Live chat offers customer service representatives yet another way to resolve customer issues and an opportunity to answer questions that customers can't answer for themselves. But not all live chats are created equal.
With increased volume of queries, the ability to identify which visitors may be at risk of abandoning the inquiry may be the difference between a happy resolution or a customer letting loose on social media with their frustrations.
So it's important to understand what live chat can and can’t do, and which level of platform works best for each business.
There are three levels of live chat technology, each of which provides progressive degrees of benefits for the company and customer involved.
At the most basic level, the chat is available to all customers. The chat agent has no contextual information about the customer and no way of telling who would benefit the most from being invited to chat. Either the customer starts the chat by clicking a button or the chat is offered by default to everyone who visits the website.
At the next level, the customer service representative (CSR) can attempt to start a conversation based on simple rules-based criteria — the so-called "if-then" format.
At the third level, visitors who meet a range of more complex criteria — for example, purchasing or browsing history, propensity to buy, history in using chat combined with real-time online behavior — can be targeted for more immediate and informed intervention.
Let’s look at each of these levels in a bit more detail.
In the Beginning
First generation chat platforms offer static, reactive button based visitor targeting. This means that suppliers put a ‘click to chat’ button on their websites in the hopes that customers click on it to start a conversation if they had a query. Some offer chat to everyone that comes onto the website.
These chat platforms don't include background information about the customer’s previous purchases, browsing sessions or preferences. So the friendly ‘Would you like to chat?’ message that pops up when visitors arrive to the site (and before they need help) might actually end up annoying rather than helping and interrupt the customer’s train of thought.
This level of software doesn't align CSR availability with customer demands, so the chat service is in effect "always on." That means that it may on occasion become overloaded with inquiries, maybe at lunchtimes when people are away from work, while at other times, remain completely unused for hours on end. Maintaining a minimum level of staffing to cope with sudden increases in demand may lead to irritation for customers who expect instant responses to their inquiries.
But customers generally prefer this basic level of live chat to "no chat at all." And from the supplier point-of-view, it is a cost effective way of providing customer service and much better than asking customers to call a toll-free helpline. These solutions are most useful for websites that receive low levels of traffic or handle fairly basic, routine inquiries.
The next level up take everything a little bit further by allowing live chat administrators to add some simple rules-based logic into which visitors to target.
In this instance, customers are targeted by simple, broad rules. This rules-based approach is based on conditional ‘if-then’ trigger which is applied to website visitors to decide whether or not to invite them to chat. The if-then format is the most basic of all the control flow statements. It tells the computer program to execute a certain section of code, but only if a particular test is true.
This can result in customers interacting in chat when they may not be ready to purchase or may not be high value purchasers. There's also the risk of a certain amount of over-selection of enquirers, in that many people who would otherwise be happy and able to find answers by browsing the website themselves, will also be targeted. This raises the cost per interaction, as well as using up valuable resources that could be used for customers who genuinely could benefit from assistance.
These solutions are most useful for websites where simple rules are sufficient and personalization and high accuracy are not completely necessary or worth the investment.
Third generation chat platforms go further than rules-based logic, to apply predictive analytics which then decide which visitors to the website to approach, at what stage of their journey to intervene and what intervention would be most useful to that particular individual based on past history. It applies predictive analytics to large amounts of customer data and real time online behavior in order to predict intent. By knowing the customer’s intent, a CSR can quickly help the customer make the purchase or resolve the issue.
This data-driven approach targets customers who are most likely to make purchases with the help of chat and helps customers that have service issues that require assisted service. It determines which customers will benefit from an intervention — not those who would have been able to make purchases or resolve issues by themselves. The results are higher agent productivity, higher service levels and higher rates of customer retention.
Before You Buy
Before making a buying decision, call centers should ask themselves if the Live Chat solution is rules-based or predictive analytics-based using sophisticated data models. Is it a simple rule to offer a chat? Or does the data model determine propensity to buy, propensity to chat, the likely value of the purchase and other algorithms? Predictive chat can determine intent based on these models.
So what questions should contact center managers ask to make sure their requirements are met before making their final decision?
1. Return On Investment
Does the chat solution drive Return On Investment (ROI)? ROI can be measured by increased revenue where CSRs help drive sales. ROI can be attained by the reduction of costs by deflecting phone calls to chat for customer service.
2. Agent Productivity
Does the solution offer the option to link CSR availability to offers to chat or identify agent availability? And if not, how does the solution manage the process of asking people to chat when agents aren't available immediately?
An availability solution should use algorithms to check agent availability, the length of waiting times, the concurrent chat capability and average handling time of each chat. The overall aim of this is to maximize the number of visitors being helped, while minimizing the number of customers abandoning the website due to long queuing times.
Does the solution also provide the ability to redirect customers to the right CSR based on their specific expertise or skills?
3. Operational and Platform Expertise
What kind of resources can the supplier provide to ensure a solution can be achieved?
A third generation platform should deliver all visitor interaction details to the agent prior to the start of the chat, so giving the CSR immediate insight and understanding into the customer’s issues and relieving them of the burden of having to start their story from scratch every time they start a new chat.
These new service channels require deep operational expertise in the CSRs that use the chat solutions. CSRs for chat service require different skillsets from voice CSRs such as understanding customer intent through text and communicating by typing. Companies with dedicated chat CSRs have the best expertise and the greatest chance of showing ROI from chat services.