Per-agent metrics: how to measure the people who answer on WhatsApp
Which indicators help evaluate each person on the support team, how to read them without penalizing whoever takes the hard cases and what decisions they enable.
In a nutshell
Measuring by person helps you spread the work better and spot who needs help—not build a leaderboard. A few well-read indicators are worth more than a dashboard full of numbers.
The indicators that actually matter
For a team answering on WhatsApp, four numbers per person are enough: how many conversations they handled, how many messages they sent, how long they took to give the first reply and how long they take on average to reply during the conversation.
With those four you can already see whether someone is overloaded, whether a person is slow to start or whether the problem shows up mid-conversation.
- Conversations handled.
- Messages sent.
- Time to first reply.
- Average reply time.
Reading the numbers in context
A high response time can mean slowness, but also that the person gets the most complex cases or works the busiest shift. Always compare within the same period and the same type of work.
Volume isn't merit on its own either: many short conversations may be worth less than a few well-closed sales conversations.
- Same period for everyone.
- Type of cases each person gets.
- Shift and time slot.
- Outcome, not just speed.
Decisions they enable
Per-person metrics help you rebalance the workload, reinforce a shift, support someone who just joined or detect that information is missing and that's why everyone is slow on the same topic.
If a metric doesn't lead to any decision, you don't need to check it every week.
- Reassign workload.
- Reinforce shifts.
- Train with real cases.
- Fill in the business information.
Satisfaction: measuring without over-surveying
End-of-chat surveys get few responses and tend to be answered by very happy or very upset customers. An alternative is to estimate satisfaction by reading how conversations ended.
In Volga, on plans with AI, satisfaction is estimated when a conversation is closed on a 1 to 5 scale, and reports show the average of the conversations that could be analyzed as estimated CSAT. It's a signal for spotting trends, not a rating the customer gave.
- Survey: direct data, few responses.
- Estimate: doesn't depend on the customer replying.
- Use it to see trends.
- Read a sample of closed conversations.
Where to see these metrics
In Volga, reports show a per-agent table with conversations, messages, time to first reply and average reply time, calculated over the period you choose.
Administrators see the table for the whole team and can also narrow the report to one person to review their activity.
- Per-agent table.
- Each person's activity in the period.
- One person's activity in detail.
- Full view for administrators.
Frequently asked questions
Questions about this guide
Should you build an agent leaderboard?
Rarely. A leaderboard pushes people to reply fast and short even if the case isn't resolved. It's more useful to compare each person with their own history and with the type of cases they get.
Can agents see their teammates' metrics?
It depends on the tool and permissions. In Volga, administrators see the table with the numbers for the whole team.