What counts as reaching a funnel stage?
Yes, they counted. Somebody counts as having reached a stage when they carry a date for it, or a date for any later stage, so a person who went from a new lead straight to an opportunity is counted at every rung in between. This article gives the exact rule, the denominators each figure rests on, and what happens when your tool records no stage dates at all. For the companion rule about people who stop moving, read What counts as stuck in a stage?
Before you start
Open Analysis, choose your tool and open the Funnel tab before you read on. The panel headed How we worked this out at the foot of that tab prints the same rules against your own numbers.
The rule, in one sentence
A person has reached a stage when they carry a date for that stage, or a date for any stage further down the ladder, or they sit in a later stage today.
The reasoning is that someone who became a customer did pass the point where a lead becomes qualified, whether or not anybody wrote it down. Counting only the people with a date for each rung would make every stage look emptier the further down you read, and the funnel would narrow because of your record-keeping rather than because of your buyers.
What MarqIQ reads to work out the furthest rung
Two things: every stage-entry date on the record, and the stage the person sits in today. The furthest of those is where they are counted to. A person sitting in Customer today with no dates at all is therefore counted at every stage, and contributes to none of the timing figures.
Skips are counted, and reported on their own
A skip is a person who got past a stage with no date for that stage. They still count as having reached it, and the skip is tallied beside the step so you can see how often it happens.
MarqIQ raises a skip as something to act on when two conditions hold together: at least 30 people got past that stage, and 40% or more of them never passed through it. The denominator is the people who went on to the next stage, not everyone who reached it. The sentence that appears reads that a stage most buyers skip is not describing how people really buy from you.
Only one skip is raised, the worst by that share. The first stage and the last stage are never raised: nobody skips the rung they start on, and there is nothing past the end of the ladder.
Take care
A high skip rate does not mean the figures above it are wrong. It means the stage is not being stamped. Both readings are true at once: the conversion through that step is real, and the timing through it is measured on a smaller group than the conversion is.
What each figure on the tab is measured on
| Figure | What it counts | Denominator | When it appears |
|---|---|---|---|
| reached this stage | People whose furthest rung is this one or lower down the ladder | None — it is a count, not a rate | Always |
| move on to <next stage> | Of the people who reached this stage, the ones who reached the next one | Everybody who reached this stage | Only when the tool records stage dates. Below 30 people it is labelled “too few people to rely on” |
| The typical time to the next stage | The median number of days between the two stage dates | Only the people carrying a date for both stages — usually fewer than the conversion above it | Ten or more people carry both dates. Below that, no time is shown at all |
| “within <n> days for three in four” | The 75th percentile of the same spans | The same people as the median | The same ten-person floor |
| here now | People whose current stage is this one | None — a count of where people sit today | Always |
| went past without ever being marked <stage> | People who reached a later stage with no date for this one | None — a count | Only when the tool records stage dates |
| From the first stage to the last | People who reached the last stage, over people who reached the first | Everybody who reached the first stage | Two or more stages, with somebody at the first |
Every rate on the tab prints its two counts beside it. Quote both.
The median is used for every time, never the average. One deal that took two years would otherwise move the number for a whole quarter of ordinary ones. The line under the chart names both the median and the three-in-four figure so you can see the spread rather than one point in it.
Tip
A conversion and a timing figure on the same step almost never share a denominator. “62% move on, typically in 9 days” usually means 62% of everybody who reached the stage, and nine days measured on the smaller group carrying both dates. Say which when you quote it.
Where the biggest drop is judged, and which groups get named
The step called out as the biggest loss is the one that loses the most people in raw count, chosen only from steps where at least 30 people reached the earlier stage. A step that loses everybody out of eleven people is not the biggest loss in your funnel.
Under it, MarqIQ names the groups moving through that step more or less often than the rest. A group is named only when all three of these hold:
- At least 30 people in that group reached the earlier stage.
- The difference from the overall rate for that step is larger than chance could explain, at 95% confidence.
- The gap is worth reading: a group whose rate sits between 80% and 125% of the overall rate for that step is left out, however certain the difference is.
Groups falling behind are listed worst first; groups doing well are listed best first. Each line prints the number of people it rests on. At most eight groups are listed under one step.
The groups read are the ones on the record: its source field, its latest-source field, its industry, its country, and the seniority worked out from its job title. No other attribute is tested.
Stages that go backwards
A person counted as sitting in an earlier stage than one they had already reached is counted separately, as having gone backwards. Ten or more of them raises a line telling you that something — an import, a workflow or a manual edit — is pushing people back down the ladder.
Their furthest rung is unchanged by the reversal, so a person who went backwards still counts as having reached the stage they got to. Only the here now count moves.
When your tool records no stage dates
MarqIQ then reads the tab in snapshot mode, and says so at the top: it can show where people sit today and not how they moved. No conversion between stages is shown, no timing is shown, and no drop-off is claimed. A conversion is never inferred from a snapshot, because a snapshot cannot tell the difference between a person who passed through a stage last year and one who never did.
Before advising you to start recording stage dates, MarqIQ looks through the fields you already have. Where date fields exist whose names read like stage dates, it names up to four of them and tells you it does not yet read them as stage dates for this tool — so nothing new has to be built on your side. Where there are none, it says that the fix is a date field per stage, stamped by an automation at the moment the stage changes.
| Tool | What it calls the ladder | Does the tab draw a dated funnel? |
|---|---|---|
| HubSpot | Lifecycle stage on the contact, on every plan | Yes. HubSpot stamps a Date entered property for every stage, custom stages included, and those are the dates MarqIQ reads |
| Marketo Engage | Person Status, and the stages of a revenue model | Snapshot only today |
| Salesforce | Lead Status, and Contact Status where you use it | Snapshot only today |
| Zoho CRM | Lead Status, and the CommandCenter journey, on all editions | Snapshot only today |
| Microsoft Dynamics | The lead's Status Reason, and the Business Process Flow stage | Snapshot only today |
| Pipedrive | No stage field of its own on a person. Build one, or use Leads and the Leads Inbox | Snapshot only today |
| Pardot (MCAE), Mailchimp, Salesforce Marketing Cloud, Adobe Campaign, Adobe Journey Optimizer | Not verified. MarqIQ says so rather than guessing at the setup, and nothing on the tab is written as a judgement about your data | No |
The middle column is what to set up in your tool, and the setup card on the tab gives the steps for it. The right-hand column is what MarqIQ reads back today.
Why only HubSpot draws a dated funnel today
MarqIQ reads the stage a person sits in from the lifecycle-stage field on the contact, and the date they entered each stage from HubSpot's Date entered properties. No other tool's stage dates are mapped yet. That is why a Salesforce or Zoho CRM workspace with a full status ladder still reads in snapshot mode, and why the tab names the date fields you already carry rather than telling you to build new ones.
What the whole tab was read on
The first line of How we worked this out names the number of people read. Where that is fewer than the number in the store, the same line says so and states both figures — an even sample drawn across your whole history, so recent arrivals do not skew it. Shares from a sample are reliable; the counts are the sample's, not the store's.
If it did not work
A stage shows more people than the stage above it
Why: People reached the lower stage without a date for the higher one, and the ladder MarqIQ read is not in the order your business uses them.
Fix: Read the skip count on the step above. Where it is large, the stage is not being stamped. Where it is small, the ladder order in your tool is not the order people really move through, and the stage list is what needs fixing.
A step shows a conversion but no time
Why: Fewer than ten people carry a date for both stages, so no median is claimed.
Fix: On HubSpot, stamp the earlier stage date by automation and the timing appears once ten people carry both. On every other tool, stamping dates will not bring the timing back yet, because MarqIQ does not read that tool's stage dates — the note above says so.
The tab shows where people sit and nothing about how they moved
Why: MarqIQ found no stage-entry date it reads, so the tab is in snapshot mode. On any tool but HubSpot that is the normal reading today.
Fix: Read When your tool records no stage dates above. Where MarqIQ can see date fields that look like stage dates on your records, it names up to four of them there, so you know the data exists even though nothing reads it yet.
The conversion says “too few people to rely on”
Why: Fewer than 30 people reached that stage, which is the floor below which a rate is not evidence of anything.
Fix: Widen the Created range at the top of the page, or wait for the population to grow. The figure is still shown; it is the reliance on it that is withheld.
Nobody appears at a stage you know people pass through
Why: That stage is being skipped wholesale, or its key does not match the ladder MarqIQ read from the tool.
Fix: Check the stage name in your tool against the row label on the tab. Where they differ, the values in the tool have been renamed since they were set up.
Next steps
Related articles
Was this article helpful?