Quarter-end totals disagree when an extract is incomplete, populations differ, or value rules differ.
Key takeaways
- Inspect the batch response body for individual errors, as the request returns 200 OK while hiding specific failures Microsoft.
- Respect the 5,000 record page size limit for standard and elastic tables when extracting data Microsoft.
- Separate opportunity creation dates from performance dates to align population windows Close.
- Verify which of the Won, Committed, and Best-case categories a total includes before comparing values Freshsales.
- Distinguish total deal value from internal revenue splits allocated among team members Membrain.
Sales report mismatch: the documented causes to check
Extraction errors are one possible cause of a mismatch. A standard HTTP 200 OK status does not guarantee that every individual request within the batch succeeded. According to Microsoft, the batch request returns 200 OK, and individual response errors are included in the batch response body. This means a report built on an automated extraction script may silently drop records if the script checks only the top-level status code and ignores the nested error details.
When reconciling quarter-end totals, treat the batch response body as the source of truth for extraction integrity. If the script logs a 200 OK but the report totals are lower than expected, check whether individual requests failed within that batch. The error details are present in the response body, not in the header status.
Do not assume that a successful batch status implies a complete dataset. The documented behavior shows that individual errors are embedded within the successful batch response. Your reconciliation process must parse the response body to identify which specific requests failed. If your extraction tool does not surface these individual errors, you cannot verify whether the extracted population is complete.
Extraction mistakes: incomplete pages and hidden request errors
When extracting sales data from Microsoft Dataverse with the Web API, the total count can be wrong if the retrieval process stops early or misinterprets paging rules. The documentation for Page results using OData from Microsoft Dataverse Web API - Power Apps | Microsoft Learn specifies that for both standard and elastic tables, you can specify a max page size up to 5,000 Microsoft. If your script requests a page size larger than 5,000, the service ignores that request, potentially leaving you with a partial dataset if you do not account for the remaining records.
A common error occurs when developers attempt to use standard OData paging parameters that the platform does not support. Dataverse does not support the $skip query option, which means you cannot use the combination of $top and $skip for paging Microsoft. Instead, the API returns a @odata.nextLink URL in the response body to guide you to the next set of results.
You must handle this next link precisely to ensure complete extraction. The guidance states not to change the @odata.nextLink URL value or append any query options to it Microsoft. Modifying this URL, such as by adding filters or changing the page size parameter after the initial request, goes against that guidance. Verify that your integration script uses the next link exactly as provided by the service.
To reconcile conflicting totals, inspect the raw API responses for your quarter-end extraction. Confirm that your script respects the 5,000-record page limit and does not attempt to use $skip. Ensure that you follow the @odata.nextLink without alteration until the response no longer includes a next link. If your extraction tool logs a successful batch response but the record count is lower than expected, check whether the tool incorrectly assumed that a 200 status code implies all individual requests within the batch succeeded. The documented paging constraints apply specifically to the Dataverse Web API; other systems may handle pagination differently, so verify the behavior against the specific tool's documentation.
Population mistakes: date ranges, stage columns and categories
Differing time windows or stage definitions can cause mismatches. Before aligning two conflicting quarter-end figures, verify that both reports use identical date ranges for creation and performance, and that they define the same opportunity populations.
Creation and performance dates
The Opportunity Funnels Report in Close requires you to select a date range for when the Opportunities you want to view were created, along with a separate date range to view their performance. These two windows serve distinct purposes. One filters the set of deals based on their origin date, while the other filters based on when activity or outcomes occurred. If one report filters by creation date and the other by performance date, the populations may differ, even if the stage filters appear identical. Check which date field each report uses for its primary filter. A report showing deals created in Q3 but performing in Q4 will exclude deals created in Q4, regardless of their current status.
Win rate denominator
Win rate calculations can also diverge if the denominator differs. According to the Opportunity Funnels Report in Close, your Win Rate is calculated by dividing the number of won Opportunities by the total Opportunities that were won or lost. This specific denominator excludes open opportunities from the calculation. If a competing report includes open opportunities in its denominator or uses a different definition of "lost," the win rate percentage may differ. Confirm that both reports use the same numerator and denominator definitions before comparing percentages. The source explicitly ties the calculation to won and lost opportunities only.
Stage column population differences
Stage columns in funnel reports may include more opportunities than you expect. The Opportunity Funnels Report in Close notes that if you look at one specific column, it not only includes opportunities which are currently in this status, but also those which moved forward. This means a column labeled "Negotiation" includes deals currently in Negotiation as well as deals that have already advanced to a later stage. If one report treats a stage column as a snapshot of current status and another treats it as opportunities currently in the status or that moved forward within the report's selected population, the totals can disagree. Verify whether your reports use current-status filtering or cumulative-stage filtering. The distinction between "currently in" and "moved forward" is critical for accurate population matching.
Value mistakes: shared credit and recognized revenue
The disputed total may count different things depending on how the tool assigns value. Freshsales shows the total deal view breakdown in three categories—Won, Committed, and Best-case Freshsales. The ability to commit to a deal is governed by Roles and Permissions. Record which categories the disputed total includes before comparing reports.
Membrain allows sales managers to allocate portions of a deal’s total value among internal team members Membrain. The deal owner is included automatically, but everyone contributing to the win can be acknowledged and measured. Revenue Splits are fully integrated into analytics and goal tracking. Identify whether the comparison concerns deal value or contributor attribution.
Stripe illustrates revenue spread across service periods. The 31 USD is recognized over 17 days in January and 14 days in February Stripe. If you inspected the account balances at the end of January, 17 USD of revenue was recognized, and 14 USD of revenue remains deferred to be recognized in February. On January 1, a customer starts a 3-month subscription for 90 USD, which generates and finalizes an invoice with a service period from January 1 to March 31. In this case, revenue recognition continues to amortize the 90 USD over the original 3-month service period. Retain the period behind each recognized-revenue figure.
To resolve these value mismatches, check the Pipeline, Best Case, Commit: Forecast Categories Explained guide to understand how different tools categorize deal stages. Additionally, review the Weighted deal amount: 5 fields to check HubSpot, Pipedrive article to ensure you are comparing consistent valuation metrics across your systems.
Illustrative example: same count, different total
These figures are illustrative only. Report A lists 200 deals totaling 90. Report B lists 200 deals totaling 75. The deal counts match, so a missing-record gap is less likely. Compare the categories first. If Report A includes Best-case and Report B includes only Committed and Won, the 15 difference may be a category scope issue. Next, check whether either report counts full deal value or revenue splits. Confirm that both reports use the same date field and window, then write down which difference you found. If the counts differ instead, check the extraction first. Keep this note with both reports so the difference is traceable.
Evidence to retain with the quarter-end reconciliation
Today, select two conflicting quarter-end reports and annotate their extraction status, date filters, opportunity populations, included categories, attribution basis and service periods using the checks applicable to their tools. Record whether your extraction inspected that response body before accepting the totals. If the body contained errors, check which requests failed and whether they affected the required records before calling the dataset incomplete.
Next, annotate the date filters. Note whether each report uses opportunity creation dates or performance dates, and whether the range includes opportunities that moved forward in a stage column during the period. These population definitions vary by tool, so a report built on creation dates may differ from one built on performance dates, and different date filters can select the same deals.
Then record the included categories. For tools that break deal value into Won, Committed and Best-case, note which categories the disputed total includes. A Best-case total can be larger only when other amounts match and the added Best-case amount is positive, which still does not rule out a data entry error.
Keep these annotations alongside the two reports so the reconciliation shows exactly which extraction, population, category or attribution difference produced the mismatch.
FAQ: investigating differences between sales reports
What causes a sales report mismatch at quarter end?
Incomplete extracts, different populations, or different value rules can cause the mismatch.
Does the Dataverse paging guidance apply to both standard and elastic tables?
Yes, the service ignores max page size requests larger than 5,000 for both standard and elastic tables. For both standard and elastic tables, you can specify a max page size up to 5,000, as stated in Microsoft. This limit defines the boundary for valid page size specifications in that query context.
How does Close calculate opportunity win rate?
Win Rate is calculated by dividing the number of won Opportunities by the total Opportunities that were won or lost. The Opportunity Funnels Report uses this specific denominator for its calculations from Close. This definition ties the metric directly to the final status of opportunities within that report.
Who can commit a deal in Freshsales?
Access to this action depends on the specific configuration of user roles within the system, according to Freshsales. This permission boundary determines which users can mark a deal as committed.
How does Stripe recognize revenue for a 3-month subscription?
Revenue recognition continues to amortize the 90 USD over the original 3-month service period, as Stripe describes.