Most healthcare data leaders aren’t struggling with analytics.They’re struggling with inherited chaos — 10+ years of conflicting definitions, EMR exports nobody trusts, and leadership demanding answers from numbers that don’t align.

I partner with the people living that world every day: VPs of Analytics, Directors of IT
and cross-functional data leaders who are drowning in Excel reconciliation, political misalignment, and dashboards that spark more debate than decisions.

If your finance, operations, and EMR teams all have different definitions of “Net Revenue,” I fix that.

I build the data foundations that make AI possible — and trustworthy.

DM me "AI in Healthcare" if you're ready to get your data foundation in order.

Migrated to Microsoft Fabric. Data went missing.
 
(One API call fixed it.)
 
Microsoft Fabric looks perfect on paper.
 
Unified analytics platform. Simplified architecture. Modern cloud solution.
 
Then you migrate and things break in ways Microsoft's docs don't mention.
 
𝗧𝗵𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺:
 
Stored procedure loads successfully into the warehouse.
 
Power BI shows different totals than lakehouse.
 
Query the lakehouse directly? Data's there.
 
Check the warehouse? Missing rows.
 
Hours of troubleshooting. Nothing wrong with the pipeline. Nothing wrong with permissions.
 
𝗧𝗵𝗲 𝗰𝘂𝗹𝗽𝗿𝗶𝘁:
 
Metadata refresh delay.
 
Fabric warehouses don't immediately expose new data.
 
There's a background process that syncs metadata. It can lag.
 
𝗧𝗵𝗲 𝗳𝗶𝘅:
 
One API call to manually trigger metadata refresh in the Fabric Data Factory pipeline.
 
Data appears instantly.
 
Problem solved.
 
𝗟𝗲𝘀𝘀𝗼𝗻𝘀 𝗳𝗿𝗼𝗺 𝗙𝗮𝗯𝗿𝗶𝗰 𝗶𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻𝘀:
 
𝗠𝗲𝘁𝗮𝗱𝗮𝘁𝗮 𝗿𝗲𝗳𝗿𝗲𝘀𝗵 𝗶𝘀𝗻'𝘁 𝗮𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗰
 
Plan for manual triggers or delays in data visibility.
 
𝗪𝗮𝗿𝗲𝗵𝗼𝘂𝘀𝗲 𝗯𝗲𝗵𝗮𝘃𝗶𝗼𝗿𝘀 𝗱𝗶𝗳𝗳𝗲𝗿 𝗳𝗿𝗼𝗺 𝗦𝗤𝗟 𝗦𝗲𝗿𝘃𝗲𝗿
 
Don't assume parity. Test everything.
 
𝗗𝗼𝗰𝘂𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 𝗹𝗮𝗴𝘀 𝗿𝗲𝗮𝗹𝗶𝘁𝘆
 
Microsoft's docs won't cover every edge case. Build buffer time for discovery.
 
𝗣𝗲𝗿𝗺𝗶𝘀𝘀𝗶𝗼𝗻𝘀 𝗮𝗿𝗲 𝗰𝗼𝗺𝗽𝗹𝗲𝘅
 
Workspace roles, warehouse permissions, semantic model access—all separate layers.
 
𝗦𝘁𝗮𝗿𝘁 𝘀𝗺𝗮𝗹𝗹
 
Migrate one critical workload first. Validate. Then scale.
 
𝗧𝗟;𝗗𝗥:
 
Fabric is powerful but has quirks Microsoft doesn't advertise.
 
Metadata refresh delays. Warehouse differences. Permission complexity.
 
Test thoroughly. Build buffer time. Don't assume it'll work like SQL Server.
 
♻️ Share this if you've troubleshot invisible data in Fabric Warehouse.
 
Follow me for real talk on modern data platforms.

Christian Steinert

I help healthcare data leaders with inherited chaos fix broken definitions and build AI-ready foundations they can finally trust. | Host @ The Healthcare Growth Cycle Podcast

"We need better analytics."
 
(Turn that into hard dollar savings executives can't ignore.)
 
Data leaders hear this constantly.
 
"We need better insights."
 
"Our reporting isn't good enough."
 
"Can you improve our analytics?"
 
𝗧𝗵𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺:
 
Executives hear vague improvement requests.
 
They don't approve budgets for "better."
 
They approve budgets for ROI.
 
𝗧𝗵𝗲 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸:
 
𝗖𝗮𝗹𝗰𝘂𝗹𝗮𝘁𝗲 𝗹𝗮𝗯𝗼𝗿 𝗰𝗼𝘀𝘁 𝘀𝗮𝘃𝗶𝗻𝗴𝘀
 
"Finance spends 15 hours/week on manual reporting. That's $45K/year. Automation reduces this to 2 hours."
 
Hard dollars. Easy math.
 
𝗤𝘂𝗮𝗻𝘁𝗶𝗳𝘆 𝘁𝗶𝗺𝗲 𝘀𝗮𝘃𝗶𝗻𝗴𝘀
 
"Current process takes 90 minutes daily. New solution takes 10 minutes. That's 6.5 hours/week recovered."
 
𝗦𝗵𝗼𝘄 𝗿𝗲𝘃𝗲𝗻𝘂𝗲 𝗶𝗺𝗽𝗮𝗰𝘁
 
"Faster reporting means decisions happen 3 days sooner. In Q3, that delay cost us $200K in missed opportunity."
 
𝗣𝗿𝗼𝗷𝗲𝗰𝘁 𝗽𝗮𝘆𝗯𝗮𝗰𝗸 𝗽𝗲𝗿𝗶𝗼𝗱
 
"$100K investment. $83K annual savings. 14-month payback. 141% ROI over 5 years."
 
Executives understand payback.
 
𝗪𝗵𝗮𝘁 𝗰𝗵𝗮𝗻𝗴𝗲𝘀:
 
Instead of "We need better dashboards"
 
Say: "Automating this report saves $45K annually in labor costs with 12-month payback."
 
Instead of "Our data quality is poor"
 
Say: "Poor data costs us 20 hours/week in reconciliation. That's $60K/year we're wasting."
 
𝗧𝗵𝗲 𝗹𝗲𝘀𝘀𝗼𝗻:
 
Stop talking about insights.
 
Start calculating dollars, hours, and payback periods.
 
Make it impossible to say no.
 
𝗧𝗟;𝗗𝗥:
 
Executives don't approve "better analytics."
 
They approve measurable ROI.
 
Calculate labor savings. Quantify time. Show payback.
 
Speak their language.
 
♻️ Share this if you've struggled to get analytics budgets approved.
 
Follow me for frameworks that actually get data projects funded.

Pipelines fail Friday morning. Presentation at 10 AM.
 
(One Discord message to your network saved the deal.)
 
It's Friday, 7:30 AM.
 
Client's revenue dashboard broken. Presentation to executives at 10 AM.
 
Pipeline failed overnight. No clear error. Clock ticking.
 
𝗪𝗵𝗮𝘁 𝗺𝗼𝘀𝘁 𝗽𝗲𝗼𝗽𝗹𝗲 𝗱𝗼:
 
Panic. Google frantically. Open support tickets that won't get answered for days.
 
𝗪𝗵𝗮𝘁 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝘄𝗼𝗿𝗸𝘀:
 
Message your network.
 
Post the issue in a private Discord community of data engineers.
 
Three people responded within 10 minutes.
 
One had seen the exact error before. Shared the fix.
 
Pipeline running by 8:45 AM. Dashboard refreshed. Presentation saved.
 
𝗪𝗵𝘆 𝗻𝗲𝘁𝘄𝗼𝗿𝗸 𝗺𝗮𝘁𝘁𝗲𝗿𝘀:
 
Technical skills get you hired.
 
Your network keeps you employed.
 
When things break (and they will), documentation won't save you.
 
Your network will.
 
𝗛𝗼𝘄 𝘁𝗼 𝗯𝘂𝗶𝗹𝗱 𝗮 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗰 𝗻𝗲𝘁𝘄𝗼𝗿𝗸:
 
𝗝𝗼𝗶𝗻 𝗰𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝗶𝗲𝘀
 
Slack groups, Discord servers, LinkedIn groups for your tech stack.
 
𝗚𝗶𝘃𝗲 𝗯𝗲𝗳𝗼𝗿𝗲 𝘆𝗼𝘂 𝘁𝗮𝗸𝗲
 
Answer others' questions. Share what you learn. Build reciprocity.
 
𝗦𝘁𝗮𝘆 𝗮𝗰𝘁𝗶𝘃𝗲
 
Don't just lurk. Engage regularly so people know you.
 
𝗕𝘂𝗶𝗹𝗱 𝗿𝗲𝗮𝗹 𝗿𝗲𝗹𝗮𝘁𝗶𝗼𝗻𝘀𝗵𝗶𝗽𝘀
 
Coffee chats. Conference meetups. Real conversations beyond tech.
 
𝗧𝗵𝗲 𝗿𝗲𝗮𝗹𝗶𝘁𝘆:
 
Your network is your backup plan when documentation fails.
 
It's faster than support tickets.
 
More reliable than Stack Overflow.
 
And built on relationships, not transactions.
 
𝗧𝗟;𝗗𝗥:
 
Technical problems get solved faster with strong networks.
 
Invest in relationships before you need them.
 
Give value. Build reciprocity. Show up consistently.
 
♻️ Share this if your network has saved you from a crisis.
 
Follow me for real talk on what actually keeps data careers sustainable.

Pipelines broke. Revenue reporting went dark.
 
(We fixed it and didn't charge the client a dime.)
 
Most consultants would've invoiced extra hours.
 
We built a fallback solution instead.
 
𝗛𝗲𝗿𝗲'𝘀 𝘄𝗵𝗮𝘁 𝘀𝗲𝘁𝘀 𝘁𝗼𝗽 𝗱𝗮𝘁𝗮 𝘁𝗲𝗮𝗺𝘀 𝗮𝗽𝗮𝗿𝘁:
 
𝗙𝗼𝗹𝗹𝗼𝘄 𝘁𝗵𝗿𝗼𝘂𝗴𝗵 𝗼𝗻 𝘀𝗺𝗮𝗹𝗹 𝘁𝗵𝗶𝗻𝗴𝘀
 
Said you'd update the Jira ticket daily? Do it.
 
Promised to coordinate a meeting? Handle it promptly.
 
Small commitments build as much trust as big ones.
 
𝗦𝘁𝗮𝘆 𝘁𝗿𝗮𝗻𝘀𝗽𝗮𝗿𝗲𝗻𝘁 𝘄𝗵𝗲𝗻 𝘁𝗵𝗶𝗻𝗴𝘀 𝗯𝗿𝗲𝗮𝗸
 
Communicate bottlenecks immediately.
 
Focus on the solution path forward.
 
Provide status updates while fixing it.
 
Hiding never helps. Clients appreciate knowing what's happening.
 
𝗕𝘂𝗶𝗹𝗱 𝘁𝗵𝗲 𝗿𝗲𝗹𝗮𝘁𝗶𝗼𝗻𝘀𝗵𝗶𝗽
 
Not every interaction should be 100% business.
 
Talk about the weekend. Add humor where appropriate.
 
Stay optimistic. Nobody wants negativity from their data team.
 
𝗦𝗵𝗼𝘄 𝘂𝗽 𝗽𝗿𝗲𝗽𝗮𝗿𝗲𝗱
 
Have agendas for working sessions.
 
Document metric logic beforehand for quick reference.
 
Review meeting notes before you join.
 
Wasted time erodes trust faster than anything.
 
𝗥𝗲𝘃𝗶𝗲𝘄 𝗽𝗿𝗼𝗴𝗿𝗲𝘀𝘀 𝗳𝗿𝗲𝗾𝘂𝗲𝗻𝘁𝗹𝘆
 
Bi-weekly sprint retros. What's working? What's not?
 
Quarterly one-on-ones with stakeholders for feedback.
 
Course-correct before small issues become big problems.
 
𝗦𝗲𝗲𝗸 𝗻𝗲𝘄 𝗼𝗽𝗽𝗼𝗿𝘁𝘂𝗻𝗶𝘁𝗶𝗲𝘀
 
Don't just execute the SOW.
 
Study the business. Listen to industry podcasts. Read domain newsletters.
 
Proactively identify opportunities clients haven't thought of yet.
 
That's how teams become indispensable.
 
𝗧𝗟;𝗗𝗥:
 
Technical work is table stakes.
 
Going the extra mile on relationships, transparency, and proactive value? That's what gets remembered.
 
P.S. - Full breakdown in this week's newsletter. Link in comments. 👇
 
♻️ Share this if you believe relationships matter as much as technical skills.
 
Follow me for real talk on what separates good data teams from great ones.

taplio

Christian Steinert's Linkedin Analytics