Skip to content
  1. Home
  2. Glossary
  3. Key event
Conversion

Key event

The funnel step you declared to be the main predictor of a sale — the earliest signal that is an action by the buyer rather than an opinion held by the seller.

Formula
leads that reached the stage declared to be key
Unit
Count
Direction
Higher is better
Across periods
Adds up across periods
In KO
keyEvents

A key event is the funnel step you have nominated as the main predictor of a sale: a demo held, a trial started, a proposal sent, an NDA signed. Every business has its own, and it is chosen once and deliberately.

The point is to get a signal earlier than the sale but more reliable than an opinion. Qualification is a salesperson’s judgement, and it can be moved by a conversation in a stand-up. A sale is a fact, but it lands weeks or months later, when there is nothing left to change in the marketing. The key event sits between them: an act by the buyer, early enough to still react to.

The formula

Key events = number of enquiries that reached the stage declared to be key

On its own the count says little — the conversions around it do the work:

CR QL→KE   = key events / qualified leads
CR KE→Sale = sales / key events

The first asks whether you get the people you selected to actually do something; the second asks whether this is a predictor at all. If the second is consistently high, the event is well chosen. If it swings from period to period, you have declared something that does not end in deals.

What it does not show

It is not a measure of money. A demo with a large account and a demo with a student are one key event each. The count rises equally from both; revenue does not. Reading it without revenue alongside works right up to the first quarter with more events and less money.

It is not a commitment. A started trial, a downloaded document, a booked meeting — all are actions with a low cost of walking away. They predict deals statistically, across a flow, not in any individual case.

It stops working once people are measured on it. The moment the number of key events becomes a team target, somebody finds a way to raise it directly: meetings booked with everyone, trials handed out without selection. The count grows, conversion to sale falls, and the correlation the event was chosen for evaporates.

It does not transfer between companies. “Demo held” means different things in two similar businesses — forty minutes with a solutions engineer in one, a five-minute call in the other. Benchmarking your conversion rates against somebody else’s by the name of the stage is meaningless.

How it breaks

The event is too early. If almost every selected lead passes it, conversion into it sits near 100% and the metric reports nothing — it merely restates the qualified count.

The event is too late. If only a handful reach it, the count is small, the noise is large, and any weekly movement looks like a collapse or a doubling. A predictor you cannot act on before quarter-end is not doing its job.

The event is recorded by hand. A stage a salesperson sets also measures how diligent that salesperson is. An automatically recorded event — a trial started, a proposal opened — is sturdier, though it breaks too whenever the way it is logged changes.

Moving the key stage retroactively. As with qualification: move it and the whole history moves. Both conversions around the event change at once, and nothing in the report distinguishes a changed definition from a changed business.

How KO calculates it

In KO, keyEvents counts the enquiries whose CRM stage belongs to the set declared to be key. Like qualification, this is a project setting: every funnel is different, and the product does not decide for you which step predicts a deal.

Two derived conversions sit around the counter in the registry — crKe (qualified to key event) and crSale (key event to sale). Both are computed from summed inputs at every level of the report and are never averaged up from the level below, so a quarterly rate equals the ratio of quarterly sums rather than the mean of three monthly rates.

The metric is additive. Where the denominator of a derived conversion is empty, the result is an empty value rather than zero: having no selected leads does not mean the conversion was nil.