Don't Migrate Every Dashboard. Migrate the Ones That Run the Meeting.

The inventory says two hundred Tableau workbooks. The migration plan says migrate all of them. The budget will not survive the plan.

Most of those workbooks do not run a meeting. They were built for a one-off ask, a departed manager, or a pilot that never died. Treating file count as scope is how lift-and-shift consumes a year and still leaves Monday on Excel.

Don’t migrate every dashboard. Migrate the ones that run the meeting. Rebuild those. Archive the rest on purpose.

Meeting use is the prioritization signal

Mid-market ops, finance, and sales teams already know which views matter. Flash. Plant stand-up. Forecast call. Inventory review. Those cadences have owners, agendas, and consequences.

Workbench sprawl does not. Usage logs help. Meeting maps help more. If nobody can name the recurring decision a workbook supports, it is a candidate to leave behind—not a candidate for careful DAX translation.

Tableau to Power BI is a rebuild. Rebuild effort is scarce. Spend it where trust changes outcomes. Converter theater across the long tail is how programs miss the meetings that justified the project.

Prioritization is not neglect. It is how you fund the semantic model behind the pack leaders actually open.

How estates get treated like sacred inventories

License renewals arrive with a workbook count. Count becomes the project charter.

Authors fear deletion. “Someone might need it” keeps empty shells alive. The same tax unused columns put on refresh—except here the tax is migration hours.

Consultants bill against inventory completeness. Complete feels safer than ranked. Safer plans finish late.

Stakeholders hear “we are migrating Tableau” and imagine every personal view surviving. Expectation management never happens. Scope quietly includes nostalgia.

IT equates access continuity with risk reduction. Turning off a unused workbook feels riskier than funding it. The opposite is true for budget and attention.

The costs of migrating the long tail first

1. Budget burns on low-stakes sheets. Months disappear into rarely opened views while flash still depends on a fragile workbook.

2. The Monday pack waits. Leaders judge the migration by the meeting they attend. If that loop is last, the program looks failed while the backlog looks “on track.”

3. Teams learn the wrong habit. Lift-and-shift becomes the default. Grain and ownership never get designed. Sprawl relocates.

4. Parallel run explodes. Two hundred dual paths cannot be validated. Spot checks miss the KPI that blows up in the board pack.

5. Best authors tire out. Talent spends energy on trivia. When the hard measures arrive, the team is already burned.

6. Archive never happens. Everything is “in progress.” Nothing is retired. Licenses stay dual longer than the business case allowed.

7. Self-service stays blocked. Certified measures for the top meetings ship late. Analysts keep forking personal Tableau logic in the meantime.

8. Leadership loses patience. A year in, workbook percent-complete looks fine. Decision quality does not. Funding gets cut mid-rebuild.

How to fix it: score, sequence, retire

1. Map workbooks to meetings. For each recurring forum—flash, stand-up, forecast, inventory—list the views on the agenda. That shortlist is tier one.

2. Score the rest with three signals. Meeting use. Refresh failure pain. Measure complexity. High complexity without meeting use is a trap: expensive to rebuild, low return.

3. Publish a three-tier roadmap. Tier one: rebuild now. Tier two: rebuild after tier one earns trust. Tier three: archive, replace with self-service on certified measures, or delete.

4. Rebuild tier one as a semantic model loop. Not a sheet copy. Measures, grain, ownership, app, then retirement of the Tableau source for that meeting. Same discipline as certified datasets vs wild west.

5. Time-box tier one to a proof. One domain, one meeting, one retired workbook beats a forty-page tool comparison. Prove Power BI where it hurts today.

6. Say no in writing to tier-three rebuilds. Nostalgia needs an owner and a budget line. Without both, the answer is archive.

7. Use certified measures to absorb ad hoc demand. Many long-tail dashboards were personal answers to questions the model should own. Give analysts a governed model instead of two hundred conversions.

8. Retire on dates, not vibes. Each migrated meeting gets a Tableau turn-down date. Communicate it. Turn off access. Hope is not a decommission plan.

9. Re-score quarterly. New Tableau builds during migration are scope leaks. Freeze casual new workbooks or require them to land in Power BI.

10. Report progress as meetings moved—not workbooks converted. Executives understand “flash runs on Power BI.” They do not feel “47% of the estate migrated.”

What good looks like

Ninety days in, the three meetings that run the business open in Power BI. Finance defends the measures. Plant managers stopped asking for the old file for those loops.

The workbook inventory shrank because tier three was archived with a note, not pretended into the backlog. Dual licenses start trending down for real.

When someone asks “what about my regional side pack?” the answer is: use the certified model, or make a case for tier two. Not “we’ll convert everything eventually.”

A practical scoring workshop

Half day. Bring finance, ops, sales ops, and the BI owners.

List recurring meetings and the decisions each must make. Attach workbooks. Anything unattached goes to a parking lot.

Rank attached workbooks by how badly a wrong number hurts. Rebuild order follows pain—not alphabetical file names.

Leave with a one-page tier list, named owners for tier-one KPIs, and a thirty-day proof target. Dull artifact. Real sequence.

Then protect the sequence. New stakeholder asks go to the tier model. They do not jump the queue because a converter demo looked easy.

Leaders should ask: “Which meetings already run on Power BI—and which workbooks are we funding that do not run any meeting?” If the second list is long, prioritization failed before migration started.

Why lift-and-shift feels responsible

It looks fair. Nobody’s dashboard is left out. Fairness is the wrong frame for scarce rebuild capacity.

It looks measurable. Workbook percent-complete charts easily. Meeting-trust does not fit a pie chart as neatly—and still matters more.

It looks reversible. “We can always clean up later.” Later rarely gets a budget after the conversion army is gone.

Responsible is moving the decisions that move cash, margin, and service levels—then giving everyone else a governed model to answer the next question without inventing another estate.

Executive takeaway

Don’t migrate every dashboard. Migrate the ones that run the meeting.

A two-hundred-workbook lift-and-shift burns budget on the long tail while Monday still waits. Score by meeting use, rebuild the top loop with a real semantic model, archive the rest on purpose, and measure progress in meetings moved—not files converted.

Ready to sequence a migration around the decisions you actually run? Start a Power BI migration conversation. Or contact Alluvium to score the estate and pick the first meeting loop.