Power BI consultant role: scoping, estimating and client delivery

What changes when there is a client

Inside a company, a BI developer can revise a dashboard every week without anyone counting hours. A consultant cannot. Every engagement has a scope, a budget in hours or a fixed fee, and a sign-off. Three consequences shape the job:

• Discovery is structured. You must learn the client's sources, definitions and users quickly and write down what you learned.
• Estimates are commitments. An underestimate is paid for by your firm or yourself.
• Scope creep is the normal state. Every demo prompts new requests. Your job is to say "yes, as a change request" clearly and pleasantly.

The artefact: an effort estimate inside a statement of work

Here is the estimate section of a hypothetical statement of work (SOW) for a distributor that wants a sales and collections dashboard from its ERP and a spreadsheet of targets.

The assumptions column does more work than the hours column. "Read-only SQL access provided within week 1" means that if access takes three weeks, the timeline moves and the SOW already says so. "Up to 25 measures" and "5 pages" draw a line that makes the 26th measure a conversation, not an argument. "One UAT round" prevents endless rounds of revisions.

To turn hours into a quote, a firm multiplies by its rate card and adds its margin; a freelancer applies their own rate. Either way, the quote is only as sound as the hours and assumptions above it, so compare actual hours against this table at the end of every engagement and adjust the next estimate.

Out of scope, stated plainly

The SOW should also list exclusions: mobile-specific layouts, data quality fixes in the ERP itself, additional data sources, Fabric or premium capacity licensing, and support after the handover period. Writing them down protects both sides.

A sample fortnight on an engagement

• Days 1–2: workshops with sales and finance heads; agree the definition of "collection" (receipt date versus bank credit date).
• Days 3–5: extract, model and first reconciliation; discover that returns are posted as negative invoices in one branch.
• Days 6–8: measures and report pages; mid-point demo, which produces four new requests. Two fit within scope; two become a change request.
• Days 9–10: row-level security, refresh setup with client IT, and draft documentation.

Estimate accuracy, utilisation and repeat clients

• Delivery within estimate and on time.
• Utilisation — billable hours against available hours.
• Client satisfaction and repeat work.
• Quality: defects found after handover, and whether reconciliations held.
• Change requests converted into paid extensions, at firms that track it.

Practice managers, client sponsors and client IT

At a firm: a delivery or practice manager, with a project manager on larger engagements. Freelancers report to the client sponsor directly. You work with client business users, client IT (for access, gateways and security), sometimes the client's data engineers, and your firm's sales or presales team.

Confused with Power BI developer and data analyst

The developer builds in-house; the consultant builds for clients and owns the commercial side of delivery. A data analyst uses reports to answer questions; the consultant builds the reporting layer so others can. Many consultants began as in-house developers or reporting analysts.

Client-facing, Fabric and PL-300 in consulting postings

• "Client-facing, requirement gathering" — discovery and communication are core.
• "Power BI, DAX, Power Query, data modelling" — the build skills.
• "Microsoft Fabric, Azure Synapse, Dataflows" — larger data platform work.
• "Presales support, estimation" — you help win work, not just deliver it.
• "PL-300 certified preferred" — Microsoft's Power BI data analyst certification.

Estimate, deliver, compare actual hours

Month one: shadow an engagement and rebuild its model yourself. Month two: own the build for one scoped engagement under review. Month three: write an estimate for a new request and have it checked against actual hours later. Next roles include senior consultant, BI practice lead, solutions architect or independent consultant. Consultants still need the build skills first; star-schema modelling, DAX and row-level security are core modules of the Data Analyst programme.