How to measure quote turnaround honestly: a simple method for chemical distributors

Measure quote turnaround from the moment a request arrives to the moment a priced quote is sent, and report the median, not the average. Alongside speed, track the uncorrected line rate: the share of quote lines that needed no human correction.

Those two numbers, kept the same way for a few weeks, tell you more about your quote desk than any industry benchmark. They also give you a fair baseline before you change anything: a new hire, a new process or new software.

What exactly should you time?

Pick a start event and a stop event that can be read from a timestamp, not from memory.

Decide two rules before you start and write them down. First, clock hours or working hours: either is fine, but do not mix them. Second, what happens while you wait for the customer to answer a question. A useful split is total time, which the customer feels, and your time, which excludes the wait for their answer.

Why the median and not the average?

One slow quote, such as a large tender spreadsheet or an enquiry that sat over a weekend, pulls the average up and hides what usually happens.

EXAMPLE · synthetic data. Ten quotes in one week, turnaround in working hours, sorted: 0.5, 0.7, 1.0, 1.2, 1.5, 2.0, 2.5, 3.0, 4.0 and 26.0.

Report the median, and next to it either the slowest 10% or the share of quotes sent within your own target, for example the same working day. The slow tail is where to look for causes: missing information, a product not in the catalog, a price that needed a manager.

How do you track quality alongside speed?

A fast quote with a wrong concentration is worse than a slow correct one. The simplest quality measure is the uncorrected line rate: lines that went out right without any human correction, divided by all quoted lines except those that needed a question to the customer.

Count a line as corrected when someone changed the product, the quantity or unit, the price, or added or removed the line after the first draft. Keep a separate count of lines that needed a question to the customer: asking is not an error, it is the right response to an unclear request.

EXAMPLE · synthetic data. Four quotes with 40 lines in total; 31 went out without correction, 6 were corrected and 3 needed a customer question. Uncorrected line rate: 31 ÷ 37 = 83.8%; the 3 lines that needed a question are reported separately.

What are the traps in quote-to-order ratios?

Quote-to-order ratio sounds simple and rarely is. Define it once and keep it:

• Which denominator? Won ÷ decided quotes (won plus lost) is a different number from won ÷ all requests received. Both are useful; label them.
• Undecided quotes. Quotes with no recorded outcome quietly lower the ratio if they count as lost, or raise it if they are left out. Decide, and chase outcomes.
• Budget enquiries. Some requests are price checks with no intention to order this time. Mark them if you can.
• Revisions. A re-quote after a price change is not a new quote; count the request once.
• Partial orders. If the customer orders three of five lines, decide whether that counts as won.

How many lines before you quote a rate?

Small samples swing. With 10 lines, a single correction moves the uncorrected line rate by 10 percentage points; with 30 lines, by about 3. Our own working rule, which is a rule of thumb and not a statistical law, is to look at least at about 30 lines, from several requests and several days, before quoting a rate, and to say how many lines it is based on whenever you do.

Why do vendor benchmarks rarely transfer?

McKinsey's 2017 article on B2B buyers reported that “slow response times are by far the biggest frustration for buyers, bigger even than pricing issues”. That is a good reason to measure speed. It is not a chemical benchmark or a target to copy: it describes buyers across industries, and it is nearly a decade old.

Figures such as a share of deals “lost after 24 hours” circulate widely but are hard to trace to a primary source, and they mix industries, deal sizes and definitions of the start and stop events. A chemical distributor quoting 25 kg bags to a regular customer and one answering a 200-line tender do not share a benchmark. Your own baseline, measured the same way before and after a change, is the number that means something.

What does Vuntus record?

• Analytics (owners and managers) shows a weekly trend of requests, quotes and won quotes, “Requests by channel” (file, email, WhatsApp), “Win rate” (won ÷ decided quotes) and “Request→win” (won ÷ incoming requests).
• The audit trail shows “Avg. automated processing” and “Avg. human review”, and timestamps events such as quote created, quote approved, quote prepared for sending and quote outcome recorded.

Vuntus does not send quotes: your team downloads the PDF, HTML or CSV file and sends it from its own email, so the stop event above lives in your sent items. Vuntus has no measured result for time saved on real customer requests, and this article does not claim one. For the checks that keep corrections down, see the chemical RFQ checklist; for what to automate in the first place, see AI for chemical quoting. The audit trail itself is described in Account security and audit trail, and the quote steps whose times you can use in Creating, approving and exporting quotes. For what Vuntus is and is not, see What Vuntus is (and isn't). Sources last checked on 10 Oct 2026.