Is Your Excel Web Workflow Affected? A Practical Migration Guide

A third-party archive of Microsoft 365 Message Center item MC1473149 reports that, beginning October 16, 2026, Excel for the web will no longer load Power Query data into its Python runtime. The original Microsoft-hosted notice was not publicly available for independent verification, so treat the date and scope as reported.

The change is narrow. Review a workbook if it is operated in a browser, runs Python in Excel, and passes Power Query output to that Python operation.

What is retiring on October 16, 2026?

The reported retirement concerns one data path: Power Query connects to and transforms data, and Python in Excel then consumes that prepared data while the workbook runs in Excel for the web. Reports identify Excel for Windows desktop as the continuing environment for this combination.

This is not evidence that Microsoft is retiring Power Query, Python in Excel, or Excel for the web as a whole. Microsoft describes Power Query as a system for connecting to, transforming, combining, loading, and refreshing data, available in Excel for Windows, Mac, and the web. Its recorded transformation steps can also be rerun during refresh, making those steps useful migration documentation.

Scenario	Status	Reason
Power Query output passed to Python in Excel for the web	Affected	This is the integration identified in the reproduced notice.
The same reported workflow in Excel for Windows desktop	Reported to remain available	The reports identify Windows desktop as the continuing environment.
Web Power Query use without Python	Outside confirmed scope	The notice does not describe a general Power Query withdrawal.
Python using data not produced by Power Query	Outside confirmed scope	Check its platform and licensing support separately.
Excel for Mac as a replacement	Requires verification	Power Query availability on Mac does not establish support for this specific integration.
Exact behavior after retirement	Not documented	The reports do not establish a specific error or calculation result.

How to tell whether a workbook is affected

Apply this decision tree to each workbook:

1. Is the workbook used in Excel for the web? - If no, it is outside the browser workflow described in the reports. - If yes, continue.
2. Does it run Python in Excel? - If no, do not classify it as affected by this specific change. - If yes, continue.
3. Does the Python operation consume data produced by Power Query? - If yes, classify the workbook as potentially affected. - If no, it is outside the retirement’s stated scope; verify its current support separately if needed.

For example, a browser workbook that refreshes a sales table through Power Query and passes that table to Python for statistical analysis or data-quality scoring is potentially affected.

Do not include every workbook that merely contains one of the technologies. A Power Query result used only in worksheet tables or formulas is outside the reported scope. So is Python analyzing manually entered worksheet data rather than Power Query output. These distinctions prevent over-classification; they are not guarantees about every platform, license, or configuration.

The supplied reports do not document an automatic Microsoft discovery tool. Teams may therefore need to inventory known owners, shared locations, process documentation, recurring reports, publication calendars, and support records manually.

Prioritize workbooks that:

• must be operated in a browser;
• are shared by multiple contributors;
• produce recurring management, research, regulatory, or public reports;
• feed other workbooks, databases, dashboards, articles, or datasets;
• lack a clearly identified owner; or
• must refresh near a fixed publication deadline.

A shared workbook supporting a weekly production process should generally take priority over an isolated experiment.

Use this workbook dependency register

A dependency register turns the audit into an accountable migration queue. Use one row per workbook, or separate rows for materially different workflows within the same file.

Field	What to record	Why it matters
Workbook name	Recognizable file title	Identifies the asset
Workbook location or URL	SharePoint, OneDrive, or other location	Helps reviewers find the correct copy
Owner	Responsible person or team	Establishes accountability
Business purpose	Decision, report, analysis, or publication	Supports prioritization
Source system	Database, API, file, service, or manual source	Reveals access dependencies
Power Query query names	Relevant query names	Identifies transformations to inspect
Python dependency	Formula or analysis consuming query data	Confirms whether the reported change applies
Web or desktop execution	Production environment	Separates browser and desktop dependencies
Refresh cadence	On demand, daily, weekly, monthly, or event-driven	Helps schedule migration and testing
Downstream consumers	People, systems, reports, or datasets	Shows disruption risk
Published output	XLSX, CSV, TSV, dashboard, article, or public table	Defines the required deliverable
Data sensitivity	Public, internal, confidential, personal, or restricted	Controls handling and publication
Windows desktop readiness	Device, account, access, license, and version status	Identifies migration blockers
Validation owner	Person approving the result	Prevents self-certified changes
Migration status	Not reviewed, affected, testing, validated, or complete	Tracks progress
Target completion date	Internal deadline before retirement	Leaves time for corrections

Where reproducibility matters, also record source provenance, transformation ownership, the last successful refresh, validation date, and publication date.

“No administrative action required” does not mean “no work required.” A reproduction of MC1473149 reports that no tenant configuration change is required while also recommending that administrators identify affected users and workbooks and move dependent workflows to desktop Excel.

Once the inventory is stable, update internal training, runbooks, onboarding instructions, and support scripts so they no longer direct users to create the affected browser dependency.

Move the affected workflow to Excel for Windows

The reported migration path is to move Power Query-to-Python processing from Excel for the web to Excel for Windows desktop. Do not assume that every workbook will transfer without changes.

Before migration, verify that each intended operator has:

• an eligible Windows desktop environment;
• access to the workbook and its source systems;
• an appropriate Microsoft account and license;
• an Excel version that supports the required workflow;
• permission to use the necessary connections and connected experiences; and
• a workable process for recurring refreshes.

The available reports do not define the exact qualifying editions, builds, update channels, account types, or licenses. Test the intended production setup directly.

Preserve the workflow sequence:

1. Connect to the original source.
2. Apply the recorded Power Query transformations.
3. Load the prepared data for Python analysis.
4. Run the required Python calculations.
5. Produce a final tabular output for review or distribution.

Keep an unchanged copy of the original workbook. Record changes to query steps, table or range references, formulas, authentication, source paths, data types, Python calculations, and output locations. Do not overwrite the browser production file until the replacement has passed validation.

The supplied reports document neither a browser-based replacement nor a Microsoft migration utility. Set an internal completion date before the reported deadline, leaving time for access requests, comparison testing, corrections, documentation, and user communication.

Validate the migrated output before it is used or published

Whenever possible, run the original and migrated workflows against the same source snapshot.

Compare:

• Row count: Are records missing or duplicated?
• Column count and names: Has the schema changed?
• Data types: Did numbers, text, dates, or Boolean values change type?
• Null counts: Are blanks, nulls, and empty strings handled consistently?
• Date parsing: Are dates interpreted with the intended locale and time zone?
• Duplicate keys: Did joins or appends multiply records?
• Leading zeros: Are identifiers such as 00127 still preserved?
• Category values: Did spelling, capitalization, or grouping change?
• Python-calculated outputs: Do aggregates, classifications, model results, and derived fields follow the accepted logic?

Document and resolve discrepancies before replacing the production or publication process. A workbook opening successfully does not establish that its results are correct.

For recurring work, define an output contract covering expected column names, data types, keys, null-handling rules, refresh ownership, calculation assumptions, and permitted variation in row counts or categories. This gives reviewers a stable acceptance standard.

Finally, separate preparation from distribution. Run Power Query and Python in the verified environment, approve the resulting table, and create a dedicated publication copy. Keep private queries, credentials, source connections, formulas, and working sheets in the preparation environment.

Publish the finished table without exposing the workbook

Moving preparation to desktop Excel does not require every reader to use desktop software.

TablePage publishes finished spreadsheets and structured datasets as public, interactive pages. It is an optional publication layer, not a replacement for Power Query, a Python runtime, a private workbook repository, or a scheduling system.

Create a deliberately limited publication file containing only the rows and columns intended for release. Where appropriate, include:

• the source organization and source URL;
• a concise transformation description;
• the refresh and publication dates;
• definitions for fields and category values;
• units, geographic coverage, and time period; and
• a short data dictionary for unfamiliar columns.

TablePage output is public. Never upload credentials, private source records, confidential business information, personal data, access tokens, connection details, or other restricted material.

Audit the publication file according to its format. For XLSX, inspect every included sheet and the values intended for release. For CSV or TSV, inspect the exported rows, columns, headers, encoding, and values. Do not rely solely on what was visible in the preparation workbook.

The practical handoff is: validate the desktop-generated output, create a sanitized publication file, add a concise data dictionary, publish the public copy, and retain the private preparation workbook separately.

Do not confuse the retirement with a #BLOCKED! error

The reported October 16 change is future-dated. A #BLOCKED! error appearing before then is not evidence that the retirement has already affected a workbook.

Microsoft documents several other causes, including unsupported platforms or versions, missing licensing, disabled connected experiences or Python functions, privacy restrictions, usage quotas, authentication failures, and some calculation-mode conditions. The error generally means Excel cannot access a required resource.

At a high level, Microsoft’s #BLOCKED! troubleshooting guidance includes confirming platform and license eligibility, checking sign-in and privacy settings, reviewing connected-experience and administrator controls, and resetting the Python runtime when appropriate. Apply only the steps relevant to the displayed error instead of broadly weakening security settings.

The retirement reports do not document the exact warning, error, stale-value behavior, or calculation result that affected browser workbooks will display after the deadline. Diagnose current failures using the actual workbook, platform, account, settings, and error text.

What to do before the reported deadline

Inventory browser workbooks that pass Power Query output into Python, assign owners, verify an eligible Excel for Windows environment, and migrate the affected processing. Compare the old and new outputs before production use. If readers need public access, publish only a reviewed, sanitized final table while keeping the preparation workbook and restricted data private.