Large business technology projects rarely fail because the team lacks skill. They fail because the project is so big that nobody can say what to do on a Tuesday morning. A system replacement, a data clean-up or a reporting overhaul is announced with a date and a budget, then sits in a status column while everyone busies themselves with smaller, clearer work. The remedy is a plain one: break the business technology project into fifteen-minute steps, each with a visible result. What follows takes the queries that tend to surface when a team first tries this, and answers them in order.
What does a fifteen-minute step mean for a business technology project?
It is a unit of work that one person can finish, and check, in a quarter of an hour. The test is whether you can describe the finished state in one sentence. “Export the customer list from the old system into a spreadsheet” qualifies. “Prepare the data” does not, because nobody can tell when it is done. Steps of this size are not an attempt to make the project shorter; the total effort stays the same. What changes is that every step is startable, and the finished steps accumulate into evidence that the project is moving.
Why does the project stall before it begins?
Because the opening move is ambiguous. Large initiatives begin with abstract verbs such as plan, evaluate, align and prepare, and abstract verbs do not tell a person where to put their hands. Each time someone sits down to start, they have to invent the first action, and inventing is tiring. People postpone tiring decisions. Slicing the project in advance moves the inventing to one deliberate planning session, so that later sittings consist only of doing.
How do I find the right size when I slice?
Cut in layers. First divide the project into a handful of workstreams, for instance data, access, reporting and training. Then divide each workstream into deliverables. Finally divide each deliverable into steps until every step passes two checks: it has a one-sentence finished state, and you can picture it fitting in fifteen minutes. If you hesitate, cut again. It is better to end with steps that take ten minutes than with steps that quietly swell to ninety.
Can you give a worked example?
Take a reporting overhaul. The workstream “monthly sales report” might split into these quarter-hour steps.
- List the five questions the report must answer, in one paragraph.
- Name the source system for each question.
- Pull one month of raw figures for one question into a sheet.
- Check the totals against last month’s circulated version and note any gap.
- Draft the table layout on paper.
- Build that one table.
- Ask a colleague to read it and write down what they thought it meant.
Each item has a result you can point to. After seven sittings you have one trustworthy table, and a clear sense of how the next four will go. For a wider orientation on this kind of work, the overview titled how to approach business technology solutions on this blog covers the broader landscape.
Who should own the slicing?
The person accountable for the outcome should draft it, because they hold the full picture, but they should not slice alone. Ask the people who will do the work to challenge each step. They will spot the hidden dependency, the missing access request or the approval that takes a week. A good slicing session ends with every step carrying a single owner and a single finished state, and with the awkward waits exposed on the page instead of buried in a plan.
What if a step overruns?
Treat the overrun as information, not failure. Stop at the quarter hour, write down what remains and split it into one or two new steps. Resist the temptation to extend “just a bit longer”, since doing so teaches the team that fifteen minutes is a suggestion and the plan loses its predictability. If steps overrun repeatedly in one workstream, the slicing in that area was too coarse, or the work depends on something unknown. Either finding is worth knowing early.
How many steps are too many?
A project of a few months can reasonably hold a couple of hundred steps, but nobody should stare at all of them at once. Keep the full list in a backlog and show only the next five to ten. A short visible queue gives the calm of a small task list while the larger plan stays intact behind it. Revisit the backlog weekly, remove what no longer applies and add what has been uncovered.
Does this work for tasks that need deep concentration?
Mostly yes, with an adjustment. Some work, such as designing a data model, resists being chopped into tiny pieces because the thinking needs a long run-up. For those items, use quarter-hour steps to surround the deep work rather than to replace it. One step gathers the inputs, another writes the question to be answered, a third reviews the output afterwards. The core thinking can then happen in a longer block, entered with everything prepared and no time wasted on logistics.
How do I report progress to stakeholders?
Count finished steps against the planned total for each workstream, and show a short list of what was completed this week. Stakeholders respond well to specifics. “Fourteen of forty steps done in the data workstream, including the customer export and the first reconciliation” is clearer than “good progress, on track”. Be candid about steps that revealed new work, because the credibility of the count depends on it staying honest.
What tools do I need?
Very little. A shared list with columns for step, owner, finished state and status is enough. Teams often reach for heavyweight project software first and then spend their first week configuring it, which defeats the purpose. Start with something that takes five minutes to set up, learn what you actually need, and only then consider more. The method matters far more than the container.
Is there a role for outside platforms and services?
Often there is, and the same slicing logic helps you integrate them. When a project involves connecting several web services, sort the work so each connection has its own short steps: an account created, a test call made, a result checked. As a small illustration of services built from separate parts, MotoShareGulf is a website for hiring and sharing two-wheelers that hands user logins to a separate shared account service. The piece of work is cleanly separated from the rest, which is just how a good step should be. For the cloud and web side of such integrations, the discussion of bridging web solutions, cloud and artificial intelligence is a useful companion read.
How do I keep momentum after the first week?
Protect a regular slot, even a modest one, and begin each session by choosing the next step from the top of the queue so that you never face a blank page. Celebrate completion in a quiet way, by ticking the item and moving on. Momentum in a business project comes from repetition of small wins, not from occasional bursts of effort. Teams that hold a fifteen-minute slot every working day usually find that the project moves faster than the ones that wait for a clear afternoon.
What is the first thing to do today?
Open a blank document and write the project name at the top. Under it, write the single most uncomfortable part of the project and cut it into three steps that each could be finished in a quarter of an hour. Do the first one before you close the document. Do that and you have started, which is the hardest part. Tomorrow, repeat it with the next uncomfortable piece, and keep the habit going until the list runs out. A project that once looked like a wall will, within a couple of weeks, look like a column of ticked boxes.

