01 / The messy start
Good ideas rarely arrive as tidy briefs.
Digital products rarely begin with a complete specification. More often, someone has a problem that keeps getting in the way, a conversation that raises more questions than answers, or a rough sketch of what might make things better.
I quite like that stage. There is something interesting about taking an idea that is not yet fully formed and gradually working out what it could become. But I have also seen how quickly a team can move from a vague problem to a very specific solution, often without stopping to ask whether it is the right one.
After years working across development, user experience and delivery, I have learnt that the challenge is not just turning an idea into software. It is understanding the problem well enough to make sensible decisions about the experience, the technology, the effort involved and the people who will eventually have to use and maintain it.
A useful delivery plan does more than give a project some dates. It gives everyone a shared understanding of what matters, what is still unknown and what needs to happen next.
02 / Start with the problem
The feature list can wait.
One of the easiest ways to make an early idea feel real is to start writing down features. We need a dashboard. We need an approval workflow. We need an app.
Those may turn out to be the right answers. But at that point, they are still answers to questions we have not properly explored.
Before discussing screens, integrations or frameworks, I want to understand four things:
Who is experiencing the problem?
Not only the end users, but also those responsible for operating, supporting and administering the product.
What are they trying to achieve?
Describe the task in ordinary language, without prescribing a feature.
Where does it get difficult?
Look for delays, repeated effort, confusing information, handovers and workarounds.
What would better actually look like?
Less uncertainty? A task completed with fewer steps? Clearer decisions? A process that is easier to support?
Take the request, “We need a dashboard.” I would rather start with, “Our operations team needs to spot exceptions quickly and know what to do about them.” Now we have something useful to investigate.
The outcome might still be a dashboard. It might also be an alert, a simpler workflow or a small change to an existing system. Understanding the problem keeps those options open.
03 / Test assumptions
The things we assume deserve a closer look.
Almost every early product conversation contains assumptions. We assume users will understand a status, that the available data is good enough, that an integration is straightforward or that a particular feature is what customers really need.
There is nothing wrong with starting from an assumption. The problem is treating it as a fact without realising we have done so.
I find it useful to write these things down and turn them into questions the team can actually answer:
Make the assumption visible
What are we assuming? Be specific. For example, “Users will understand the difference between these two statuses.”
Understand the risk
What happens if we are wrong? Is it a minor inconvenience or something that could undermine the entire experience?
Find the smallest useful test
A conversation, a sketch, a quick prototype, a technical spike or a closer look at existing evidence.
Keep what we learnt
Capture the result and the decision, not just the activity.
Not every question needs a formal research programme, and not every idea needs a fully polished prototype. The point is to spend time where the uncertainty could be costly.
I would rather discover that an approach does not work while it is still a sketch than after a team has spent weeks building it.
04 / The first useful slice
Make the first version smaller, not incomplete.
Once the problem and the important assumptions are clearer, the temptation is often to put everything into the first release. Every stakeholder has something they would like included, and each addition can sound reasonable on its own.
But a first version needs a clear purpose. I prefer to think of it as the first useful slice: one coherent journey that gives the intended user a meaningful outcome and gives the team something real to learn from.
That usually means identifying:
A clear need
One user or operational problem we understand well.
An end-to-end task
One meaningful task someone can actually complete.
Decision-making context
Enough information to understand what is happening and what to do next.
Proper feedback
Thoughtful loading, empty, error and permission states.
A way to learn
Evidence of what works and where users still struggle.
Room to grow
A sensible way to extend the product without starting again.
There is an important distinction here. Small does not mean unfinished. It is perfectly reasonable to leave secondary features for later. It is much harder to justify leaving out the feedback, accessibility or recovery paths someone needs to complete the primary task.
A short scenario is often enough to bring everyone together: who is using the product, what they are trying to do, what they need to know, and what should happen when they finish.
05 / Map the experience
Work through the journey before polishing screens.
A collection of screens can look convincing without telling us whether the experience makes sense. I have found that a simple journey map or service flow often exposes more than a polished interface does at the start.
Walk through the experience in order:
Entry
How does someone arrive, and what do they already know?
Orientation
What must they understand before they can get started?
Decision
What choices do they have, and what information helps them choose?
Action
What happens when they do something, and how do they know it worked?
Recovery
What if information is missing, the connection fails or they make a mistake?
Completion
How do they know they are finished, and what happens next?
It does not need to be complicated. A few boxes and arrows can be enough to reveal that an important decision is missing, an exception has no owner or someone will need to move between three systems to finish one job.
Those are useful discoveries to make before the design starts looking finished.
06 / Make delivery real
A plan should help the team make decisions, not just track dates.
This is where product thinking, user experience, technology and delivery need to work together. They cannot be separate conversations that only meet when something goes wrong.
A practical delivery plan should make the shape of the work visible: the outcome we are aiming for, the first useful scope, the decisions we still need, the dependencies we cannot ignore and the people responsible for moving things forward.
I want to know what is technically uncertain, which integrations or data sources need investigating, what needs to be reviewed with stakeholders, and how the product will be tested and supported once it is live.
It is also worth being clear about what we are not doing in the first release. A decision to leave something out is still a decision, and it deserves a reason.
Milestones matter, of course. So do budgets and resources. But a plan becomes much more useful when those things are connected to actual decisions, evidence and working outcomes, rather than simply a list of activities marked complete.
07 / Keep learning
Clarity is something you keep working at.
Even a well-considered plan will change. Research reveals something unexpected. A technical constraint becomes clearer. Someone using an early version shows us a better way.
That should not automatically be seen as a failure of planning. The purpose of doing this thinking early is not to predict every detail. It is to give the team enough shared understanding to respond sensibly when new information arrives.
For me, that is the difference between a plan that looks reassuring in a presentation and one that genuinely helps people deliver.
An idea becomes easier to build when we can explain the problem, the outcome, the first meaningful experience, the important unknowns and the next decision. Not everything needs to be certain. We just need to be clear about what we know, what we do not, and what we are going to do about it.