Count the Leads That Got a Conversation
Every owner knows last month's lead count. Almost nobody knows how many got a real conversation. How to find that number in an afternoon.
Ask any owner how many leads came in last month and they'll tell you. Forty-two. A hundred and ten. Whatever the CRM says. Ask how many of those leads actually got a conversation with a human, and the room goes quiet.
That second number is the one that matters, because the gap between the two is money you already spent. Every lead was paid for: an ad click, a Google Local Services charge, a referral you earned, a truck wrap somebody noticed. The ones that never got a conversation are acquisition cost with nothing on the other side of the ledger.
This is how to find the number. It takes an afternoon, and you will not like what you see.
Step one: pull every lead source into one list
Start with a single spreadsheet and one row per lead for the last full month. Pull from every place a lead can land:
- Inbound calls (your phone system's call log, not your memory)
- Web forms (the form tool's submissions, not the emails that survived your inbox)
- Text messages to the business line
- Facebook and Instagram messages
- Google Business Profile messages and calls
- Chat widgets
- Third-party lead platforms, if you buy from any
Most owners find a lead source they forgot about during this step. A form that routes to an old employee's email. A chat widget nobody has opened in months. Those are leads you paid for and never even saw.
Step two: define "conversation" strictly
A conversation means a two-way exchange where the lead's question was answered and a next step was proposed. That's the bar. Under that definition, these do not count:
- A voicemail left for the lead
- A "thanks for reaching out, we'll get back to you" auto-reply
- A single text with no response
- An email that was never opened
Be strict here, because a loose definition hides the problem. An auto-responder is not a conversation; it's a receipt.
Step three: mark each row
Go down the list and tag every lead with one of four outcomes:
- Conversation, booked. They talked to someone and an appointment, estimate or sale was created.
- Conversation, not booked. They talked to someone and said no, or went quiet after.
- Attempted, no conversation. Somebody tried once, left a message, and moved on.
- Never touched. Nobody reached out at all.
Rows three and four are the leak. In our experience running this exercise with owners, those two categories are rarely small, and the "never touched" row is usually where the after-hours and weekend leads live. Industry research backs that up: roughly 27% of business calls go unanswered on average, and the form and text leads are often worse.
Step four: put a dollar figure on rows three and four
Take the number of leads in rows three and four. Multiply by your normal booking rate (the rate from row one and two, not a hopeful one). Multiply by your show rate. Multiply by your average ticket.
That's the recovery figure: the revenue sitting in leads you already paid for, run at the close rate you already have. Notice that nothing in the calculation assumes you get better at selling. It changes one variable only: how many leads get a conversation. That's the whole claim, and it's why the number holds up when you check it against your calendar. If you'd rather have it computed for you, the recovery calculator runs exactly this math.
Step five: find the pattern before you fix anything
Before you buy or build anything, sort the leak by time of day and by source. The pattern will tell you what to fix:
- Clustered after 5 PM and on weekends means the leads are arriving when nobody's on the phones.
- Clustered on web forms and social messages means your process treats calls as real and everything else as optional.
- Spread evenly through the day means you're at capacity, and the front desk is triaging by whoever is loudest.
- Heavy in one source means that source deserves either a faster response or a smaller budget.
Each of those has a different fix, and only one of them is "hire someone."
What an AI front office changes about this exercise
The reason most businesses can't answer the conversation question is that no system is counting. Calls live in one log, forms in another, texts on someone's phone. An AI front office sits across all of them, has the conversation in seconds regardless of channel or hour, books what can be booked, and reports the counts back every morning. The number nobody tracks becomes a line in a daily brief.
Run the audit first, though. Owners who run it themselves believe the number. Owners who are told the number argue with it.
Frequently asked questions
How far back should I audit? One full month is enough to see the pattern and small enough to finish in an afternoon. If your business is seasonal, pick a month from the busy season, since that's when the leak is largest.
What if I don't have a call log? Your phone carrier or VoIP provider almost certainly has one, even if you've never opened it. For a cell number, the carrier's online account shows inbound calls. It won't show forms or texts, which is a reason to fix the logging alongside the response.
Should I count spam and wrong numbers? Strip them out before you calculate. Counting junk as "never touched" inflates the leak and makes the exercise easy to dismiss.
Isn't this just a speed-to-lead audit? Speed is part of it, and being first still wins. But a lead that got a fast auto-reply and no follow-up still shows up in row three. This audit measures whether a conversation happened at all, which is the harder and more useful question.
Bring your row-three and row-four counts and your average ticket, and we'll run the recovery figure live. See if you qualify.