Win rate by rep, pipeline by stage, from the export you already pull.
Every CRM has a report builder, and every sales lead still ends up in a spreadsheet — because the question they actually have needs two objects, a date range the builder will not express, and a definition of "won" that someone changed last quarter.
The reason pipeline numbers drift is rarely bad data. It is that "win rate" means three things to four people, and whichever spreadsheet is open decides which one you get this month.
The metrics come pre-defined
Every formula, with the assumptions written down — because the assumptions are where these numbers go wrong.
| Metric | Definition | The assumption behind it |
|---|---|---|
| Win rate | won deals / closed deals | Counts deals, not value, and excludes still-open deals from both sides. |
| Won amount | value of deals closed won | Windowed on close date — a deal created earlier counts in the period it closed. |
| Lost amount | value of deals closed lost | Deal value as recorded when the deal was lost — the last quoted figure, not what the customer would ever have paid. |
| Closed amount | won + lost | The value that left the pipeline, not the value booked. |
| Won deals / closed deals | counts of each | A quarter can win more deals and less money. Both are published so you can see which happened. |
| Average deal size | won value / won deals | The average won deal, as summed value over summed count — not the average of individual deal sizes, which differ whenever the group mixes segments. |
| Open pipeline | value of deals not yet closed | Unweighted: every open deal at full value regardless of stage. As-of by nature, but the close-date window still applies, so a deal with no expected close date is not counted. |
| Weighted pipeline | sum of amount × stage probability | Your CRM's own stage weighting, not a commit or a plan. It moves whenever a deal changes stage. |
| Average sales cycle | days to close / closed deals | Closed deals only, so it is survivor data — a deal open for a year contributes nothing until it closes. |
Read the win-rate row again, because it is the one that causes arguments. “Of the deals we finished, how many did we win” and “how much of our pipeline converted” are different questions with different answers, and most spreadsheets quietly answer the first while the person reading it believes the second. Quiriz picks one, states it, and computes it the same way every month.
What it breaks down by
Every metric above breaks down by any of these, over any window you name:
- Deal owner, owner team and owner region — from your users or owners export
- Stage — so open pipeline splits across the board the way your reps see it
- Lead source — win rate by channel, computed the same way each time it is argued about
- Loss reason — with deals that carry none in their own bucket rather than dropped
- Deal type — new business, renewal, expansion, as your CRM classifies them
Each one is published here only once it is implemented end to end and verified against a sample pipeline, because an answer that quietly groups by the wrong field is worse than no answer.
What it will not do
Two things a sales leader asks for that a deal export cannot answer, said here rather than discovered in the product.
- A forecast. Weighted pipeline is arithmetic on the stage probabilities your CRM already stores. A commit, a management forecast and a quota are none of them in a deal export — nothing there records what anybody committed to — so Quiriz will give you the weighted figure under its own name and decline to call it a forecast.
- A lead or marketing conversion rate. Win rate divides deals by deals. Visits, leads and MQLs are not in a deal export, so a rate from first touch to closed-won needs the marketing data beside it.
Both are stated the same way in the product: the answer names what it would need instead of producing a number that looks close enough.
Questions, and what comes back
Four from the ten that are re-run against a sample pipeline every time these definitions change.
| Ask | What comes back |
|---|---|
| Win rate by deal owner, deals closed this year | One row per rep — won deals, closed deals, and the ratio. All of them, including the rep who closed least. |
| Open pipeline by stage | One row per open stage, summing to total open pipeline. Closed stages are absent by construction, not by filter. |
| How much did we lose, by loss reason? | One row per reason, summing to lost value. Deals with no reason recorded get their own bucket rather than vanishing. |
| Win rate by lead source | One row per source. This is the one that usually starts an argument with marketing, which is the point of computing it the same way twice. |
The exports it reads
- Deals — from HubSpot, Salesforce, Pipedrive or Zoho. Quiriz recognises each one's column names for the same field, so
dealstage,hs_object_idandclosedateneed no renaming. - Owners or users — so results break down by the person or region that owns the deal.
- Anything else at deal grain — a territory mapping, a product list, an ARR spreadsheet the CRM does not hold.
An export, not an integration. No CRM admin has to approve anything, no OAuth screen, no app in your instance. Sales ops pulls the report they already pull. The trade is that figures are as current as the last export, which fits a weekly pipeline review rather than a live leaderboard.
Why the number holds still
The failure mode with a chat assistant is not a wrong sum. It is that it re-decides what "closed" means each time you ask, so your Q2 win rate moves between the Tuesday you asked and the Thursday you presented. Quiriz routes the question to a stored definition that compiles to SQL, and every answer states whether it came from those definitions or was answered freehand — so a number in a board pack can be traced back to the rule that produced it.
That is a claim we check rather than assert. Ten questions are run against a sample pipeline whenever these definitions change, each figure compared against the same total computed independently from the raw export. When the two disagree, the definition changes or this page does.
Pricing
You pay for workspaces and the people who build them — not per seat asking questions. Every rep can check their own pipeline without another licence. See pricing →
Your files stay yours, and we do not train on your data. Privacy policy →
Related reading
- Why your win rate keeps changing — three definitions, one word, twenty points apart
- Win rate by sales rep — the breakdown, and the reps who silently vanish from it
- Weighted pipeline is not a forecast — what the figure can and cannot know
- Why is my open pipeline showing zero? — the date field, and our own four-day defect
- How to analyze a sales pipeline spreadsheet
- How to analyze a CRM export in Excel
- All guides →