With all the hype around the development of AI driven chatbots – you would be forgiven for thinking that messaging is now common place in every organisation’s contact centre strategy? Whilst Customer Service use cases are emerging from the early adopter pioneers using Messenger and Whatsapp, the risk remains that automated messaging responses could fail to meet customer needs. So what are the lessons that we can learn from the ‘traditional’ text messaging via SMS and ‘established’ customer support via web chat that we can apply as a filter to remove the noise created by all the AI hype? Understanding the nature of synchronous and asynchronous communications is – and always will be – the essential pre-requisite to predicting future customer contact trends.
Outside of the voice and email channels which have been in use since the dawn of contact centres, messaging can take several different forms:
SMS text messaging has been around since the early 1990s – and was first created to provide a way for cell phone users to message each other. The ‘S’ in SMS stands for “short” and initially messages were limited to 160 alpha-numeric characters.
The first synchronous text chat service came along a few years later as a means for online users to communicate in real time and get an immediate response from the recipient. Call centres having grown up with dealing with the real time world of phone calls were comfortable supporting web chat, indeed the delays between text messages being sent and responded to an operating model where agents supported several concurrent chats. But in the drive to create high levels of agent utilisation, the inherent weaknesses of web chat as a synchronous communication method became apparent, with customers having to hang around waiting for replies to their messages that help resolve their issue.
Next there was social media messaging – through apps like twitter – where users could vent their frustration at an organisation when service lets them down. As this is a public broadcast communications service then these issues are often quickly diverted by businesses to a private message channels so that the customer issue can be resolved outside of the public glare.
And most recently there is Instant Messaging through apps such as facebook Messenger and WhatsApp. This is very much an asynchronous service allowing users to:
• send a short message, pausing only to check the message was delivered before leaving the app
• receive a notification of a reply, which is a prompt to re-open the app and carry on the conversation
• see the benefit of an asynchronous communication from the flexibility to naturally move between a live conversation and one that’s spread out across hours and days. And then next time to pick up the conversation right where they left off.
• take advantage of the richness of content that can be included in each message – emojis, photos and links can all be included leading to a conversational style of use.
The other difference that is now apparent between sync and async comms is the nature of the enquiry. A synchronous webchat session works best for complex enquiries where the user is really determined to resolve and complete their online activity – booking a ticket, checking-in for a flight or resolving a service issue. With async messaging then short, simple informational transactions are more the norm.
What drives customer satisfaction?
When we start to think about contact centre business outcomes such as customer satisfaction, then different benchmarks can be applied. For a synchronous web chat session then NPS type measures are appropriate to rate the quality and advocacy generated by resolving the issue online. I remember a project with a global airline where webchat was the jewel in the crown of their service operation. This was because consumers were spending a high amount of disposable in income in booking their holiday and it was such a key moment or truth for the family when travelling internationally.
Immediately after each sales or service chat session was completed a customer survey was pushed out. As the customer was still online then the response rate was remarkably high and 75% of customers rated the agent who had dealt with their enquiry as five star. This was because the ethos of the service was to do everything possible to support the customer to complete their transaction online, and thus avoid them falling out into the phone channel. As sales conversion rates and Average Order Values were significantly high for ‘assisted’ online sales – where a chat session had taken place – versus ‘unassisted’, then it made commercial sense to.
With async messaging support then the recommendation nature of NPS is less relevant, so instead consider surveys based on metrics such as customer effort. This means that you measure how easy it was for the customer to get support as a part of their daily activities.
Decision making criteria
Because of the inherent differences between these messaging communication methods then the use cases that you consider for SMS text messaging, live chat and IM messaging need to reflect the distinct nature of each channel. Use the table below to understand the factors to consider:
Photo credit: Multichannel Customer Experience Limited
As you can see there are several different factors to consider ranging from customer access, messaging functionality, customer segment preferences, implementation costs etc.
So when I see arguments for chat bots being deployed as a universal front end to all messaging interactions, I often pause, draw breath and think “that’s not always going to work”. Going back to my airline example then the last thing you would want to do for a customer trying to spend several thousand pounds on booking a family holiday is deflect them to a robot. Understanding the context of where a customer is in their purchasing journey is vital. If they have spent 15 minutes on your site and are now on the booking page when they click to chat, then do everything in your power to get your best agents to help them.
Now that doesn’t mean that I am against robots per se, it is just a question of how best to harness the power of the technology. If AI was being used to learn about the customer journey i.e. the pages visited and the values entered into forms and then decide on the ‘next best action’ in terms of how to respond then I would be a massive advocate in terms of the power of analytics to get the best outcome for both the customer and for the business.
Indeed we need to remember that contact centre managers now have even more choice in how they resource their support teams. In a previous Customer Think article I talked about how chat bots need to work alongside contact centre agents as well as crowd sourced customers who can be rewarded to help their peers. Indeed some of the most exciting development in AI are coming from vendors who are trying to optimise the routing of support requests between these three types of resource – bots, agents and the crowd – using the different messaging channels.
Impact on contact centre strategy
So my message is please don’t get caught up thinking about chatbots and messaging in isolation from understanding the nature of communications channels and how customers want to use them. Look for practical examples where the different characteristics of messaging best fit your digital support strategy and then test, learn and iterate these new ideas using customer feedback as your measure of success.