Measure proposal software by revenue or by reply, the choice that changes your numbers

What to take away

A Sheffield sales manager has a spreadsheet of 340 quotes sent last quarter and a renewal conversation with her finance director on Friday. She can say how many documents went out, but not which ones led to money.

That gap is why proposal software measurement exists as a discipline rather than a feature list, and this guide closes it.

Why reply rate beats open rate

Open tracking is the weakest signal in the stack. Apple Mail Privacy Protection pre-fetches images, so a large share of "opens" in a UK business inbox are automated, not human. A number that rises when nobody reads the document is not a metric; it is noise with a chart.

Reply rate is harder to fake. Either a named person answered, or they did not. For a team sending 40 quotes a month, a reply rate moving from 18% to 26% is eight extra conversations, which is a pipeline change you can act on this week.

That does not mean discard opens entirely. Use them as a diagnostic on delivery: a quote with zero opens and zero replies usually went to the wrong address or the wrong person. Use them as a prompt to check the contact record, never as a headline figure in a board pack.

The four numbers worth defending

Revenue per quote sent

Take closed-won revenue in a period and divide by quotes sent in the same period, accepting that the two groups do not match one-to-one. It is crude. It is also the only figure that connects document volume to money, and it survives a finance review.

Reply rate

Replies divided by quotes delivered. Define "reply" in writing before you report it: a one-line acknowledgement counts, an out-of-office does not. Ambiguity here is how two teams end up arguing about the same dashboard.

Win rate by segment

Split by deal size band or sector rather than reporting a single blended figure. A 22% blended win rate can hide 40% on small deals and 8% on enterprise ones, and those two facts demand opposite responses.

Time to first reply

Measured from send to the first human response. It is the leading indicator that moves before win rate does, which makes it useful when a quarter is going badly and you need to know why.

If you want a worked model of how these figures behave across a full quarter, the benchmark research on proposal performance sets out the ranges teams report and how they were collected.

Build the report in nine steps

Step three is where most builds die. A company that appears as three names across three systems produces three partial deals and a win rate that is wrong in both directions.

Attribution choices and their trade-offs

First touch credits the campaign or source that created the contact. It flatters demand generation and long nurture programmes, because the first touch is often a webinar or a downloaded guide rather than a quote.

Last touch credits whatever was in the room when the deal closed. It flatters the sales development or account management motion, and it will make a quoting tool look more valuable than it is if the tool happens to be the last thing a buyer saw.

Even split divides credit across every recorded touch. It is the least wrong option for most small teams, and the easiest to explain. Its weakness is that it treats a thirty-second email view and a two-hour pricing workshop as equal.

Whichever you pick, report the rule next to the number. A win rate attributed by last touch is not comparable to one attributed by even split.

Swapping rules mid-year without saying so is the fastest way to lose trust in the whole report. The guide to attribution methods compares the common rules in more detail, including what each one does to a small deal pipeline.

Where the data comes from

Most of your inputs already exist. The quoting tool holds send dates, view events and document versions. The CRM holds stage changes and owner. Finance holds invoices and payment dates.

The joining problem is commercial, not technical. Companies House recorded 5,113,048 incorporated companies on its register at 31 March 2022, according to its annual report and accounts for 2021 to 2022, which is a reminder of how many legal entities sit behind trading names your CRM may store inconsistently.

One practical fix is a golden record: one row per buying organisation, with every trading name and registration number recorded against it. Building that takes a week. It saves a quarter of arguing about whose number is right.

Compliance when you track buyers

Tracking who opened a proposal and when is personal data processing. If the contact is an individual at a sole trader or an unincorporated partnership, direct marketing rules bite harder than they do for a limited company contact.

The ICO's step-by-step guide to direct marketing for small businesses sets out the consent and soft opt-in tests in order, and it is the page to read before you start scoring individual recipients on engagement.

The Data (Use and Access) Act 2025 changes parts of the marketing data regime, so verify your lawful basis rather than relying on a policy written three years ago. The text is on legislation.gov.UK.

Two rules follow. Keep the tracking proportionate to the purpose: knowing a document was delivered is defensible, building a behavioural profile of an individual buyer is a different exercise. And keep a record of the basis you relied on, dated.

Report design that survives a review

A dashboard is a communication tool, not a database. Three principles keep it usable.

Lead with one number per audience. The sales director needs win rate by segment. The finance director needs revenue per quote sent and quote-to-cash days. The marketing lead needs source mix. Putting all three on one screen serves none of them.

Show the denominator. "Win rate 24%" is arguable. "Win rate 24%, from 96 quotes sent and 23 closed-won deals" is checkable.

Version the definitions. When the win-rate definition changes, the chart should show a break, not a smooth line that quietly rewrites history. A worked example of layout and filter design is in the walkthrough of a proposal reporting dashboard, which shows how the same data is presented to two different audiences.

The mistakes that flatter the numbers

Counting drafts as quotes sent inflates volume and depresses win rate. Exclude anything never delivered.

Reporting a blended win rate across deal sizes hides the segment that is failing. Always split.

Treating opens as intent. A pre-fetched image is not a read.

Changing the attribution rule mid-year without flagging it. The trend line becomes fiction.

Measuring only the happy path. If you exclude lost deals from the dataset because the record is thin, you have measured your best quarter and called it a year. The common measurement mistakes article collects the recurring ones with the corrections teams apply.

Turning numbers into document changes

Measurement only pays when it changes what you send. Two loops are worth running.

First, a content loop. Group quotes by win outcome and compare structure: how early pricing appears, how many options are offered, whether a case study is included. You are looking for patterns, not proof, and a run of twenty quotes per group is enough to start.

Second, an audience loop. If reply rate is low but views are high, the document is reaching people who are not deciding. That is a targeting problem, not a copy problem. The audience research guide for England teams covers how to identify the actual decision maker before the quote is written.

Quality management frameworks make the same point in a different language: customer satisfaction is built by consistently delivering what the customer values, not by measuring what is convenient for the supplier. BSI's work on enhancing customer satisfaction is a useful reference when someone argues that proposal quality cannot be defined.

Governance and the monthly rhythm

A measurement system decays without an owner. Give one person the job of reconciling totals with finance each month and publishing a one-page note: the four figures, the movement, and the reason.

Where the reason is unknown, say so. A dashboard that admits uncertainty is trusted more than one that always has a story.

Quarterly, review the definitions themselves. Are you still measuring the decision you set out to support? If the quoting licence renewal is settled, the report should change shape, not keep producing the same charts out of habit.

Common questions

What is the single best metric for proposal software?

Revenue per quote sent, because it links document volume to money and survives a finance review. Pair it with reply rate so you can see whether the problem is reach or content.

Should we report opens at all?

Only as a delivery diagnostic. Apple Mail Privacy Protection and similar features mean a large share of business inbox opens are automated, so opens cannot carry a performance claim on their own.

How long before a measurement build shows something useful?

One quarter of history is usually enough to set a baseline, provided the join key is fixed first. Expect two to three weeks of data cleaning before the first credible report.

Do we need consent to track who opened a quote?

It depends on who the recipient is and what you do with the data. The ICO guide sets out the tests, and the Data (Use and Access) Act 2025 changed parts of the regime, so check the current position rather than an old policy. Keep tracking proportionate and record your basis.