1. Inventory the files and the people behind them
Identify the reports people actually use, their owners and the decisions each one supports. Record source files, update frequency, manual steps and who receives the result. Several similarly named workbooks may contain different definitions or reporting periods.
Choose one bounded report for the first migration. An inventory helps distinguish source data from presentation files and avoids rebuilding obsolete sheets simply because they exist. Keep a copy of the agreed reference version for reconciliation.
2. Capture the logic before changing the format
Review formulas, pivot tables, lookup rules, macros and manual overrides with the person maintaining the workbook. Document measures, date logic, exclusions and exception handling. A cell containing a typed adjustment may represent a real business rule that is missing from the source system.
Decide which logic belongs in data preparation, which belongs in the reporting model and which needs a business decision. Do not replace an unexplained calculation with a new one and assume that a visually similar result is equivalent.
3. Prepare tables with a clear reporting grain
Create consistently structured tables with clear field names and a known meaning for each row. Review merged cells, subtotal rows, mixed date formats and values stored as text. Keep units and identifiers consistent across files. If the workflow receives a new file every period, define a predictable naming and update process.
Tableau supports connections to Excel workbooks. Its documentation explains connection behavior and the options for combining tables. Test the structure against the installed version and your actual files rather than assume every presentation sheet is analysis-ready.
4. Check relationships and totals deliberately
Before combining tables, identify keys and check whether they are unique at the expected grain. An order header and its order lines do not represent the same level of detail. A fixed join that duplicates header values can change an aggregate even if the rows look plausible.
Tableau documents relationships and joins as distinct modeling approaches. Choose the approach around the data and analytical requirement, then test it. Compare record counts, distinct keys and selected totals before and after the combination. Include unmatched keys and records with missing values in the review.
5. Validate the report with business users
Reconcile the agreed period, population and filters against the reference spreadsheet. Investigate differences instead of forcing the new result to match an unexplained total. Some differences may reveal a previous error, while others indicate that a required rule was missed.
Ask users to carry out realistic tasks: find a segment, change a reporting period and explain a variance. Check empty results, unusual dates, access roles and performance with representative data volume. Tableau’s workbook performance checklist provides additional technical review points.
6. Assign ownership before retiring the spreadsheet
A published dashboard still needs data updates, monitoring and a change process. Agree who owns the source, who checks refreshes and where users report a problem. Document file locations, connection ownership and the response to a late or missing input.
Run the old and new reporting processes in parallel for an agreed validation period. Retire the previous report only after its owners approve the new definitions and results. Keep reference documentation so later changes to the source or workbook can be checked against the same business purpose.
What a first migration should deliver
The first release should leave more than a workbook: an agreed source inventory, documented measures, reconciliation evidence and operating responsibilities. Use what the team learns to decide which report to migrate next. This keeps the sequence connected to business value and the readiness of the data.