Telecaller KPIs that predict results, and the ones that only look busy

Every calling team measures something. The trouble is that the easiest numbers to collect are rarely the ones that tell you whether sales are coming. A dashboard full of dial counts and hours signed in can look healthy for weeks while follow-ups slip and new enquiries wait until the next morning. Then the month ends short and nobody can say why.

This guide is about choosing a small set of numbers that move before revenue does, and reading them in a way that helps callers rather than watching them.

Three tests for a useful KPI

• It leads results. It changes before deals do, so you can act this week rather than explain next month.
• The caller can influence it. A number that depends mostly on lead quality or the weather is a management metric, not a caller’s KPI.
• It is hard to game. If a caller can improve the number without improving the work, they eventually will, usually without meaning to.

Six numbers that predict results

1. Connect rate

The share of calls that someone answers. It sets the ceiling on everything else: no conversation, no sale. A low connect rate is rarely an effort problem. It usually points to timing, list quality or a number that people have learned not to pick up. In LumaKol, Reports → Coaching shows each caller’s answered calls in the Caller scorecard and highlights answer rates well below or above the team’s, once a caller has at least 10 calls. The Best time to call heatmap shows the share answered by weekday and hour, so a low rate can be tested against the calling window before anyone is blamed.

2. Speed to lead

How long a new enquiry waits for its first call. Someone who filled in a form a few minutes ago still remembers why; the same person tomorrow may have spoken to three other companies. Coaching shows the median Time to first call for new leads and how many were called at all, with a breakdown from “Under 5 minutes” to “More than a day” and “Not called yet”. The last two groups are where money leaks quietly.

3. Follow-ups kept

A follow-up is a promise: “I will call you on Thursday afternoon.” Keeping it is the cheapest way to build trust, and missing it hands the deal to whoever calls first. Coaching’s Follow-up discipline counts every follow-up that fell due and whether the lead was called on time (from 30 minutes before to 2 hours after it was due), late, or not at all. Follow-ups on time also appear for each caller and in the Team score on the dashboard.

4. Talk time and average call

Total talk time shows how much real conversation happened; average call length shows its shape. Neither is good or bad on its own. Very short average calls with a high connect rate often mean callers are being cut off at the opening line, which is a script problem. Long average calls with few next steps can mean a caller who is friendly but never asks for the visit. Read both next to the outcome mix.

5. Outcomes logged

Every call should end with an outcome. This is a data-quality number rather than a sales number, but every other metric depends on it: without outcomes you cannot see why leads are lost or which source produces visits. The attendance report shows outcomes logged per person per day, and the dashboard leaderboard ranks the team by outcomes logged in the last 7 days.

6. Conversion by stage

Deals won is the result, but it arrives late. The leading view is where leads sit in the pipeline and where they stall. The Pipeline funnel in Reports shows how many leads are in each stage right now, and a stage that keeps growing week after week tells you exactly which conversation the team is struggling with. The forecast below it estimates how many open leads should close from each stage’s chance to win, and monthly targets for calls, connected calls, deals won and revenue show each caller as ahead, on track or behind.

Two numbers that mislead

Raw dial counts

Dials are the easiest number to collect and the easiest to inflate. Redialling a number that never answers, hanging up at the first hesitation, or working the shortest calls first all push the count up without moving anything that matters. A dial target also rewards the wrong behaviour at the edges: the caller who spends ten minutes rescuing a difficult enquiry looks worse than one who skims. Treat calls as a floor that confirms work is happening, never as the goal. Attempt limits in Settings → Calling policy help here too, because a lead that reaches its maximum attempts leaves the queue instead of being dialled again for the count.

Login hours

Being signed in is not the same as working. A browser tab left open over lunch or an app installed on a phone in a drawer says nothing about calls made. LumaKol’s attendance report separates time on site from active time, which comes from the web app while a tab is open and the calling app while it is in the foreground, and shows breaks and idle time alongside. That is a far better presence measure than hours signed in, but it is still presence. Use it to spot a day that went wrong, and judge the work by the six numbers above.

A few others deserve caution. Talk time set as a target invites padded calls. Conversion rates on a handful of leads swing wildly from week to week. And a team average can hide a spread where two callers carry the rest, which is why the scorecard compares each caller with the team.

A weekly review that does not feel like surveillance

Daily numbers are noisy; a single bad afternoon can halve someone’s connect rate. A weekly look at the last seven days is enough to see a pattern and still recent enough to fix it. A useful review takes about thirty minutes for the manager and ten per caller.

1. Start with the team. Read the Team score and the Coaching numbers for the week before looking at any individual. If everyone’s connect rate fell, the cause is a list or a calling window, not a person.
2. Find the outliers, then ask. The scorecard highlights callers well above or below the team. Before drawing a conclusion, ask what happened: a new lead source, a run of wrong numbers, a sick day.
3. Pick one number per caller. Nobody improves six things at once. “Call new enquiries within the hour” or “keep Thursday’s follow-ups” is a goal someone can act on tomorrow.
4. Pair the number with a call. Listen to one recording that explains the number and score it in call reviews against your own criteria; the caller sees the score and your note, and the QA score joins their scorecard.
5. Compare people with their own last week, not with the best caller on the team.

Callers should see the same numbers before the review, not during it. In LumaKol, agents have My performance, with their calls, connected share, talk time, pipeline and attendance, along with their call reviews and their monthly targets. Nothing in the review should be a surprise.

What to leave alone

The live board is for running the day: who is available, on a call, on a break or idle, with overdue callbacks per person. It is not a reason to message someone the moment their status changes. Idle on the board means no activity for a set number of minutes while checked in, which can be a long note, a colleague’s question or a break someone forgot to start. Look at patterns across the week instead of single moments, and talk about the work, not the minutes.