How to do sales funnel analysis and find where leads drop off
Use sales funnel analysis to measure stage conversion, find where leads drop off, and separate observed losses from deals that are still open.
Free app. Runs on your Mac with the Claude or Codex plan you already have.
What sales funnel analysis is
Sales funnel analysis follows a group of leads or deals through the stages that precede a sale. It shows how many reached each stage, how often those eligible to advance did so, and where the evidence stops. The starting group is an entry cohort, a set of records that entered the funnel under the same rules during a named period. A conversion funnel report follows that group through a stated date. A tally of today's pipeline answers a different question. It shows where work sits now.
Funnel drop-off analysis helps you decide which handoff or stage deserves a closer look. A lead can disappear from a later stage count because it was closed lost, because the team is still working it, or because the move was never recorded. Treat those explanations separately. Use the counts to locate the problem and read the deal history to understand it. A usable result lets a sales manager trace a finding back to the records and decide what to test in the sales process.
What it does for the business
- Focus the next sales reviewSend managers to the stage where a material group failed to advance and where records can explain what happened. The meeting can end with a specific follow-up, such as reviewing the qualification handoff, and an owner.
- Fix the records that obscure decisionsWhen missing stage events explain a suspicious drop, give sales operations specific records and fields to inspect. Better logging makes the next report easier to interpret and lets reps spend the review on deals that need attention.
- Separate follow-up from loss reviewUse open records to identify work that still has an owner and a possible next action. Route closed losses into a separate review of documented reasons, so the team can study missed sales and pursue live opportunities in parallel.
- Choose a change you can judge laterKeep the cohort definition and observation rules with the result. After changing a handoff or coaching reps on discovery, you can run the same analysis on a later cohort and see whether the pattern deserves more investigation.
Sales funnel analysis, step by step
Set the cohort and the decision
A funnel has a useful denominator when everyone agrees who entered it and when.
What to do
- Start with the decision the team can make, such as which handoff to review before changing its qualification process. Name the person who owns that decision.
- Choose the record you will count. If the report follows opportunities, carry an opportunity ID through every stage. If it follows leads that become opportunities, establish how those IDs link before joining the exports.
- Set the entry event, entry dates, pipeline and observation cutoff. Use the same inclusion rule for the entire cohort and record how you handle items that entered before the period.
- Locate the current export with stage history and any records that document outcomes. Check whether the export covers the whole group and whether definitions changed during the period.
Why
A group that entered during the period can be followed consistently. A group made up of whatever happened to be in the CRM at the cutoff blends fresh leads, old open deals and survivors from earlier periods. Use that current view for a workload question and a cohort for a progression question.
Turn stage history into usable evidence
A stage event records a move and can occur more than once for the same deal.
What to do
- Write the stage order in business terms and map the source labels onto it. Keep the original label beside the mapped value so someone can check the mapping later.
- Sort events by record ID and timestamp. Check repeated rows, missing dates, moves back to earlier stages and ties whose order you cannot resolve. Keep a record of what you set aside.
- Agree how to treat reentry and skipped stages. For stage reach, count an item once per stage. For a handoff, define which moves qualify and keep skipped transitions visible.
- Check whether the data reports first entry into a stage or only the current position. If the source only gives the current position, produce a pipeline snapshot and label it as a count of current records by stage.
Why
Many teams move deals backward when new information comes in. An automatic rule that discards every backward move can erase a real workflow. Document how you interpreted it, and keep incomplete records visible so the audience can tell which calculations still have support.
Calculate conversion and keep open work visible
Each stage conversion rate needs the eligible group and the observed advances from that group.
What to do
- Count distinct cohort records observed reaching each stage by the cutoff. Reconcile those counts to the records that entered the funnel and retain a path back to the contributing IDs.
- For each adjacent pair, count eligible records that reached the earlier stage. Use that denominator and count the members of the same group observed advancing to the next stage under your transition rule.
- Set a conversion window if the decision needs one. Show which records have had enough observation time and make any immature cohort explicit. Label the rate as observed through the cutoff or through the fixed window you used.
- Use documented outcomes to distinguish closed losses from open deals and unresolved records. Compute elapsed time only where the timestamps needed for that interval are valid.
Why
A slow stage can look like a leaky stage when the cohort has only just arrived. Decide how to handle incomplete follow-up before looking at the results. Keep the size of each group visible so a weak-looking rate on a small group gets the scrutiny it deserves.
Reconcile the report and inspect the apparent leak
A low conversion rate tells you where to look for evidence.
What to do
- Recompute the key handoff from the source and inspect examples that advanced, went backward, closed lost and stayed open. Check that each example landed in the right group.
- Reconcile entry counts and exclusions, and check the handling of records with missing or conflicting data. Look for one rep or import source that uses stage labels differently.
- Compare the count of observed losses with the rate and the maturity of the cohort. Pick the handoff whose evidence is strong enough to support a useful review.
- Write hypotheses separately from findings. If several losses contain the same documented objection, state what you observed and identify which records could support or weaken that explanation.
Why
A missed logging step can create the same shape as a selling problem. Reviewing exceptions helps you avoid coaching reps on an issue caused by the export or changing a process because recent deals had too little time to progress.
Deliver a result that can drive a decision
A funnel report earns its place when the next owner can act on its evidence.
What to do
- Put the strongest supported observation first. Name the cohort dates and cutoff nearby, and be explicit about whether the table measures progression or current-stage workload.
- Include the stage counts, the numerator and denominator, and the open and unresolved records alongside the result. Give the sales lead a way to find the underlying examples.
- Name a first action that matches the evidence. Ask sales operations to repair an export with gaps or ask a manager to inspect examples from a handoff with documented losses.
- Save the definitions and input dates with the report so the next reviewer can repeat the calculation on the same basis.
Why
Managers often forward the result without the workbook or meeting context. A self-contained explanation keeps a provisional finding tied to the period and records it actually represents.
Common mistakes
- Counting stage events as dealsA deal can enter a stage repeatedly, so use distinct record IDs when you calculate reach.
- Mixing cohortsSet one entry rule so the numerator and denominator describe the same group.
- Calling all unadvanced leads lostUse a documented loss outcome and keep open records visible.
- Timing transitions with broken datesExclude intervals whose entry or exit timestamp is missing or invalid.
- Treating the weakest rate as its own explanationRead the source records and check whether the apparent problem comes from the workflow, the data or the observation window.
What a good result looks like
- The entry event, cohort dates, stage order and observation cutoff are easy to find.
- Each stage pair has an eligible denominator and an observed numerator you can recompute.
- Observed losses, open records and unresolved outcomes remain visible.
- Exclusions, skipped stages and incomplete timestamps have documented treatment.
- The headline points to a supported finding and a named action.
And then what? Using the result
- Assign a manager to inspect records from the selected handoff and record the reasons that the evidence supports.
- Give sales operations examples of missing stage events or inconsistent labels and verify the next export against those records.
- Choose one workflow change to test where the evidence supports it and record when it starts.
- Repeat the report on a later comparable cohort and review the limits before interpreting a change.
When to run it
- Before changing qualification or routingTrace the path from entry through the next stage and read the records where work stalled. The review can show which rules or logging steps deserve attention.
- After a CRM migrationCheck whether historical events and IDs survived before comparing conversion across periods. A pipeline snapshot can still answer where open work sits if stage history is unavailable.
- When the pipeline looks full and sales are hard to explainFollow a cohort and inspect the difference between documented losses and open items. Use the distinction to plan a loss review or a follow-up session.
Questions people ask
How do you do sales funnel analysis?
Choose one entry cohort, follow distinct records through dated stage events, and calculate which eligible records advanced before the cutoff. Review the weak handoffs and document observed losses, open deals and uncertainty before recommending a change.
How do you calculate stage conversion?
Divide the number of eligible records observed advancing from the earlier stage to the next one by the number of eligible records in that earlier stage. Apply the same cohort and observation rules to both counts and publish them with the rate.
Can you measure conversion from a pipeline snapshot?
A single pipeline snapshot shows current stage counts and cannot establish historical conversion. To follow the same records across stages, obtain dated stage history or another source that records transitions and their timing.
How should you handle open deals?
Show open deals separately from observed losses and label the cutoff that determines their status. If older and newer deals had different chances to progress, make that difference visible before interpreting rates.
How do you find the biggest funnel leak?
Look for a handoff with a material number of documented losses and enough follow-up time to judge progression. Check the underlying records and logging before making it the focus of a sales process change.
What if leads skip a stage?
Record the skip and apply the stated transition rule without inventing an entry event. A skipped stage can explain gaps in the stage counts and may deserve its own review if reps use a different path in practice.
More jobs in Sales & outbound
Find where the sales funnel loses leads, on your own Mac.Free app. Runs on your Mac with the Claude or Codex plan you already have. Your files stay on your computer.
Download for Mac