Funnel dashboard: stage counts, conversion and leakage

A funnel dashboard compares stage counts, conversion, and elapsed time for a defined cohort using supplied stage-history data. Its first job is to make those figures reconcile. Keep cumulative stage entrants separate from current stage inventory, and label opportunities still open at the cutoff.

Start with the history the dashboard needs

Request opportunity ID, original source, cohort entry date, stage entry timestamps, and dated outcomes. Include the export cutoff and the stage definitions. Use the existing revenue operations metrics handoff definitions rather than introducing a second meaning for qualified opportunity.

A current CRM snapshot cannot tell you which earlier stages an opportunity visited, when it entered them, or whether it returned. A won opportunity may have skipped a stage. Obtain history or a prepared, verified entry-date summary; do not fill missing history from its current status.

Choose how repeats work before counting. This example follows first qualifying entry and the first subsequent proposal, retaining a record once at each stage. Reopened deals stay in their original cohort. A separate report can study repeated visits, but those visits should not inflate the dashboard's unique opportunity counts.

The account funnel glossary gives the broader progression concept. Here the task is building a view whose displayed counts and dates can be checked against the export.

Prepare one row per opportunity

A US services company has the following verified history summary. All opportunities qualified on October 1, 2026. The cutoff is October 15, 2026, so every record has the same fourteen days of observation. Sources are fixed at original acquisition.

Put ID in A, source in B, qualified date in C, proposal date in D, outcome in E, and outcome date in F. Store dates as spreadsheet dates, not text; a dash below represents a blank cell.

Keep a copy of the original export and state the extraction date on the dashboard. Before summarizing, verify unique IDs, valid dates, and proposals occurring after qualification. Confirm every won record on this particular ordered path has a recorded proposal. Investigate missing entries instead of silently adding them.

Calculate the three count columns

Use a compact summary with source in J2, qualified entrants in K2, proposal entrants in L2, and wins in M2. Put Organic in J2 and Paid in J3. Microsoft's COUNTIFS reference describes counting records that satisfy all supplied criteria.

=COUNTIFS($B$2:$B$9,J2,$C$2:$C$9,">0")
=COUNTIFS($B$2:$B$9,J2,$D$2:$D$9,">0")
=COUNTIFS($B$2:$B$9,J2,$E$2:$E$9,"Won")

These formulas assume the prepared table is already restricted to the cohort and cutoff. On a larger export, apply those restrictions before counting or include them explicitly as additional criteria.

The rate formulas are =IF(K2=0,"",L2/K2) and =IF(L2=0,"",M2/L2). For the total row, divide total proposals by total qualified entrants and total wins by total proposals. Display percentages to one decimal place and retain full precision underneath.

Overall cohort conversion is 3 / 8 = 37.5%. It answers a different question from 3 / 5 = 60.0%, which measures wins among proposal entrants. Keep denominator labels visible beside both.

Add elapsed time without hiding the unfinished work

In G2, calculate qualification-to-proposal days only for records with a recorded proposal:

=IF(D2="","",D2-C2)

Fill down. The completed transition times are 2, 5, and 6 days for Organic, and 3 and 7 days for Paid. Microsoft documents AVERAGEIFS as an arithmetic mean restricted by criteria. For source J2:

=AVERAGEIFS($G$2:$G$9,$B$2:$B$9,J2,$D$2:$D$9,">0")

The total is 23 / 5, not the average of the displayed source averages. This timing measure describes completers only. It excludes O03, P02, and P04 because none has a proposal date. Keep their non-progression count visible so a quick completed transition does not imply every opportunity moved quickly.

Reconcile progression and current inventory separately

Three qualified opportunities never reached a proposal by the cutoff: one lost and two remain open. Of the five proposal entrants, three won, one lost, and one remains open. Therefore:

Qualified entrants = No-proposal losses + No-proposal open + Proposal entrants

8 = 1 + 2 + 5

Proposal entrants = Won + Proposal losses + Proposal open

5 = 3 + 1 + 1

The current inventory tells a different story: two open in qualification, one open in proposal, three won, and two lost. Those mutually exclusive positions sum to eight. Cumulative proposal entrants remain five even though only one is currently open in proposal.

Place both checks beside the dashboard. Use counts and reconciliation status before adding color or chart shapes. The conversion funnel optimization workflow then helps prioritize an investigation; dashboard construction itself does not explain the source differences.

Where it goes wrong

• History inferred from a snapshot. Current stage alone cannot establish first-entry dates or prior visits.
• Cumulative entrants added together. The same opportunity appears at several stages, so summing stage columns counts it repeatedly.
• Open records labeled lost. Non-progression at a cutoff includes unfinished business.
• Timing without an eligible count. A completed-transition average hides opportunities that have not completed the transition.
• Uncontrolled dashboard filters. Apply cohort, source, and cutoff rules to every card. A numerator and denominator from different selections produce a misleading rate.

Covirage is not affiliated with Microsoft.

Do this on your own file in Covirage

Bring a CRM stage-history export or a verified summary of stage-entry dates, sources, and outcomes. Covirage maps the columns, and its tools compute and check stage counts, conversion, and elapsed time; the external AI model explains the result with every figure traced to its rows. Create a free Covirage account, which takes about a minute, then set up your business and bring a sample of your own data.

See Covirage for sales insights for the supporting workflow.