How to Vet a Microsoft Fabric / Power BI Consultant (7 Questions)
- doramadhusudan
- Jun 30
- 4 min read
If you've ever searched a forum for "recommended Power BI consultants" or "should we hire a firm to set up our data warehouse," you already know the real fear: paying for months of work and ending up with a slow, half-finished platform nobody trusts.
It happens constantly. The data world is full of firms that win the pitch on buzzwords "unified analytics," "single source of truth," "AI-ready" and then deliver dashboards that break, pipelines that silently fail, and a bill that doubled mid-project.
The good news: you can spot a weak partner in a single scoping call. You don't need to be technical. You just need to ask the right questions and listen for specific, confident answers instead of vague reassurance.
Here are the seven that matter most.
Question 1. "Can you walk me through a Fabric (or Power BI) project you delivered - with the actual before/after numbers?"
Why it matters: Real practitioners talk in specifics. Weak firms talk in adjectives.
A strong answer sounds like: "We migrated a retailer from a nightly SQL Server warehouse to a Fabric Lakehouse. Report refresh dropped from 6 hours to 20 minutes, and we cut their cloud spend ~30% by consolidating three tools into OneLake."
A weak answer sounds like: "We've done lots of Fabric work, it's been very successful, clients love it."
Ask for *one* project, end to end. If they can't name a metric - refresh time, rows processed, cost saved, hours of manual reporting eliminated - they either didn't own the outcome or it didn't go well.
---
Question 2. "Who specifically will do the work, and what are their certifications?"
Why it matters: Many firms sell with senior architects and deliver with junior contractors you never met.
Ask for the *named* people on your project and their Microsoft certifications (e.g. **DP-700 Fabric Analytics Engineer**, **PL-300 Power BI Data Analyst**). Ask whether the person on the sales call will actually be hands-on.
---
Question 3. "How do you scope and price the work and what happens when scope changes?"
Why it matters: The #1 way data projects go wrong financially is open-ended time-and-materials with no guardrails.
Strong partners offer a fixed-scope first phase - often a proof of concept or discovery sprint so you can judge them on a small, defined deliverable before committing to a big build. They'll also tell you, plainly, how change requests are handled.
Red flag: Reluctance to put any phase on a fixed price, or vague "we'll figure out scope as we go."
---
Question 4. "What's your approach when the project is *not* a good fit for Fabric?"
Why it matters: An honest consultant will sometimes tell you Fabric (or a full warehouse) is overkill. A firm that recommends the same expensive platform to everyone is selling, not advising.
A strong answer: "If your data is small and your needs are simple, we'll tell you a lighter setup is cheaper and we'd rather you trust us for the next project."
Red flag: Everything is always a Fabric Lakehouse, no matter what you describe.
---
Question 5. "How will you make sure we can run this after you leave?"
Why it matters: A platform you can't maintain is a hostage situation. Some firms intentionally build complexity so you stay dependent on them.
Ask about documentation, knowledge transfer, and training your team. A good partner wants you self-sufficient and earns repeat work by being good, not by locking you in.
Red flag: No mention of handover, docs, or training and just "we'll manage it ongoing for a monthly fee."
---
Question 6. "How do you handle data governance, security, and refresh reliability?"
**Why it matters:** Anyone can build a dashboard that looks good in a demo. The hard part is a platform that's *secure, governed, and doesn't silently break at 6am.*
Listen for concrete practices: row-level security, workspace/permission structure in Fabric, monitoring and alerting on pipeline failures, source-of-truth definitions. These are the things that separate a real data engineer from a "report builder."
**Red flag:** Hand-waving on security and "the reports just refresh automatically."
---
Question 7. "Can I talk to a past client?"
Why it matters: This is the simplest test and the one weak firms dodge. A consultant with happy clients will happily connect you with one.
If references are "confidential," slow to materialize, or never offered — assume the worst.
Question | Strong signal | Weak signal |
Past project + numbers | Named metric (refresh, cost, hours) | "Very successful" |
Named team + certs | DP-700 / PL-300, real names | "Right resources" |
Scoping & pricing | Fixed-scope first phase | Open-ended T&M only |
When Fabric isn't right | Will recommend something cheaper | Always Fabric |
Handover | Docs + training your team | "We'll manage it" |
Governance & reliability | RLS, monitoring, alerting | "It just refreshes" |
Reference | Connects you to a client | Dodges |
Hiring a Microsoft Fabric or Power BI consultant isn't about who has the slickest pitch it's about who can show specific, proven outcomes and who's honest enough to tell you when *not* to spend money. The seven questions above will surface that in under an hour.
At Aptocoiner Analytics, we answer every one of these with names, numbers, and a fixed-scope proof of concept so you can judge the work before you commit to the build.
Comments