Reviewing calls from sales managers: what to look for and why
Calls get reviewed either all of them, which is abandoned within a week, or selectively, which finds exactly what you went looking for. Here is what reviewing conversations actually gives you, and how to keep it from becoming a report nobody reads.
Why spot-checking does not work
The usual arrangement: once a week the manager listens to a few calls and draws conclusions. The problem is not laziness, it is the sample. Out of three hundred conversations he gets through five, and those five are either random or the ones a customer complained about.
Five calls will not show you a pattern. They will show you one bad conversation, from which you then draw a general conclusion that does not apply to the other two hundred and ninety-five.
A machine listens to all of them. Not because it is smarter, but because it does not care whether there are five or five hundred.
What to look for in a conversation
A score of “seven out of ten” is useless. It tells the salesperson nothing about what to do tomorrow, and it does not survive the question “why seven and not eight”.
What works are things you can check and fix:
Was a next step agreed. A conversation with no date for the next contact is a lost conversation, however warmly it went. This is the most common and the most expensive thing to go missing.
Was the key question asked. Every business has its own: dimensions, deadlines, volume, site address. If the salesperson did not ask it, the quote has to be redone and the customer bothered a second time.
What the customer objected to and what they were told. Objections repeat from one conversation to the next, and they show what is missing in your materials, not only in the salesperson.
Who talked more. Not as a mark of guilt, but as a sign of whether the task was being understood or the company was being talked about.
An excuse and a real objection
A useful distinction that only shows up at volume. A customer with a real reason to decline repeats it from conversation to conversation in the same words. A customer who is simply uncomfortable saying no changes the reason every time: too expensive, then needs to consult someone, then later.
A single salesperson cannot see this difference: they have one conversation in front of them. Across a hundred calls it is obvious, and the two cases need different handling — the first dealt with on the merits, the second not talked round but helped to a decision.
What call reviews give the company, not the salesperson
Call review is usually sold as staff monitoring. That is the narrowest of its uses and the most unpleasant for the people being reviewed.
Something else matters more. The conversations show what customers ask about most often, and that is a ready-made list of what is missing from the site, the proposal and the training of new people. They show which objections come up for every salesperson at once, which means the problem is not the people but the product or the price.
And they show where the company loses money for no reason at all: the customer asked, nobody answered, the customer left. Cases like that never appear in any report, because there is simply no deal to report.
Transcription and review are different things
Transcription turns sound into text. It is cheap, it runs on your own server and nothing leaves it.
The review is reading that text with questions in mind: was anything agreed, what was asked, what was objected to. This is where the meaning appears, and where the cost appears too.
Keeping them apart is useful for more than cost. Transcripts accumulate and stay yours for good — you can come back a year later and look for anything, even something you were not looking for at the time.
What a review will not do
It will not teach anyone to sell. It shows where a conversation went wrong; fixing that is done by a person and their manager, not by a report.
It will not replace listening entirely. A manager should still hear a couple of conversations a week with their own ears, otherwise they lose the feel of how their team sounds.
And it must not turn into a stick. A review that produces nothing but reprimands ends with salespeople talking for the report instead of for the customer. That is worse than not reviewing at all.
Where to start
With one question, not twenty. Take a week of conversations and look at one thing only: how many of them have a date for the next step. That is usually enough to see where most of it leaks away.
The other signals come later, one at a time, and each has to change somebody's behaviour. A signal nobody acts on gets removed.
Frequently asked questions.
Do customers have to be told the call is recorded?
Yes. Recording a conversation is processing personal data, and the notice has to be given at the start of the call or set out in the terms on your site. It is done once and then stops being a bother.
How much does reviewing one call cost?
It comes in two parts: transcription and the review itself. Transcription can run on your own server, in which case it costs almost nothing. The review is priced by the volume of text — for an ordinary five-minute conversation that is pennies.
Can written conversations be reviewed too, not just calls?
Yes, and it often gives more: in writing the customer puts the question into words themselves, and those words are usable afterwards — for the site, for replies, for training new salespeople.
Isn't this surveillance of employees?
It depends on what is done with the result. If the review exists to find grounds for reprimands, salespeople will start talking for the report. If it exists to find what is missing from your materials and where customers are being lost, it works for everyone.
What do you do with the result?
Start with one signal and one action. For example: every conversation with no date for the next step goes back to the salesperson the same day. That changes behaviour within a week, unlike a report with twenty metrics.