Abhishek KC
Notes from delivery

Notes from delivery

What five years in support queues taught me about scoping

March 2026 3 min read Scoping

The requirements nobody writes down are the ones support inherits. Here are five questions I now ask during scoping, all of which I learned the hard way on the other side of the handover.

I did not come into project management from business analysis. I came from support, and I spent about five years there before I ever wrote a scope document. For a long time I thought of that as the part of my history I would eventually grow out of.

It turned out to be the most useful thing I brought with me. Not because support taught me technical detail, though it did, but because support is where you find out what a scope document left out. Every gap between what was agreed and what the customer actually needed lands in a queue eventually, and somebody has to answer for it in writing.

The gap nobody plans for

A scope document describes what will be built. What it rarely describes is what happens when the thing built behaves in a way nobody anticipated. Those are two different documents, and only one of them usually gets written.

So the pattern repeats. A project ships, everybody signs off, and three weeks later support starts seeing a class of problem that was never in scope because nobody thought to ask about it. It is not anyone’s fault exactly. The people writing the scope are thinking about the happy path, because that is what they are being asked to price.

Scoping asks what should happen. Support asks what happens when it does not.

Five questions I now ask

These all come from tickets. Each one is a question I wish somebody had asked before a build I later had to defend.

  1. What does the user see when this fails? Not what the log records, what the person on the screen sees. If the answer is “nothing” or “we have not decided”, that is a scope item, and it is usually a cheap one to add at this stage and expensive to retrofit.
  2. Who is the first person to hear about it? If a feature breaks quietly, the first report comes from a customer to support, not from monitoring to engineering. Knowing that changes what you instrument and who you train.
  3. What does this depend on that we do not control? Third-party services, a client-side integration, someone else’s API rate limit. Every external dependency is a future incident with a date not yet assigned.
  4. What is the manual fallback? When the automated path is unavailable, is there a way for a human to complete the task? Sometimes the answer is legitimately no. But it should be a decision, not an omission.
  5. Who will own this in six months? A build with no named owner after handover becomes everybody’s problem, which in practice means nobody’s. Naming the owner during scoping is free. Finding one later is not.

This makes scoping slower

I want to be honest about the cost. Asking these questions adds time to a phase that clients already experience as a delay before anything visible happens. It also occasionally surfaces something uncomfortable, like a dependency the client did not realise they were exposed to.

I still ask them, because the alternative is discovering the same information in a queue six months later, with a customer waiting and less room to move. A scope that takes two extra days and holds is cheaper than one that is agreed quickly and renegotiated.

The one adjustment I have made is to be explicit about why I am asking. If a client hears a list of failure questions with no framing, it can sound like pessimism about the project. Framed as “here is what we will not be surprised by”, the same questions read as care.

What support actually gives you

People sometimes describe a support background as an unusual route into delivery, and it is. But the specific advantage is narrow and worth naming precisely: support teaches you the difference between a system working and a system being usable.

Those come apart more often than you would expect. A build can meet every acceptance criterion and still generate a steady trickle of contacts, because the criteria described the mechanism and not the experience. Once you have spent years answering that trickle, you scope differently. Not more cautiously, just with a clearer picture of where the cost lands after everyone has moved on to the next thing.

A NOTE ON THIS PIECE

Written from my own experience across support and delivery roles. No client is named, and nothing here describes a specific engagement.

I am Abhishek KC. I manage delivery on client software projects from Kathmandu, working with teams in Australia and the US.

More about my work