Telesales call metrics: the seven that move conversion

Call volume is not a metric

Most telesales reports open with one number: how many calls were made this week. It shows that work happened and says nothing about the result — because call volume can always be increased, and the effect on conversion is frequently negative.

The trouble is that call volume is the easiest number a team can control. Put a target on it and the team fills it: short calls, redials, a third attempt on a number that never answers. The report improves; the pipeline does not.

This article sets out the seven metrics that actually govern conversion: what each measures, when to worry, why combining them is dangerous, and how targets are set.

The seven metrics

These seven form a chain: each measures one step and depends on the result of the one before it.

1. 1Response timeFrom an enquiry arriving to the first call. The first link in the chain, and it affects everything downstream.
2. 2Conversation rateWhat share of dialled numbers ends in a real conversation. It measures list quality, calling hour and number accuracy at the same time.
3. 3Average talk timeNeither good nor bad on its own — it is read in context. On qualification calls short is good; in decision conversations long is normal.
4. 4Qualification rateWhat share of conversations produces a fitting lead. The main indicator of targeting and of the script.
5. 5Next-step rateWhat share of qualified leads becomes a meeting, a proposal or a booking.
6. 6Attempts per contactHow many dials one conversation takes on average. When this rises, either the list is ageing or the calling hour is wrong.
7. 7Outcome code completionWhat share of calls closed with an outcome code. This is the meta-metric that determines whether the other six can be trusted.

How to read each one

The value of a metric is in its movement rather than its level. The rules below say what a given movement points at.

• Response time up, conversation rate down — a capacity problem, not a script problem
• Conversation rate flat, qualification rate down — targeting or the lead source changed
• Qualification rate high, next-step rate low — the offer or the qualification criteria are the problem
• Average talk time suddenly short — the script is being cut off early, or the numbers are wrong
• Attempts rising while conversation rate holds — the team works harder for the same result
• Outcome code completion falling — every other number is now in doubt

The last line matters more than the rest. If outcome codes are not written, the report reflects only the part that was, and that part is not random: successful calls tend to be logged and unsuccessful ones do not. Everything then looks better than it is.

Combining metrics hides the problem

The most common reporting mistake is dropping the middle of the chain and joining the ends: «call-to-sale conversion». It looks convenient and is practically unusable.

The reason is simple: that single number hides six different problems. When it falls you cannot tell whether response time lengthened, the list aged, the script changed or the offer drifted out of step with the market — and the team starts making corrections on guesswork.

The practical rule: every link in the chain stands on its own in the report. Overall conversion appears in the quarterly review as a trend; the weekly work reads the step-by-step numbers.

Cadence: what to look at, and when

Each metric needs a different amount of time to settle, and looking at one on the wrong frequency leads to the wrong decision.

• Daily: response time and the state of the queue, nothing else
• Weekly: conversation rate, qualification rate, outcome code completion
• Fortnightly: the effect of a script change
• Monthly: next-step rate, the trend in attempts per contact
• Quarterly: overall conversion and comparison between segments

Limiting the daily view to two numbers is deliberate. Daily variation is mostly noise, and reacting to it forces a team to change direction every day — which is the most expensive management error available.

Without segments, the numbers mislead

An aggregate figure is the average of several different flows, and the average frequently describes none of them.

• By lead source: advertising, referral, cold list — their numbers differ naturally
• By new versus returning customer: conversation rate with returning customers is always higher
• By product or service category
• By operator or team — this split is for the script, not for individual judgement
• By hour and day: the calling hour is the largest single explanation of conversation rate

That last split produces the most and is done the least. Most teams that look at the hourly breakdown find a two or three hour window they have been losing.

Measuring individuals: what can be compared

Individual numbers are the most delicate part. Used well they improve the script and the training; used badly they teach a team to dress the numbers up and cost the whole report its credibility.

• Compare only inside the same segment — two people working different lead sources cannot be compared
• Call volume is not used as an individual measure
• Qualification rate misleads at the individual level: criteria can be softened, so it is read together with next-step rate
• Outcome code completion can and should be measured individually — it is a behaviour, not a result
• The most useful individual measure is listening to a sample of calls, not a number

The practical approach: individual numbers are compared against the team average rather than against each other. Where someone sits far from the average, look for the reason — sometimes it is the person, and sometimes it is the quality of the list they were given.

That distinction matters, because list allocation is rarely random: newer people tend to get the harder or older lists, and their numbers come out lower as a result. Check the list before explaining the number.

How targets are set

There are no universal target numbers here, and borrowing another company's is useless: the sector, the list, the offer and the city are all different.

1. 1Take a two-week baselineMeasure the current state without changing anything. This is the comparison point.
2. 2Pick one metricFixing three things at once hides the cause of the result. The usual first choice is response time, because it sits at the head of the chain.
3. 3Build the target from your own numberNot an «industry average» but «how much better than this month» — that is a realistic and checkable target.
4. 4Hold the change for two weeksShorter is noise; longer loses the link between cause and effect.
5. 5Write the result down and move to the next metricAn experiment that is not written down gets run again in six months.

Limits

Metrics do not improve a process — they only say where to look. In a team that tracks numbers and changes nothing, reporting becomes a use of time rather than a source of results.

How Vexvon produces these numbers

Calculating these depends less on the calls themselves than on what they leave behind.

• Each call's outcome code, attempt count and close reason are stored on the lead record
• The fields to extract are defined in the scenario, so a qualification answer is a field rather than prose
• One timeline shows the order of calls, notes and status changes
• The Excel export covers the overview, working-hours analysis and an hourly heat map — the hourly split comes from there
• Daily and weekly reports can be sent automatically to Telegram

Which cuts the reporting is built on is shown on the analytics page. The cost side — cost per conversation and return on investment — is a separate subject and is covered in telesales ROI.

First step

Do not try to stand all seven up at once. Start with outcome code completion: if that is not high, the other six numbers are unreliable and measuring them is a waste of time.

To discuss how these would be built on your own call flow, get in touch.