The concept of “deflection rate” in the customer service world was first introduced when help centers became a trend. At the time, they were the best way to help customers help themselves, allowing them to solve their questions on their own, which in turn saved companies time and resources.
High deflection rates were associated with complete, accurate and up-to-date knowledge bases. Customers could search for a solution before submitting a support request which would save companies time and money. For that reason, measuring how content was performing and identifying knowledge gaps was key to offering a good self-service experience.
But that wasn’t always the case, and customers resorting to a help center couldn’t always find a solution to the problem they were trying to solve. And more often than not, even when there was information available to answer a particular question, customers would simply not search for it and contact support directly. Or search was not as advanced as it is today and finding the answer to a question was just time-consuming that customers would easily give up and put that ball on the customer service court.
Luckily, the technology has evolved and customer service teams can now offer comprehensive self-service solutions. By proactively integrating the flow of submitting a request for help with reviewing relevant self-service content, usually using AI-enabled capabilities, the entire self-service experience became less frustrating for customers and way more efficient for companies.
For that reason, deflection rate has become one of the most important metrics in customer service organizations.
What is deflection and how has it evolved?
In customer service, deflection means that a decreased number of tickets are submitted for support agents to handle, and the deflection rate is what shows you the percentage of requests that were deflected because the customer was able to solve their problem without contacting support.
Before, deflection would mean getting a help center in place and have users search for articles and read them. There was no way of measuring its effectiveness other than compare help center page views versus the number of support requests.
When we refer to deflection today, we can mean different things, because it’s no longer dependent on help centers alone:
- Self-service forms, that allow customers to ask a question and get an answer based on your help center and/or knowledge base;
- Other techniques, like automatic replies, might be used to resolve support requests without user interaction.
(For consistency’s sake throughout the rest of this article, when we refer to deflection, we are referring to self-service deflection and not on automations.)
In addition to this change in the concept of deflection in customer service, in recent years, there’s been an emergence of AI-based solutions customer service teams can incorporate into their customer self-service flow to improve the customer experience. This in turn means that self-service metrics calculated purely on help center are no longer complete.
Cleverly’s technology, for example, detects the category of the question the customer is asking as the customer is filling out the contact support form, and suggests the best self-service content to answer the question. It also allows support teams to configure specific forms that can be matched with each category, to collect relevant information, instead of collecting information from the customer through a generic form that often just asks for contact details and not much else.
How do you calculate your deflection rate today?
The way we calculate deflection today is basically by looking at both the percentage of customers who submitted a ticket and the number of those who didn’t.
Back in the so-called Help Center Era, to properly calculate your deflection rate, you had to look at metrics such as page views, engagement, the number of searches that returned good results versus those that returned zero results, among other things.
With the rising popularity of chatbots and AI-based self-service solutions, deflection has gained a new significance in this context.
While in the first one, it’s difficult to get benchmarks of what good looks like, for AI-enabled self-service solutions or automatic reply workflows, deflection rates can go up to 60%, as per our customers’ data and industry-backed research. But these should not be the only metrics you calculate — drop-off rates, for example, are useful as they might indicate the customer contacts you via another channel. Another important metric is resolution rate, which indicates how successful you are at resolving customer issues with your self-service content.
To make it easier for you to calculate your deflection rate and other relevant metrics — resolution rate, form drop-off rate, and cost savings — and evaluate your self-service strategy, we have put together a Self-service Deflection Calculator that you can download here. And if you want to understand how AI can enable a better self-service experience and improve your key metrics, feel free to drop us an e-mail!