A fix to one broken report takes hours. A single dashboard on clean data takes days. A first reporting system for a small or mid-sized business takes about three weeks, which is why my build sprint is fixed at exactly that. A multi-department estate takes two to six months, and anyone quoting it tighter without a discovery phase is guessing.
Those are build times, though, and the build is rarely what makes a project late. Data quality, access to systems and decisions nobody has made yet are what stretch a Power BI timeline, and none of them show up in the quote.
Here are the honest ranges by project shape, and then the part that actually determines whether you hit them.
How long does each type of Power BI work take?
Every duration in this table is published on my site, alongside the price, as of September 2026. They're worth using as anchors even if you never buy from me, because a supplier who won't put a duration next to a price is telling you the duration is negotiable in one direction only.
| Project shape | Published example | Typical duration | What moves it |
|---|---|---|---|
| Getting unstuck on one problem | Unblocking session, £100 | One session | Whether it's really one problem |
| A ready-made dashboard on your own data | The Setup Hour, £195 | One hour | How your data is stored |
| A scored review of your whole estate | Health Check, £950 | Results within five working days | Access granted before day one, or not |
| Training a team | Training day, from £750 | Half-day prep plus one or two days | Licences and installs, which the prep exists to absorb |
| A first reporting system | Reporting Sprint, £9,500 | Three weeks, live at the end | Data quality and decision speed |
| Migrating existing reports | Migration, £600 per report | Free one-day assessment, then per report | How many versions of the truth the old reports hold |
| A Fabric go-or-wait decision | Fabric Readiness Assessment, £5,950 | Five days of work over about three weeks | Stakeholder availability |
| A multi-department estate | Scoped per project | Two to six months, discovery first | Everything in the next section |
The three-week sprint is capped at three source systems and eight report pages, and that cap is the reason the timeline holds. When someone quotes you three weeks with no cap on scope, one of the two numbers is fiction.
Why is the build the short part?
Connecting a system Power BI already understands, modelling the data properly and designing pages people will actually read is skilled work, but it's predictable work. A consultant who has done it many times knows roughly how long a three-source model takes to build, the same way a kitchen fitter knows how long a kitchen takes.
What the kitchen fitter can't predict is discovering the wall is damp. In reporting, the damp wall is your data. The build estimate assumes the data says what everyone believes it says, and in most businesses it doesn't quite. Customer names duplicated, dates missing, codes reused, a field everyone swears is populated that stopped being populated in 2023. None of that is visible from outside, and all of it lands inside your timeline.
That's why I treat the build estimate and the project estimate as different numbers. The build is mine to control. The rest is shared.
What actually stretches a Power BI timeline?
Four things, in the order they usually bite.
Data quality. The single biggest variable. If margin is calculated three different ways in three spreadsheets, someone has to decide which one is right before a model can encode it, and that conversation can take longer than the modelling. This is one reason six reports showing six different numbers is such a common starting point: the disagreements were always there, and a modelling project is the first time anyone has to resolve them rather than work around them.
Access. Getting a consultant read access to your systems sounds like a five-minute job and routinely takes two weeks. There's an IT provider to go through, a licence to assign, a security review, a person on holiday. Every day of it comes off the front of the project while the clock, in everyone's head, is already running.
Decisions. A reporting project generates questions: which definition of revenue, which customer grouping, who should see what. Each one is quick to answer and impossible to skip. If the person who can answer them is in one meeting a fortnight, the project moves at one meeting a fortnight.
Your team's time. The people who understand the data are always the busiest people in the business. The month-end they can't skip, the audit that lands mid-project. This isn't a complaint, it's physics, and a realistic plan works around it rather than pretending it away.
Notice what isn't on the list: the technology. Power BI itself is almost never the bottleneck.
How do you keep a Power BI project on schedule?
Four things, and they're all yours to do rather than the consultant's.
Sort access before the start date, not on it. Ask your consultant for the exact access list in writing, hand it to whoever controls your systems, and don't agree a start date until it's done. This one habit removes the most common fortnight of dead time in the industry.
Name one decision-maker. Not a committee. One person who can say "margin means this" and make it stick, and who commits to answering questions within a day or two. Committees are how three-week projects become three-month projects.
Accept a smaller first release. Eight pages that ship in three weeks beat thirty pages that ship in June. Reporting is one of the few purchases where the smaller version genuinely tells you whether the bigger version is worth buying.
Ask what happens when the data is worse than expected. It's the likeliest surprise, so ask up front: does the price move, does the timeline move, or does scope come out? A supplier with a ready answer has hit the problem before. A supplier who says the data will probably be fine hasn't looked at much data.
When is a fixed timeline honest, and when is it a guess?
A fixed timeline is honest when the scope is capped and the assumptions are written down. My sprint can be three weeks because it's capped at three sources and eight pages, and because week one is deliberately front-loaded with the discovery that surfaces data problems while there's still time to react.
A fixed timeline on an uncapped scope is a guess dressed as a commitment, and it fails in a predictable way: the deadline arrives, the work isn't done, and now you choose between a late project and a rushed one. For anything beyond the small end, the honest structure is discovery first, then a fixed proposal, which is exactly how I price larger builds. Paying for a small discovery phase to get a real number beats getting a fictional number free.
If you want the money side of the same question, what Power BI work costs in the UK covers day rates and project pricing in the same spirit.
How much of your own time should you budget?
Less than you might fear, but more than zero, and concentrated at the edges.
At the front: a scoping conversation, the access admin, and an hour or two walking the consultant through how the business actually runs. In the middle: quick answers to definition questions, ideally within a day or two of being asked. At the end: reviewing the numbers against what you know to be true, and a handover session so your team owns what was built.
For a three-week build, that's perhaps a day and a half of your people's time in total. The projects that go wrong are rarely the ones where the client gave too little time overall. They're the ones where the answers came slowly.
Common questions
How long does it take to build a single Power BI dashboard?
On clean, accessible data, a competent consultant builds a single well-designed dashboard in two to five days, including proper data modelling behind it. If the data needs significant cleaning first, or nobody can agree what the numbers should be, it can take two or three times that, and the extra time is spent on the data rather than the dashboard.
Can a Power BI project be delivered in a week?
A small one, yes, if access is sorted before day one and one person can answer definition questions same-day. A fix, a single report or a proof of concept fits in a week comfortably. A reporting system doesn't, and a supplier promising one in a week is either scoping something much smaller than you think you're buying or planning to skip the reconciliation.
How long does a Power BI health check or audit take?
Mine returns a scored report card within five working days: 48 checks across six areas, each scored against a written standard, with a walkthrough call after. Larger estate audits take longer, but five working days is a reasonable benchmark for a review of a typical SME estate, provided access is granted at the start.
What is the slowest part of a Power BI project?
Almost always the data, and specifically the decisions the data forces. Duplicates, missing values and disagreeing definitions all need a human ruling before a model can encode the answer. The technology is quick. The rulings are slow, and they're slowest when nobody on the client side owns them.
Why do Power BI projects overrun?
The usual causes, roughly in order: access to systems arriving weeks late, data quality being worse than anyone admitted, scope growing quietly because nothing was capped, and decisions queuing behind a committee. Note that none of these is a technology problem, which means all of them are preventable in the first week.
How long does it take to migrate existing reports to Power BI?
The honest answer starts with an inventory. My version is a free one-day assessment that catalogues every report, bands the complexity and fixes the total price at £600 per report, so the duration becomes simple arithmetic rather than a hope. Migrations stretch when the old reports disagree with each other, because each disagreement needs resolving before the new version can ship.
Do I need to free up my own staff for a Power BI project?
Yes, but modestly and predictably: scoping and access at the start, quick answers to definition questions in the middle, review and handover at the end. Budget a day or two of your best-informed person's time across a small project. What derails timelines isn't the total hours, it's answers that take a fortnight to arrive.
Durations and prices for my own services are published on the linked pages, as of September 2026. If you'd rather talk through what your project actually is before believing anyone's estimate, the free 20-minute intro call is for exactly that.