Sales Funnel & Conversion Analyser

What this tool does

CM8-280 measures conversion. You record how many enquiries entered each stage of your sales process in each period, for each channel; it works out the conversion between consecutive stages, the end-to-end win rate, which step loses the most, which channel actually converts, how long deals sit at each stage, and how many leads a win target requires. Everything runs inside this single file — no account, no upload, no network request of any kind — so your conversion rates never leave the computer you are using.

Funnel or pipeline?

Two different questions needing two different records. A pipeline is a list of live opportunities: who, how much, what stage, when it might close. It answers what is in play right now, and the companion free tool, the Sales Pipeline & Forecast Tracker (CM8-60), does exactly that — one row per opportunity, weighted by probability.

A funnel counts, in aggregate, how many records moved through each stage in a period. It answers where do we lose people, and how many do we need at the top. A pipeline cannot answer that, because it shows only the survivors: the enquiries that quietly failed to be qualified last March are missing from it, and they are what the funnel is about. Most teams need both — the pipeline weekly to run deals, the funnel monthly to run the process that produces them.

The row model

Each row is a count: this many records entered this stage, in this period, from this segment. Six stages, two channels and one month is twelve rows. That is what makes conversion measured rather than guessed at — both sides of the division are written down and dated.

Use the month end as the period date, the same way every time, and spell segment names identically — the comparison groups on that text. Value is optional but worth the effort: it separates a big stage from a valuable one, and recording it against the lost stage prices what walked away.

Counting stages honestly

Read this twice: one measurement error ruins everything downstream, and it is by far the most common one. Count what entered a stage during the period. Do not count what happens to be sitting in it today. Those sound similar and are completely different numbers. Export "everything currently at quoted" on the last day of the month and you get a snapshot: things quoted yesterday, things quoted four months ago that nobody has closed, and none of what was quoted and then won or lost during the month, because those have moved on. Divide that by a lead count and the answer means nothing — it is a ratio of a stock to a flow, and it flatters the top of the funnel while hiding the stage that is silted up. What you want at every stage is the flow: how many crossed into it between the first and last day of the period. Most systems report that as "stage changed to X between these dates". If yours cannot, count entries by hand — three honest months beat three years of snapshots.

Two consequences. The same record legitimately appears at several stages in one period — a deal that arrived, was qualified and was quoted in March is counted once at each — which is why the stacked segment chart is a shape comparison, not a total. And consecutive stages in one period are not a strict cohort: some deals quoted in March were leads in January.

The formulas

Stage conversion is the local question: of everything that reached "qualified", what share went on to be quoted? Cumulative conversion is the global one: of everything that arrived as a lead, what share ever reached this stage? Both sit side by side in the funnel table because they are routinely confused — a stage can look healthy locally, converting 80% of the handful that reach it, while cumulatively almost nothing arrives there at all. Lost is not a step in the chain, so it carries no conversion from the stage above and is shown against leads only; its value is deliberately kept.

How many leads you need

This is the most useful output here, and it is one line of arithmetic. If you need ten wins next quarter and your end-to-end win rate is 4%, you need 250 leads. Not "more focus", not "a push on closing" — 250 leads, or a materially better conversion rate. Those are the only two levers, and no amount of pipeline discipline changes the arithmetic.

Read the win rate off the tile, decide how many wins you need, divide, then look honestly at the answer. If the required volume is three times anything you have generated, the plan is not a plan. Use a win rate drawn from several periods — the tile pools every period in the current filter for that reason.

The biggest drop-off

The tile names the consecutive step with the worst conversion in the latest period. It tells you where the work is. The instinct when wins are short is to push harder at the top — more campaigns, more calls, more leads. If the weakest step is near the bottom, that can work. If it is at the top, volume poured into a broken stage is waste at full price: every extra lead pays the same acquisition cost and dies in the same place. Improving a weak middle step is usually cheaper than doubling the top — moving one step from 45% to 60% lifts everything below it by a third, with no extra lead spend.

Comparing segments

The segment table pools every period in the current filter and gives each channel its leads, quoted, wins, win rate, value won and cycle length, with a verdict: Best converting, Volume but poor conversion or Low volume. The channel producing the most leads is very often not the channel producing the most wins. High-volume inbound is cheap per lead and expensive per win; a slower channel with a tenth of the volume can out-convert it several times over, and until the funnel is split by segment the average hides both. Take the verdicts as prompts — "best converting" on eleven leads is noise.

Days in stage

Average days answers what conversion cannot: where do deals sit? Add the averages for stages 1 to 5 and you have an approximate sales cycle, which tells you how far ahead leads must be generated. With a long cycle, the leads you need are generated a cycle before the target date.

Watch for a stage where the days climb while conversion out of it falls. That is the signature of deals quietly dying: nobody has said no, so nothing is marked lost, and the stage silts up with records that will never close. A deal four months "in negotiation" is usually a loss nobody has written down.

Drop-off reasons

Every row carries the main reason records at that stage did not progress. Use a short phrase and reuse it exactly — "price above budget", "no budget confirmed", "not the decision maker". Ten consistent phrases beat a hundred unique sentences, because the reasons turn a percentage into an action: "qualified to quoted is 46%" is a fact; "qualified to quoted is 46%, and the reason is nearly always that no budget was confirmed" tells you to change the qualification questions.

Review cadence

Monthly suits most businesses. Quarterly is better if you close fewer than about ten deals a period, because below that a win rate swings wildly on one deal.

1. Close the period and enter every stage and segment in one sitting, from the same source. Mixing sources between stages is how impossible conversion rates appear.
2. Read the funnel and the biggest drop-off. Has the weakest step moved? If you acted on it last month, did the number respond?
3. Check the win-rate trend against the target line — the direction, not the last point.
4. Do the required-leads arithmetic against the next target, and ask whether the effort is going into the channel that converts.

FAQ

Can conversion be over 100%? Yes, and it is usually a warning: more records entered a stage than the one above it in the same period. Normal with a long cycle, where this month's quotes came from last quarter's leads; a mistake where deals skip a stage or are entered straight in at "quoted". The tool cannot tell those apart from one row, so it does not block it. Investigate before you quote it.

Should a deal be counted at every stage it passes? Yes — a deal moving through three stages in a month is counted three times, once at each.

Must won and lost add up to quoted? No: some quotes stay open, and some losses come from deals quoted long before. If won plus lost is persistently far below quoted, you have deals that are neither closed nor real.

Why record value on the lost stage? It prices the problem. Losing forty small enquiries and losing four large ones need entirely different responses.

Saving your work

Stage counts, settings and the report header are written to this browser's local storage as you type. That storage belongs to one browser on one computer: another browser, a private window, a second machine or a clean-up tool that clears site data will not have it.

Treat Export .json as the real save — one file containing everything, which Import .json restores anywhere. Export CSV gives you the counts for spreadsheet work. Reset asks twice, then erases everything. There is no undo.

Accuracy & disclaimer

The arithmetic is division, and the tool does it faithfully. Everything that matters sits underneath it: whether you counted entries into a stage or a snapshot of it, whether every segment was measured the same way, and whether the sample is large enough to divide at all. A conversion rate built on a handful of deals moves a long way on one result — a month's funnel is a hint, a trend across several is evidence.

Required-lead arithmetic assumes the future converts like the past: the most useful assumption here and the most fragile. This is a measurement and prioritisation aid, not a forecast, not a commitment to anybody, and not an accounting record.