Efficient staffing is key to a successful live chat strategy
Our data shows that teams with over 50 agents have the highest satisfaction rate and the fewest chats per agent, while teams with 26 to 50 agents have a customer satisfaction score more than 6 points lower, more than twice the wait time, and more than double the number of chats per agent. This latter group also has the longest chat duration. Clearly, these organizations have the most work to do toward improving customer satisfaction and managing agent capacity. Some effective methods to consider include reducing the number of chats their agents handle by deflecting more to chatbots or introducing agentfacing AI solutions like Agent Assist to help agents find information more quickly. These tools will help reduce chat duration and drive higher customer satisfaction.
While teams of 1 to 5 agents also have a lower satisfaction score, these organizations often have their agents handle multiple channels at once. When teams grow beyond 10 agents, they started to segment agents by channel. No matter the chat volume or team size, our data shows that finding the right ratio of agents to chats and staffing accordingly is invaluable for a successful live chat strategy.
Don’t sacrifice quality for quantity
It may be tempting to look at teams of 11 to 25 agents and draw the conclusion that high per-agent chat volume and the shorter wait times and chat durations that come with it are good for the customer experience – after all this band has the second highest benchmark CSAT score. But caution is in order here, as an 84% CSAT score is not the pinnacle of achievement it may seem. As we know from experience, higher agent productivity does not necessarily yield higher customer satisfaction on its own. There are always other factors in play. We would urge everyone reading this report to never sacrifice quality for quantity.
56 million chats – and that doesn’t include December – tell a significant story for each of the industries and team sizes covered in this report. When reduced to simple averages, sometimes the story offers surprises – higher CSAT for teams with fewer than 26 agents but longer wait times – and sometimes it doesn’t – the largest teams produce metrics that reflect a more sophisticated customer service operation.
But there are certain parts of the 2019 story that speak the loudest:
- Canned messages are a double-edged sword: what you gain in speed you may lose in personalization. Organizations must strike a fine balance between prepackaging responses and ensuring that their customers still feel loved
- Chatbots are not only getting more popular, they’re also getting better at handling conversations from start to finish. This is a combination of two critical factors: improvements in the way chatbots understand natural human language, and deployment scopes and integrations that set them up for success in the first place
The contrast between these two conclusions is interesting. On the one hand the data points to decreased answer automation via declining use of canned messages, yet on the other hand chatbots are getting more and more work done. How do we explain this paradoxical reality?
I think it’s quite simple. When a customer knows it’s chatting with a chatbot, it expects and accepts a higher degree of standardized responses. However when chatting with a human agent, the same customer expects more originality, and the experience will feel less genuine if they sense too much standardization.
So for 2020, organizations have to double down on their effort to meet seemingly divergent yet actually fully rational customer expectations: automate away with chatbots in the right circumstances and make sure they can get things done, but ensure human interactions stay genuine. Do that, and customers will reward you with greater loyalty
Finally, a prediction (you read it here first): as your customers get more and more comfortable chatting and co-browsing, they’ll start to engage with you on other channels including SMS and social media. Are you ready?
Data and Methodology
Comm100 researchers gathered live chat data for this report from January 1st, 2019 to November 30th, 2019. GDPR and other data protection regulations were strictly enforced by our Information Security Management team during data collection – no personally identifiable data was downloaded for analysis.
The sample size includes 56,784,708 chat interactions from organizations all over the world representing 14 industries using live chat for customer service, support, sales, and marketing. Only customers with established, ongoing live chat accounts were included. Trial and free accounts were excluded from our analysis.
These criteria are in alignment with past Live Chat Benchmark Reports from 2016, 2017, 2018, and 2019 to allow for an accurate year-over-year comparison.