Funnel math for small business: how to find the stage that leaks most

Step 1: collect five numbers

Use the last full month, or better, an average of the last three. Estimates are acceptable as long as you write down how you estimated them.

1. Qualified visits. Website visits from your market, plus walk-ins if you have a shop. Exclude obvious junk such as traffic from countries you don't serve.
2. Enquiries. Every call, text, email, social message, cart or walk-in that asked about buying. Count a person once even if they used two channels.
3. Enquiries answered within your target. Pick a target, such as one business hour, and count how many met it.
4. New sales. First-time customers in the month.
5. Repeat sales. Purchases by customers who had bought before.

You'll also need average gross profit per sale: average sale value multiplied by your gross margin.

Step 2: turn them into rates

Rates let you compare stages and months. Calculate:

• Enquiry rate = enquiries ÷ qualified visits
• Response rate = enquiries answered within target ÷ enquiries
• Close rate = new sales ÷ enquiries
• Repeat share = repeat sales ÷ all sales

A worked example

Consider a home-based catering business in Ontario, described here as a template rather than a real client. In an average month it has:

So the enquiry rate is 2.0%, the response rate is 50%, the close rate is 20% and repeat share is 3 ÷ 13, about 23%. Monthly gross profit from those sales is 13 × $450 = $5,850.

Step 3: put a dollar value on each stage

Now ask a fair question of each stage: if this rate improved by a realistic amount, what would it add in gross profit per month? "Realistic" means a change you have seen elsewhere in your own business, in a competitor or in a reasonable test, not a fantasy.

Reach. Increase qualified visits by 20%, to 3,000. At the same rates that adds 10 enquiries and 2 new sales: +$900 a month. It also costs money every month to sustain.

Enquiry. Lift the enquiry rate from 2.0% to 2.5%, through clearer prices and easier contact options. That gives 62.5 enquiries and 12.5 new sales: +2.5 sales, about +$1,125 a month.

Response. Half of the enquiries wait more than an hour. Suppose enquiries answered quickly close at 28% and slow ones at 12%, which averages to the 20% we saw. If you answer 80% quickly, the blended close rate becomes 0.8 × 28% + 0.2 × 12% = 24.8%. On 50 enquiries that is 12.4 sales: +2.4 sales, about +$1,080 a month, with no extra traffic.

Sale. Add a second follow-up on every open quote. If that lifts the close rate from 20% to 23%, sales rise by 1.5: about +$675 a month.

Repeat. Send a consent-based reminder to past customers before their usual reorder time. If repeat sales rise from 3 to 5 a month: +$900 a month, compounding as the customer base grows.

Step 4: rank by value per dollar and effort

Dollar value alone isn't enough; divide by cost. In this example:

The faster response fix wins: it is nearly as valuable as the enquiry-rate fix and much cheaper. Notice also that fixing response first makes every later reach dollar more productive, because more of the new enquiries will close.

Step 5: set a target and a test window

For the chosen fix, write one sentence: "We will raise the share of enquiries answered within one business hour from 50% to 80% by [date], and expect the close rate to rise from 20% to at least 23% over the following eight weeks." That sentence is your test. When the date arrives, you compare the actual figure with the target and decide whether to keep going, adjust or move to the next stage.

This is the same discipline the Conv.ca method uses for page tests, applied to a business process instead.

Common mistakes

• Mixing new and repeat buyers. If repeat customers are counted as new sales, the close rate looks better than it is and the repeat stage looks empty.
• Counting the wrong visits. Visits from outside your service area or from job seekers inflate reach and depress the enquiry rate.
• Using a single month. One unusual month can send you after the wrong stage. Average three, or compare year over year.
• Ignoring capacity. If you can't serve more customers next month, fix margins or repeat business rather than adding reach.
• Assuming the biggest number is the cheapest fix. Always divide by cost and effort.

A one-page template

Copy these lines into a spreadsheet and fill them in each month: qualified visits, enquiries, answered within target, new sales, repeat sales, gross profit per sale. Add four calculated rows for the rates, then one row for monthly gross profit. After three months you'll have a trend; after twelve, a seasonal baseline.

When to bring in help

You can run this calculation yourself in an afternoon, and many owners should. Bring in help when the numbers don't exist, when different reports disagree, or when you know the stage but not the fix. Our funnel audit does this calculation with your real records, and the analytics and reporting service keeps the rates updated monthly. For the response stage specifically, see lead handling and follow-up. The other guides cover testing with low traffic and follow-up rules in Canada.