WhatsApp agent performance: which metrics to measure and which to skip

Measure what an agent controls, over enough work to mean something, with the fraction shown beside the rate. For support that is first-reply time against a target, tickets resolved, close rate and queue depth. For sales it is deals won, win rate, days to win, open book and overdue tasks. Skip anything that rewards speed at the expense of the customer.

What makes a comparison fair?

• Compare like work. A person given the hardest chats should not rank below one given easy ones; use separate queues for different skill.
• Compare the same range, and enough of it. A week is noise for a rate, a month is better.
• Count in working hours. Clocks run on your working calendar, so an overnight gap is not a miss.
• Let the person see their own figures. Reports show an agent their own book.

Which figures should you skip?

• Messages sent. It rewards chatter.
• Chats closed. It rewards closing.
• Speed alone. A fast wrong answer is worse than a slower right one.
• Attainment against a quota. WavesFlo does not show it, because it cannot compute it honestly.

A target is a goal you set; the product states none and nobody guarantees a response time. The point of measuring is to find the hour or the queue that needs help.

WavesFlo counts these figures from records, shows each rate with its fraction, and leaves out any figure it cannot compute honestly. The features overview lists the support and sales reports they come from.