Digital Squad

Strategy · 19 November 2026

Why Last Month’s Revenue Keeps Changing: Data Analytics in Hong Kong

Explain why past marketing revenue changes in your CRM. Preserve reporting snapshots, trace deal edits and give Hong Kong teams a clear process for corrections.

By Digital Squad

A hand holding a pencil beside two stacks of papers and a closed folder.

Key Insights

  • A report built from today's CRM values can change last month's totals whenever a deal is edited, reopened or moved to another period.
  • Preserve an auditable reporting snapshot with its extraction time, definitions and deal identifiers so the original decision can be reconstructed.
  • Give corrections an owner and show their effect separately from new commercial performance.

The leadership team agreed October's marketing-attributed revenue at the monthly review. Two weeks later, the same dashboard shows a different total. No one has changed the chart, yet a campaign that appeared to have reached its target now looks short.

The explanation may sit in ordinary CRM activity. A salesperson changes a close date after contract discussions, an operations colleague corrects an amount, or an integration updates the market assigned to a deal. A report that reads current values can apply those edits to a period your team thought it had already reviewed.

For a Hong Kong business reporting local and regional pipeline, this can unsettle budget decisions. The first task is to establish what the report is intended to show: the latest commercial position or the position your team knew at a particular reporting cut-off. Both can be legitimate views, provided their definitions are explicit.

Decide which version of the past the report should show

A current-state report answers a question such as, “What do we now believe about deals with an October close date?” If someone corrects a deal, the report should reflect that correction. It is a living view of the commercial records.

A reporting snapshot answers a different question: “What records and values supported the October review when we prepared it?” Preserving that view lets you reconstruct the decision made at the time. It doesn't require salespeople to stop updating the CRM.

Consider an illustrative deal with a value of HK$240,000. It appears in the October review, then its close date moves to November. A current-state October report may lose HK$240,000 while the saved October snapshot retains it. The difference is explainable if the two reports state their cut-offs and treatment of date changes. Without those labels, it can look like a calculation error or a missing sale.

Agree whether leaders need both views. A budget review may need the original decision evidence alongside the latest corrected position. Give each a clear name and date so a refreshed report doesn't silently replace the record of an earlier meeting.

Trace the fields that can move the total

Start with a handful of changed deals and compare their stable record identifiers. Examine the amount, currency, stage and close date, then the fields that determine inclusion in the marketing report. An attribution model or market filter can change a total even when the deal's commercial value stays the same.

HubSpot, for example, lets users inspect a record's property history, including its changed value, date and source. That can help establish whether a user, workflow or integration changed a relevant field. Use the history to investigate the actual records behind the difference, rather than guessing from the chart alone.

Keep time meanings separate. The time a record was edited may differ from the commercial close date entered by the salesperson. If an update made in November assigns a deal to October, your report needs a rule for that backdated change. Applying a local reporting cut-off also requires consistent time-zone handling across exports and dashboards.

For a regional pipeline, preserve the distinction between Hong Kong demand and the market served by the deal. If those labels are overwritten during account cleanup, the apparent market performance may shift without any new customer decision.

Capture enough evidence to reproduce the review

Save the underlying rows used for the agreed reporting period, with the fields that determine inclusion and value. Include record identifiers, the extraction timestamp and the version of the reporting rules. A picture of the total preserves what the meeting saw, but it can't explain which deals produced that number.

Document how the calculation handles currency conversion and attribution. If marketing receives a share of deal value, preserve the credit applied at the cut-off as well as the total deal amount. A later change to that share should be distinguishable from a salesperson correcting the amount.

Set access according to the information captured. A reproducible marketing report rarely needs every contact field in the CRM. Retain the minimum evidence that supports reconciliation, keep it in an appropriately controlled location and decide who can replace or approve a snapshot.

Property history can assist investigation, but it has product-specific limits. HubSpot's history export documentation states that company, deal, ticket and custom-object properties retain up to 20 revisions per property, while contact properties retain up to 45. It also says change dates in the export use UTC. Those limits are a reason to verify your own platform's retention and export behaviour before relying on it as the permanent reporting record.

Make corrections visible to the people using the numbers

Choose an owner who can coordinate sales operations, marketing and the dashboard team. Their role is to establish why a figure changed and approve how the reporting view represents it. The salesperson should still correct a wrong deal; the reporting team needs a controlled way to explain the resulting difference.

For each material correction, retain the record identifier, previous reporting treatment and revised treatment. Explain the reason in business language. “Close date corrected following contract confirmation” gives the review a basis for judgement; “data updated” leaves the same question open.

Present the movement from the original snapshot to the corrected view. Separate late additions, value changes and reclassification so leaders can see which movement reflects new information. Avoid silently labelling the entire difference as campaign improvement or decline.

A correction policy also needs a threshold and a timetable. Decide which changes warrant a revised management report and which can be explained at the next review. The right threshold depends on the decisions the report supports. Agree it before a disputed number arrives, so the team doesn't invent a different rule for each campaign.

Test the report with a reopened deal

Walk through a controlled example in a test environment or an approved reporting copy. Move a deal out of the closed stage, change its close date and correct its amount. Observe what happens in the current report and whether the saved snapshot remains reproducible.

Check the reverse journey too: an incorrectly excluded deal is restored. The reconciliation should show why the total rose without presenting the correction as a newly won opportunity. Ask the people who use the report to interpret both examples; if they reach conflicting conclusions, improve the labels and definitions before the next review.

This is management reporting for marketing decisions. The organisation's finance team owns accounting treatment and revenue recognition. Keep marketing-attributed deal value labelled according to its definition rather than assuming it is interchangeable with recognised revenue.

Make Data Analytics in Hong Kong Reproduce the Decision

Bring the original report, its latest version and several deals that changed to the next measurement discussion. Agree the reporting cut-off, the fields to preserve and who signs off material corrections.

Ask Digital Squad's Hong Kong marketing analytics team to trace that comparison. The objective is a report your team can reconcile: what it knew at the review, what changed afterwards and which commercial decision the revised evidence supports.