Note: This post is based on the Forrester report “How To Diagnose Channel Performance Issues In Chat,” prepared by an experienced analyst. Get the full report here.
Chat programs are convenient and mimic how individuals interact in their everyday lives, making them valuable tools in delivering the service experience that customers want. However, launching and maintaining a successful chat channel isn’t without its challenges.
When issues arise, it can be difficult for customer service leaders to identify what’s at fault and what steps to take to fix them. The good news is that, in general, most fledgling programs experience similar challenges and, by adopting a diagnostic approach, leaders can gain the skills and insights to drive success.
Common issues organizations face
The steps that leaders take when first setting up their chat programs will influence the experience and efficacy down the road. As we’ve mentioned before, the strategic groundwork must be in place before launching your program.
Assuming your strategy is good and your KPI expectations don’t contract your channel goals and you still come up against issues, it’s likely that you’ll be experiencing one (or more) of four issues: high abandonment, low issue resolution, low efficiency, or poor adoption.
Overcoming high abandonment rates
Generally, there are two reasons you may be experiencing high levels of abandonment on your chat channels: the initial time spent waiting for an agent is too long or the chat duration is too high. Either way, duration is the main factor you need to look out for with high abandonment rates. But what are the reasons for this?
Poor workforce planning
While periods of high demand are normal, if your agents are consistently at full capacity and long queues form, there’s likely an element of bad planning involved – or you need to hire more agents.
If the number of agents is theoretically enough to meet the demand, you will need to take a look at your efficiency processes. For example, if there are consistent times of high and low demand, this is an issue of forecasting and you will need to track arrival patterns to forecast chat volumes. This will allow you to better organize your agents.
Underutilization of automation
Conversational AI tools are essential tech partners to help alleviate the workload and increase efficiency. Without them, agents will likely spend time on conversations with simple resolutions that didn’t require live support.
Agents should be sorted into skill groups that avoid creating artificial bottlenecks. It’s true that most agents should be able to handle almost all situations, but organizing language skills, specialized technical knowledge, or chats with specific regulatory requirements can help avoid long wait times. If this is a consistent issue, consider training agents in the skills that are in highest demand.
At times, you may experience a high abandonment rate even though the queue is, theoretically, manageable. When this happens, look at the concurrency as if each agent is juggling too many cases at a time, it can result in similarly frustrating wait times.
Finally, the last reason you may be experiencing abandonment is due to start-stop conversations. Even if the average response time looks good on paper, if the chat tempo is off – meaning the pace of message and response – the customer gets the feeling that they’re being ignored and will leave as a result.
Addressing low issue resolution
The next problem that many chat programs face is the problem of low issue resolution. When it comes to chat channels, engaging how well you’re doing isn’t easy. As a rule of thumb, compare resolution rates across different channels. In theory, they should be similar (for similar inquiry types) across video, voice, and chat. If not, you may be experiencing the following problems.
It’s common to limit the nature of the inquiries that your team deals with in chat, often for understandable reasons. However, this can lead to a service pain point where customers need to invariably be transferred to another channel. Make sure to audit your rules on a regular basis to see if they are customer-centric or that there’s a genuine need for them. Otherwise, try to set channel expectations as clearly as possible to cut down on transfer frustration.
Premature chat resolution
Do your chats occasionally end before a satisfying resolution has been achieved? When this happens, it’s often due to an automated action such as automatic timeouts. If a customer clicks to a different tab while waiting for a response, this can trigger a timeout – which sometimes ends after just 20–120 seconds.
Incorrect chat routing
While agents should be able to deal with a large number of issues, specialized support is a natural part of customer service, whether technical, language-related, or something else. If the wrong agent is assigned an issue they’re not trained to deal with, they will be forced to transfer them to a colleague. A certain amount of transfers is normal but an unnaturally high number points to an issue with intent capture and routing. Consider replacing complex pre-chat forms with a chatbot that’s better able to gauge intent and review the logic or train periodically.
Agent knowledge gaps
If there isn’t an issue with intent capture and routing, it may be that your agents don’t have the skills or knowledge necessary to properly help customers. Contact centers often have high turnover rates and ensuring agents have the necessary level of training is an ongoing task. To help mitigate this, consider developing a database of canned responses for the most common situations.
Efficiency is a balancing act and the triggers put into place to increase it can, if not implemented properly, have the opposite effect. To achieve higher levels of efficiency, service leaders need to combine a number of metrics to ensure easier calibration at the agent level, while ensuring that customer-centric metrics like first contact resolution are taken into account. When the balancing act isn’t right, it may be because of:
If not properly accounted for, concurrency can lead to a phenomenon where an agent appears occupied – as they are actively engaged in chats – but are not operating at their full capacity. For voice or video calls, an agent can only deal with one issue at a time, but when it comes to chat channels, one agent can be expected to hold two or three conversations at the same time.
As mentioned, efficiency is a balancing act and demanding too much from an agent can have similar results as not using them enough. If you’re finding a high AHT but a quick response rate, it may be an issue with concurrency. A high AHT is a sign of too many concurrent chats, and, when under pressure, agents tend to keep customers warm with soothing messages as they try to contend with backend systems and other conversations. However, this will only work for so long and unless they find a resolution, customers will become frustrated and leave due to low efficiency.
A high AHT coupled with a long response time is unlikely to be a concurrency issue as something else is taking agents’ time and attention. Commonly the root of this problem is in a lack of available resources or an inefficient knowledge base system that takes time to sort through. Addressing this issue is either a matter of further training or streamlining internal processes.
Struggling to get your customers to actually use your chat channel? No matter how good your program is, if the customers don’t use it, it’s all for nothing. If you’re experiencing this problem, there are a number of likely issues at play.
Hidden or inconsistent chat button
When launching a chat channel, companies sometimes make it hard to find their chat channel to avoid overwhelming the system before agents are properly changed. However, after a certain time, this needs to be expanded and even promoted to ensure proper uptake. Similarly, there’s the ongoing debate as to whether you should queue or not. Toggling on and off the chat button as agents reach capacity seems logical, but it can have adverse effects down the line. Experimentation is the key here and if you’re having trouble related to inconsistent chat buttons, try setting expectations instead by clearly labeling an approximate wait time.
Insufficient channel promotion
Once your agents are properly trained, the efficiency that’s possible with chat channels needs to be properly leveraged. For this to happen, it’s important to promote your chat channel as a simple, easy-to-use service. What occasionally happens is that a broad offering of chat-based touchpoints leads to a fragmented experience. Having options is fantastic – provided they don’t confuse the customer. If you’re having issues, consider developing an information campaign to promote proper usage of your channels.
Self-service is doing its job
Finally, what may at first appear to be a problem, might not be. One issue that service leaders have is apparently poor adoption that is directly related to a well-performing self-service platform. Before you begin addressing different issues, it pays to look at how well your chatbot or similar is resolving issues.
Want more in-depth insights into the issues discussed in this post? Read the Forrester report “How To Diagnose Channel Performance Issues In Chat” here!
Unblu is a leading Conversational Engagement Platform that’s fine-tuned for the financial services industry. Centered around three pillars – texting, video & voice, and collaboration – we help retail banks, wealth management firms, and insurance companies to provide faster, more secure customer service experiences.