The problem
Every revenue team has a coverage number. Total open pipeline for the quarter, divided by the quarter’s target, expressed as a multiple — 3x, 2.4x, whatever the board deck says this month.
That number is honest exactly once: on the first day of the quarter. After that it decays, because its two halves run on different clocks. Deals leave the numerator when they close, when they slip, and when they die. The denominator — the full-quarter target — does not move at all. By week six the ratio has drifted for at least three unrelated reasons, and the number itself cannot tell you which one moved it. A team that has closed half its target and a team that has lost half its pipeline can print the same coverage ratio.
The two obvious fixes both miss. Adding a process step — another pipeline review, a weekly scrub — inspects deals one at a time and throws away the aggregate signal; you leave the meeting knowing more about six opportunities and nothing about the quarter. Buying a forecasting tool returns a point estimate of where you land, which is a different question from whether the remaining gap is still closable — and usually returns it as a score nobody in the room can argue with.
The structural shape is the whole problem: a ratio whose numerator moves for several reasons and whose denominator does not move at all. Fix the denominator and the ratio starts speaking again.
How it works
The practice is set out in a Kellblog post published 29 April 2021, which the author describes as part of a three-part series.
The convention it starts from. Kellogg opens on the standard view: most folks say coverage should be around 3.0, and the usual refinement is that a team needs to start the quarter with 3.0x pipeline coverage. He treats that as an improvement on a flat “3.0x, always” and still not enough.
What he wants instead. The target is stated in words rather than as a formula:
at all times during the quarter I'd like to have 3.0x coverage of what I have left to sell to hit plan.
Two things in that sentence do the work: what I have left to sell, and at all times during the quarter. The measurement is against the remaining gap, and it is recomputed continuously rather than snapshotted once.
The worked example. The post follows an example company, SaaSCo, through a quarter in a spreadsheet, week by week. SaaSCo starts with a 3.2x coverage ratio, which Kellogg calls pretty healthy. The post also states the scrub cadence the example assumes: sales management is supposed to scrub the pipeline in weeks 2, 5 and 8, so that the data presented in weeks 3, 6 and 9 is scrubbed. Here the week-2 scrub bled off $3,275K in pipeline, which Kellogg says he would not be happy about. By the nine-week mark:
To-go coverage has dropped from 3.2x to 2.2x during the first 9 weeks of the quarter.
The interpretation rule. This is the operative line of the post, and the reason the metric is worth computing at all:
If to-go coverage increases, we are closing business faster than we are losing it. If to-go coverage decreases we are "losing" (broadly defined as slip, lost, no decision) business faster than we are closing it.
Note how broadly losing is defined: slip, lost, and no decision are pooled together as the same thing happening to the pipeline.
What the post does not resolve. It ends unresolved by design. Kellogg writes that the quarter is not looking good for SaaSCo but that he still cannot tell what happened to the pipeline or what it means for the future, and defers that to the final post in the series. There is no outcome data here: no claim that teams adopting to-go coverage forecast more accurately, and no evidence that a particular level predicts hitting plan.
The companion argument. In an earlier post from March 2021, Kellogg inverts the ratio into a statement about win rates:
My primary coverage metric is snapshotted in week 3, so week 3 pipeline coverage of 3x implies a 33% week three pipeline conversion rate.
He then turns that on anyone asking for more coverage:
If you need 5x pipeline coverage because you convert only 20% of it, maybe the problem isn't lack of pipeline but lack of winning.
That post lists several readings of a demand for more than 3x — unvetted pipeline, weak win rates, low-quality prospects, sales deflecting accountability to marketing, missing nurture programmes, and reps squatting on accounts — and reports meeting one company that said it needed 100x.
Why it works — our read
Everything below is RevGuild analysis, not a claim made in the source.
What changes at the input. Swapping the denominator from the full target to the remaining target means both halves of the ratio now move — but they move on different triggers. Winning a deal pulls the numerator and the denominator down together. Losing one, or slipping it, pulls only the numerator down. Those are the only two cases, and they leave different fingerprints.
What that does to the mechanism. Because the two events affect the ratio differently, the ratio’s direction encodes which one is dominating. That is the informational trick, and it is why movement in this ratio tells you something neither the pipeline total nor the bookings number tells you alone. A pipeline book can read identically week over week while its composition rots underneath — same total, entirely different deals, and nothing on the dashboard flinches.
Why that moves the outcome. It converts a start-of-quarter vanity number into a weekly diagnostic with a decision attached to it. Falling coverage in week six is a build-pipeline-now signal arriving while there is still a quarter left to act in. The identical fact, read off a full-quarter ratio, arrives in week thirteen as a post-mortem.
There is a second-order effect in how losing is scoped. Pooling slip, loss and no decision means the metric counts the ways pipeline actually dies, rather than only the one that generates a loss reason. Most CRM reporting looks away from no decision and slip precisely because neither produces a tidy record.
The transferable principle: when you want direction of travel, measure against the gap that remains, not the goal you set.
The counterfactual is the ordinary quarter. Snapshot coverage once at the start, and every mid-quarter conversation reverts to anecdote, because the only number on the table is stale and the loudest deal story wins by default.
A ratio you can only read on day one is a scorecard; a ratio you can read every week is an instrument.
Where it breaks
Our assessment, not the source’s — the post raises none of these.
- It inherits your pipeline hygiene completely, and this one is intrinsic. The ratio is assembled entirely from opportunity records, so it is exactly as true as they are. Trigger: no enforced discipline about closing dead deals out. Earliest symptom: to-go coverage sits flat through a quarter in which bookings are visibly slipping — the numerator is held up by opportunities that no longer exist.
- The quarter boundary is an assumption, not a fact. The practice assumes deals resolve inside the quarter they sit in. In long enterprise cycles that span three quarters, or in land-and-expand motions where the target is not a stack of closed-won contracts, the denominator stops being meaningful. Symptom: the ratio lurches on single deals.
- Goodhart arrives fast. The numerator is the half your team controls: pulling optimistic close dates into the quarter and leaving zombie deals open both improve coverage without improving anything. Symptom: coverage climbs while win rate stays put.
- Small numbers make it noise. With a handful of large deals, the ratio is one deal’s opinion dressed as a trend.
Worth stating: the source demonstrates the mechanic on an example company. It does not report that any team adopted this and forecast better.
The anti-recommendation: do not do this if it is going onto a board slide or anywhere near compensation. It is a diagnostic. The moment anyone is measured on it, it stops measuring anything.
What has to be true at your company
Roles. Somebody has to own the pipeline dataset and recompute it weekly. At small scale that is a founder with a spreadsheet; once more than one team is contributing pipeline it needs a RevOps owner, because the definition must stay stable across quarters or the trend is meaningless.
Data and latency. Open opportunities need an amount and a close date, updated at least weekly. Latency matters more than sophistication: a CRM refreshed every two weeks turns a leading indicator into a lagging one, which is the exact failure the practice exists to fix.
A target that is actually fixed. If the quarterly number is renegotiated mid-quarter, the denominator is not a denominator and weeks cannot be compared.
Culture. A rep has to be able to mark a deal no decision without it being a career event. This is the precondition most teams fail: the metric’s entire value depends on losses being recorded when they happen rather than at quarter end.
Sign-off. None to measure it. The CRO’s to make it the standing metric in the weekly forecast call.
Translating across shapes: at eight people this is one spreadsheet column, and the value is the conversation it forces, not the precision of the number. At eight hundred the risk inverts — computing it is trivial and so is weaponising it, so the work is keeping it a diagnostic. In PLG or consumption businesses, decide what “left to sell” means first; the direction-of-travel logic survives the translation, the quarterly framing often does not.
Try it this week
One person, one hour, no approval needed.
Export every open opportunity with a close date inside the current quarter, with its amount. Sum them. Separately, take the quarterly target and subtract closed-won to date — that is what is left to sell. Divide the first by the second. That is one reading. Do it again next Monday, and the Monday after. Three points is enough to see a direction. Budget about thirty minutes the first time and ten thereafter.
What you should see if it is working: by the second or third reading, somebody asks why the ratio moved — and the answer is findable in the deal records. That question is the output. The number is just what provokes it.
How you would know it is not working: if the ratio moves and nobody can explain the movement from the records, stop — you have a data problem, not a coverage problem, and this metric will only launder it. If the ratio does not move at all across three weeks, your CRM is not being updated, and that is the finding.