Chat-based customer service is growing quickly. But there’s one central question that I think is not being addressed as critically as it should: Is chat acting as a substitute for phone calls? This is a harder question to answer than you might think.
First, to be clear, I’m not talking about chat-bots, i.e. automated self-service, I’m talking about a conversation with a human agent via some form of text-based channel. (Chatbots are an important topic too, which we’ve covered separately here and here, and will revisit shortly.)
If it is proven that a significant amount of conversations that would have been voice are moving to chat, that has huge implications for the 40,000 US call centers and the $25 billion BPO industry. But note the italicized part of the sentence; That crucial point is often glossed over by analysis of this topic. Why is it hard to get a definitive answer?
Is Chat Replacing Phone Calls?
Before we dig in, it’s worth restating the focus. It’s a really unfortunate coincidence that “chat” and “chatbots” ended up so lexically similar. This post is only concerned with human assistance via chat, not automated responses.
[It’s doubly unfortunate that we also use “chat” to refer to interpersonal communication (like “chatting over Skype”) and to team collaboration tools (like Slack, Microsoft Teams, Jive, and others). That one little syllable is carrying a lot of weight right now!]
So we’re really focusing on the two ways human agents can deliver real-time customer service: voice and chat. Once again, the key question is:
Is chat becoming a substantial substitute for phone calls?
If so, it leads to a more powerful question:
Can we project this forward and expect a decline in voice agents and call centers?
Of course, we can find lots of anecdotes. Everyone can tell you about a great chat experience they’ve had (or a bad one). Chat vendors love to tell anecdotes about great deployments. But the plural of “anecdote” is not “data”. What does the real data tell us?
Well, we have lots of data that answers similar questions, but none that really hits it square-on. Some examples, in increasing order of helpfulness:
1) Data Showing The Overall Volume of Messaging
“There are 20 billion SMS messages per day.” [source]
“Chat is more popular than social media.” [source]
Doesn’t distinguish customer service from person-to-person communication; or between human-generated messages from automated ones. Counts each “turn” in the conversation separately, leading to a really high number. (How many individual messages would have been required to replace the last phone call you had with a company?)
2) Data (Via Surveys) About the Growing Consumer Preference for Chat
“42% of consumers say that they prefer live chat functions because they don’t have to wait on hold.” [source]
Preference doesn’t always translate into usage. Someone might have that preference, but not actually act on it when the time comes, others might have the preference, but chat isn’t available from the company, or they don’t know how to chat with the company.
By the way, if you’re interested in fixing consumers’ aversion to waiting on hold, call-backs are just the thing.]
3) Data Showing More Companies Deploying Chat
“1 in 4 US contact centers are looking to implement web chat within the next 12 months.” [source]
Doesn’t translate directly into usage.
4) Data Showing Growing Usage of Chat
“Contacts handled by chat increased from 1.6% to 2.6% between 2015 and 2016.” [source]
Often fails to distinguish between the use case.
Let’s explore that last example: remember what we’re trying to understand is the substitution factor. Consider the following use cases:
1. Customer uses chat to find the store’s opening hours, or if an item is in stock. This is replacing what might have been a self-serve web interaction or a self-serve IVR interaction.
2. Customer uses chat, but their issue isn’t resolved and a subsequent phone call is needed.
3. Customer chats with an agent and then it escalates to a voice call.
In all three of these cases, a phone call was not replaced by chat.
I am keeping an eye out for data that tackles this definitively. Please let me know if you see any!
Another Perspective on Chat vs. Voice
We can also look at this issue from the “other side”: Are customer service voice calls declining?
Here, data seems to say no.
That doesn’t mean that chat-substitution isn’t happening, because the total volume of all transactions might be growing. But, if we did have a drop in phone calls, that coincided with the rise of chat, that would be a tempting correlation to make.
Chat is Vital and Here to Stay
This post is not intended to disparage the value of chat. There exists lots of solid and positive data regarding chat deployments: High satisfaction scores from chat interactions; agent head-count reductions; cost savings; increased NPS. I am overall bullish on chat as a big part of the customer service mix.
Even if there isn’t a substitution effect, live chat is important.
If there is a substitution effect, it’s a big story, but we simply don’t know yet. I think it’s important to keep that in perspective and not get ahead of the data.
Analyst Donna Fluss summed it up well in a recent post:
Chat is an excellent communication channel for certain types of interactions, but in most cases, chat remains a valuable complement to, not a replacement for, the voice channel for customer service. Despite chat’s many advantages and strong following, it cannot be viewed as a replacement for phone calls. DMG research has shown that in times of need, when emotions run high or time is short, people of all generations (Millennials, Generation Xers, and Baby Boomers) prefer to call a company. No matter how good a chat agent is, it’s much easier to pick up subtle cues and personalize a conversation’ when you’re actually speaking with someone.