acting: what does this touch, how do I undo it, and who else is affected? If the answers take more than a moment to find, move the item to the proper process. Teams that adopt the rule should write this down as a boundary for everyone, not leave it to individual judgement on a busy afternoon.
How does this help with response times?
It shortens the lag between request and first useful reply. Many support complaints are not about the fix; they are about silence. A prompt reply that says “received, here is what happens next” costs a minute and changes the customer’s experience. Fewer open items also means fewer reminders and follow-ups, which are a hidden source of workload in any desk.
Does it fit with core delivery principles?
Yes, and the connection is worth noting. Short feedback loops and shared ownership are recurring themes in material such as a primer on the guiding ideas of dependable software delivery. Clearing small items promptly is a practical expression of those themes at the support desk: problems are resolved close to where they appear, by the people who see them.
Who is responsible for small things on a team?
Everyone who notices them, within the limits of their role. That can feel vague, so it helps to see how responsibilities are divided in practice. A guide to the range of DevOps job titles and what each role covers is a good reminder that titles differ between organisations. Agree locally who may fix what on the spot, for instance whether support staff may edit runbooks directly or must propose changes, and put it on the wiki.
What should never be done on the spot?
- Changes to production configuration, access rights or credentials.
- Replies that commit the company to a deadline or a refund.
- Any message that describes an incident to a wide audience.
- Deleting data or tickets that someone else may need.
These carry consequences if you get them wrong, so they deserve review and a second pair of eyes.
How do we introduce the rule without a big programme?
- Pick a pilot week. Ask the team to note how many tickets were created and closed within ten minutes of arriving. That figure shows how many items were clutter.
- Write a one-page boundary. Five items that can be done immediately, five that must go through the process.
- Add a “small changes” note. One sentence per quick fix, kept in a shared document and read at the weekly meeting.
- Review after a month. Adjust the boundary where people hit grey areas.
Is it fine to read outside material for ideas?
Of course. Support people borrow ideas from all over the place. Some of them follow a forum site for people talking about Indian listed companies, which carries a notice that its content is not investment advice. For a support team, the sensible reading of that notice is general: forums are conversation, and the decision about what to change in a customer’s environment always rests on your own verified process.
What if people stop using the ticket system altogether?
Watch for it. If the number of tickets collapses overnight, the rule has probably been stretched too far. The healthy outcome is a moderate drop, mostly in trivial items, with larger work still being logged as before. Compare ticket types before and after, not just the total.
Where do we begin tomorrow morning?
Open your queue and look for the items you could finish in two minutes with what you already know. Finish them and close them with a short note. Then, as new requests arrive tomorrow, apply the same test before creating anything new. That is the whole technique, and the answers above only guard its edges.