The recent drop in canned message utilization is interesting but not entirely surprising because brands often feel as though they face a trade-off between leveraging the efficiency of canned messages vs. the authenticity of allowing agents to type free form. However, we find that the best brands are able to avoid this trade-off by using machine learning and closed loop experimentation to continually improve their canned content so that it performs better than unscripted content.
Michael Housman Chief Data Science Officer, Rapportboost.ai
With the demand for personalized interactions increasing, canned messaged adoption has decreased by 9 points compared to 2018, reversing the upward trend since 2015.
Our findings also show that the average canned message utilization rate per chat session has gone down across the board, with the exception of the 50+ agent bracket which actually saw a 0.48-point increase (driven perhaps by efficiency goals that often accompany larger teams).
On the one hand this trend is promising as it demonstrates organizations are getting more personalized with their communication – only breaking out canned messages as needed when chat volume gets too high. There is always a risk of canned messages sounding too ‘robotic’ and impersonal if not scripted properly.
On the other hand, well-written canned messages can offer significant time savings. The best of both worlds is to use canned messages to get you 90% of the way through the response with the click of a button, then edit the message before hitting send to personalize the interaction.
While their greatest utility is to help answer frequently asked questions more quickly and ensure adhesion to tone and brand guidelines, organizations are recognizing when it is appropriate to use canned messages and when it is not.
Average Canned Messages Per Chat by Team Size
The best of both worlds is to use canned messages to get you 90% of the way through the response with the click of a button, then edit the message before hitting send to personalize the interaction.