The short answer
- Lead volume and lead quality usually move in opposite directions, so counting enquiries tells you very little.
- The useful split is intent: did this person want to buy, want a price, want an update, or want a job.
- You can read intent from what was actually said, which is why transcripts are worth more than call logs.
- Once every enquiry carries an outcome, “which campaign works” stops being an argument and becomes a report.
Every business with a phone has had this conversation. Marketing says enquiries are up. Sales says nothing is landing. Both have numbers. Neither can prove anything, because they are measuring different things and calling them the same word.
Here is how to settle it.
Why lead count is close to useless on its own
Turn any campaign broader and enquiries go up. Bid on more general terms, widen the match types, add more locations, and the phone rings more. It rings with people who are further from buying, but the count goes up and so does the sense that things are working.
The reverse is also true. Tighten a campaign and the count drops while revenue holds or rises. On a volume metric that looks like a failure.
So the count moves in the wrong direction under exactly the changes that improve the business. Any metric that does that is worse than no metric.
The four things an enquiry can be
Almost every inbound enquiry falls into one of four groups, and they are worth completely different amounts.
Ready to buy
They know what they want, they have a timescale, and often a budget. They ask about availability and lead times rather than about price alone. These are rarer than anyone expects and they are the entire point.
Comparing
They want a price and they are getting three. Genuinely valuable, but only if you follow up properly, and they are the group most damaged by a slow response.
Not a new customer at all
Existing customers chasing an order, suppliers, recruiters, people who rang the wrong number. Worth nothing to marketing, and in most accounts they are counted identically to the first group.
Wrong fit
A real enquiry for something you do not do, or in an area you do not cover, or at a size you cannot serve. Worth knowing about, because a lot of them usually means a targeting problem rather than a sales problem.
How to tell them apart without listening to everything
The information is all in the conversation. The obstacle is that nobody has time to listen to two hundred calls a month, so in practice nobody listens to any of them and the split never gets measured.
Three things make it practical:
- Transcription. Reading is roughly six times faster than listening, and it is searchable, which listening never is.
- A summary. Most of the time three lines is all anyone needs: what they wanted, what was agreed, what happens next.
- An outcome from your own list. The categories have to match how your business actually describes an enquiry, not a generic taxonomy someone else wrote.
That last point does more work than it sounds. “Quote requested” means something specific in your business. “Interested” does not mean anything anywhere.
Clear Ring does this on every call, form and chat. TalkTrace transcribes and summarises, ConverseIQ scores the opportunity, and OutcomeIQ applies your own outcome list automatically. See it on real enquiries.
The signals worth looking for
When you start reading conversations rather than counting them, the same handful of signals separate the groups nearly every time:
- A timescale. “This week” is a different customer from “at some point”.
- A budget mentioned unprompted. People who say a number have usually decided to spend it.
- A specific problem rather than a general question. “My consumer unit keeps tripping” converts. “What do you charge” often does not.
- Who they are. A landlord with six properties is worth more than one job, and only the conversation tells you that.
- An objection. Counterintuitively a good sign. Objections come from people who are considering it.
What to do with it once you have it
Three things, in order of how quickly they pay:
Ring the good ones back first. On a busy day the strongest enquiry is rarely the earliest one, and working in arrival order means your best lead waits behind three wrong numbers.
Judge campaigns on the split, not the total. A campaign delivering 20 enquiries of which 14 were quote requests is beating one delivering 35 of which 8 were. That comparison is impossible without classification and obvious with it.
Feed the good ones back to your advertising. Covered properly in the piece on Google Ads producing leads but no sales, and it is the step that compounds.
What this does not fix
Worth being straight about. Classifying enquiries tells you what arrived. It does not make your follow-up faster, your quotes better or your prices right. If good leads are arriving and not converting, this will show you that clearly, and then the problem is yours rather than your marketing’s.
Plenty of businesses find exactly that, and it is more useful than another month of assuming the leads were rubbish.
Common questions
How many enquiries do I need before this means anything?
Per campaign rather than in total. Around 30 classified enquiries in a campaign is enough to see a pattern. Below that you are reading noise.
Can this be done without AI?
Yes, by hand, and that is how it used to be done. It works until volume rises, then it stops happening on the busy weeks, which are the ones you most want data from.
Is lead scoring reliable?
Treat a score as a way to sort a list, not as a verdict. The transcript and summary should always sit next to it so anyone can check the reasoning.