A chat about the wrong KPI

Are you using the right KPIs for measuring the effectiveness of chat software deployment in your organisation? When I was undertaking recent research into the crowded, commoditised and convoluted market I started examining the KPIs in use.


Let me share them with you, and see if you can spot what's really missing here.

Vistor KPIs

  • Live visitors on calls
  • Current visitor pages or source, and page views
  • Lifetime value
  • Churn rate
  • Return visitors
  • Satisfaction
  • Campaign source (UTM/Social)

Chat KPIs

  • Total chats
  • Chats per day (average)
  • Chat duration (min/max/average)
  • Cost per chat
  • Revenue per chat
  • First contact resolution
  • Missed calls
  • Offline chats
  • Lost/unqualified leads
  • Deflection and leakage
  • Transfer to other agent

Agent KPIs

  • Number of agents handling calls
  • Agent utilisation rate (per hour)
  • Agents online (skill, country, available)
  • Time to answer
  • Longest chat waiting
  • Agent occupancy
  • Agent call quotas
  • Schedule adherence
  • Staff attrition rates
  • Accepted chats
  • Star rating (performance)
  • Median first reply time per operator
  • Availability of agent per hour
  • Productive on chat
  • Email follow up
  • Post call wrap up
  • Idle/waiting for chat
  • Training
  • Administration and paperwork

What's missing?

So, did you spot it...perhaps I'm just a bit old fashioned, but surely we really need to be looking at the value of the chat and objectively rate the quality of the visitor/agent interaction. I'm not talking about the often fuzzy 'sentiment analysis' but the real content of the chat. For example are we effectively looking at how the language and process adopted by agents is driving a profitable outcome from the chat interchange...or is the focus still on producing reams of easy to produce statistics to feed the management hierarchy?

If you really want to make a difference, we need to be looking at intra-interaction content and spotting patterns. For example, wouldn't it be great if we could understand more about different aspects of the actual content. Perhaps in terms of:

  • Readability quotient
  • Buying signals
  • Frustration and displeasure
  • Idioms, buzzwords and sarcasm
  • Technical competency
  • Persuasive phrasing
  • Unrealistic entitlement demands
  • Ambiguity and clarity
  • Purchase intent (now and future)
  • Achievement of customer goal

If you want to achieve higher value chats, make them truly effective, and optimise agent productivity on profitable tasks; let's have a chat about what can really be achieved. Contact naomi@amo.ai today.