Last week we had a webinar on CX boost & contact center metrics held by
- Olena Iosifova, the CEO, and founder of Ender Turing
- Christina Kachnova, an organization growth expert
During the webinar, we covered the following touchpoints:
- What is customer experience? Definition and real-life example.
- Contact centers and customer experience: how are they connected?
- 5 contact center metrics that impact customer experience
- Speech analytics and customer experience
- ROI on speech analytics implementation
- Processes, tech, and mindset in an organization to boost CX
- Short overview of Ender Turing Speech Analytics
The customer-centric approach in business seems to prevail in many international companies, like Target, Nike, and Starbucks. According to Gartner, even Chief Marketing Officers shift to Chief Customer Officers because now companies prioritize meeting the needs and expectations of their customers as well as strengthening their customer relationships, which play a crucial role in customer experience (CX).
What is customer experience?
The answer is in the question - it is what the customer experiences when interacting with the business at every touchpoint throughout the customer journey.
It'll be easier to view CX through a real-life example:
Imagine you want to get a loan from a bank.
First, you google the top 5 banks in the area and find 5 names. Then, you start experiencing your interaction with the bank by reading their name and seeing their logo.
After that, you click the first name and appear on a bank's website, and you browse the web in search of specific information - how to get a loan.
And this is your experience of getting the first information you need:
- How was it?
- Was it easy, complete, and comprehensive?
- From what you've read, how the process looks like?
- Does it seem straightforward and fast, or have you started getting it as complicated and unpleasant for some reason?
Ultimately you find this customer journey a good one, and you fill out the form to start the process. Was the form easy to navigate and fill in?
And then you get a call from a banker. Their goal is to get a deal with you. And your experience of this call is also a part of your customer experience.
Basically, every step is a touchpoint in the customer's experience.
Where is the contact center placed in CX?
According to Olena, "I might not google but see an advertisement and make a direct call because I saw a good offer and I want to get more info. Or I browsed the webpage and got lost, I didn't fully understand some points and needed clarification, so I clicked for a live chat with the bank representative or just called. And contact centers are a part of this experience at every step." So this is the complete сustomer experience, where a сontact сenter can be a part of that or influence the overall customer experience through multiple touches.
But what can we do to deliver an exceptional customer experience, and how can the contact center put its part in it?
'If you can't measure it, you can't manage it.' - Peter Drucker
Olena supports the idea by P. Drucker, but she noted that 'earlier it was a little bit different: "If you don't measure it, you cannot improve it." It's by a well-known physicist Kelvin.
But it is also applicable to business performance.
To manage customer experience, we have to measure it. Ideally, we understand the customer journey and can measure CX at every touch point.'
Thereupon comes a question of what CX metrics should be measured in a contact center. How do we measure it?
For sure, the customer satisfaction score (CSAT) is the indicator. But it doesn't give us an idea of what we must change to improve. That's why we have to look out for other metrics.
Let's have a look at the top 5 key indicators in a contact center enumerated by Olena:
1. First Call Resolution or First Contact Resolution (FCR)
Recently, we created an online poll, and more than 13 hundred people voted. FCR has been chosen by 81% of voters as the most influential on customer satisfaction.
2. Percentage of blocked calls
That's the number of calls when the client listens to a busy tone.
3. Average Time in Queue (ATQ)
Which is the average amount of time that a customer spends waiting in line before being served.
4. After-call work time (ACW)
When the agent finalizes call results, if necessary - sends out e-mails, registers cases, pushes conversation summary, etc.
According to Olena, ACW influences customer satisfaction enough to be in the top 5 metrics determining CX.
The more time agents spend finalizing after-call work; the more pressure is on them. They need to be on the line helping clients, not stuck in the complex systems or processes that do not add to the customer service. An unhappy agent produces lower customer service quality.
5. Service Level
The service level threshold is the number of calls answered within a specific period measured in seconds.
There are at least 5 more metrics to look at, but these mentioned already are the top 5.
Although the most measured metric across many contact centers is still Average Handle Time (AHT), Olena doesn't put in the top 5:
"Average Handle Time would be number 9 or even 10. If a customer's need has been satisfied within the first contact, it is okay to take a more extended effort. The customer would be much happier to solve their problem longer but in one call than to make multiple calls, even if they are short. AHT is more the metric for operational efficiency but not to make customers happier. So AHT is a good indication of operational efficiency, but it cannot be Agent's KPI."
KPIs are just a step in a CX strategy.
Even if a contact center measures all the needed metrics and KPIs, the question arises, what do we need to change to improve customer satisfaction?
"From a contact center perspective, this is a good set. But what is necessary is to have a tech stack that allows measuring these metrics in real-time. Then we can take action timely.
Also, for some metrics, it is possible to take action just from the real-time numbers, but for others, we need to conduct a fast root cause analysis. Actually, to get an idea of what causes deviation from a set number for this KPI." - says Olena.
"Partially, conversation analysis or speech analytics can assist in identifying the deviation, but speech analytics doesn't give a clue about many KPIs and metrics. It would help if you had a great contact center solution. Like, for example, TalkDesk is a powerful system, and I would recommend it.
But speech analytics is an excellent tool to go deeper into root cause analysis for metrics like FCR, AHT, Sales to Service Ratio, Conversion Rate, and Customer Effort Score. Without the overhead and understanding of drivers of contact center metrics behavior."
Do you believe NPS is a good call center KPI?
NPS is a great KPI, but not solely for a call center. The KPI should be spread over the company for every customer-facing team.
How do you define a customer-facing team?
Those who interact with the client directly or through the technology they create. Let's assume there is a team in a bank that develops the mobile application. So people who work on UI/UX are a customer-facing team.
The same are the people who plan the process of approving a loan. They should be responsible for creating a better customer experience at every touch point where the client meets the business.
Speech analytics (SA) and customer experience (CX)
Following a recent report by Accenture about unstructured data analytics where they say that 80% of the information in enterprises is unstructured. It is also applicable for voice and chats data stored in contact centers.
When we are talking about the global silo of information in the companies, contact centers are in a great position because they can convert unstructured data of conversation recordings into insights more cheaply than other business units.
Why does it happen, and how exactly the insights from conversations can be used to drive positive change in CX?
We already mentioned the First Call Resolution as the number one driving customer satisfaction in the contact center. To analyze what causes repetitive contacts, Speech Analytics can be used.
First, to track all the conversations where customers explicitly say they have already contacted the company. We need to know the topic of the conversation and better if this topic classification also comes from speech analytics and not from the dropdown menu that the agent makes a selection from.
Olena highlights that it's crucial that the topic should be identified by SA, since it is great if the list of conversation topics is always up to date and created as clearly as possible so that agent doesn't experience a hard time choosing.
In many cases, SA is a more precise way to classify interactions by topics, and it can be done in a more detailed way than the dropdown menu without overwhelming agents.
So how speech analytics helps to find out the root cause of repetitive calls?
We analyze all the conversations that mention not the first contact with the company - this is done automatically in SA where the keywords system can identify if customer contacting not the first time.
Then these conversations are divided by topics, and we need to check all the other metadata available in a visual form:
- most frequent keywords
- time of those conversations
- silence period in a call (it does not only hold time, but if an agent doesn't put the client on hold but still a long silence present)
- agent or touchpoint connected with the repetitive contact to find possible big chunks of causes
All of the mentioned above data is identified by speech analytics in a few minutes.
What can we get out of this data?
When we select a meaningful batch of such conversations, we can dig in by manually skimming the conversation's main parts.
I suppose we discovered that all 30 repetitive calls have the same issue. We can say with a high probability that the rest of the calls within the batch will belong to the same problem.
Usually to check the point of interest in every conversation take seconds. In 5 minutes, we can get an impression from approximately 30 conversations.
How is it possible to get through the conversations so fast?
For many, it may seem unrealistic: 30 conversations in 5 minutes - it is 10 seconds per conversation. But as Olena explains, with the proper speech analytics setup, it takes one click to get to the needed conversation. Then you see the full transcript of the conversation and see all the necessary marks/ tags right in the text. As a result, to scroll and find a place you look for takes a few seconds. To read a line or two to get the idea - a few seconds more. Then move to the next conversation.
How long does it take to set up speech analytics to perform such a complex analysis fast?
There isn't a universal answer; all speech analytics systems are different. Speaking of the Ender Turing system, it may take a week or two. But ET is a very flexible system.
At the same time, it has a great user interface, and what's very notable: ET doesn't require engineers for the design setup.
But there are a lot of systems on the market that have existed for 15 years and didn't change a lot. For such systems, you definitely need a dedicated team to work with, and it may take up to several months for the setup, plus you regularly make changes.
As I read earlier, many companies report that they didn't get positive ROI on speech analytics implementation. How would you comment on it?
Olena clarifies that many failed projects in this area are mainly related to the old and complex speech analytics systems that need a dedicated team to work full-time on continuous changes in speech analytics setup. And in most cases, a company that buys this solution pays for the setup and consulting hours.
In the end, it appears insufficient to complete the proper setup, and the project stops where it is. No internal expertise in the company was left on how to develop the process of speech analytics and adjust it for changing business conditions.
In short, there're two main reasons why SA doesn't bring a positive ROI:
1. Old systems are complex and require a dedicated team to work with them.
2. Companies are left without internal expertise in adjusting systems for changing business conditions.
One more reason: if speech analytics comes as a part of the contact center solution, usually, it is treated by the vendor as an add-on, not the main product, so its development is not a priority. When a company needs new features or improvements, it can take years. Or to pay a lot for customization. So then, additional regular payments drive ROI to negative numbers.
We had feedback from a big governmental call center that used to have a big-name speech analytics solution. When COVID appeared, and they had to track all mentions of COVID in conversations, it took 4 months to add this new word in the system to be recognized properly.
How to make speech analytics implementation successful and get a positive Return on Investment?
Olena mentions 4 main requirements:
1. It should be really easy to navigate for all users. If the system allows different departments to get value out of it, ensure everyone gets a whole idea of how they will use it in a few hours. That's about a straightforward User Interface and navigation.
2. It integrates seamlessly into the infrastructure and current processes in the company. It has to be connected to call center solutions, CRM systems, and other systems if the process requires it. Standalone speech analytics doesn't work - it is a tool for the data science department and has nothing to do with the fast business results or value.
3. High speech recognition accuracy, 80% of words recognized correctly, is a gold standard. Also, the system should be able to adapt different accents - ideally in a week's term, not months.
4. Avoid the systems that need IT team involvement in everyday use. It is about new words/terms to add to the system, new tags, analytics dashboards, etc. So if setting up FCR analytics requires getting engineers involved - no go. If seeing the dashboard on proper customer verification analytics needs some hands out of the contact center - no go.
Speech Analytics is something the company plans to use for years. It's not a tool to take in, make necessary analyses, and forget. It's an integral part of Quality Assurance, where it cuts time for QA processes, Training, and Coaching.
Also, a huge time saver, Compliance, Marketing, Operations, and Customer Experience. It's a big problem when somebody underestimates how fast everything changes, and the external team's need to adjust speech analytics settings will be a blocker.
The ecosystem of processes, technology, and mindset to boost Customer Experience
At the beginning of our conversation, we explained what customer experience is and what I wanted to repeat. It is an experience that customer gets at every touchpoint of their customer journey related to a company. So it means that we have to know what happens at every step.
Recently I had a conversation with a Head of Contact Center in a Fintech company from Singapore. What he said - Contact Center itself doesn't produce problems or complexity. It is the point where a customer gets when other paths don't work out for them. I couldn't complete a purchase through self-service, the product or service didn't work as it should, and something happened outside a contact center, so I had to reach out.
The main idea is that Contact Center often sits on data on why customers get in trouble at some point. This data is unstructured. Big companies have separate business analysis departments or data science departments that work with this data and provide insights to the relevant business units. First of all, not every company has it. Secondly, with current AI and Machine Learning progress, there is less and less need to have an additional layer between business and data.
Proper systems set up exclude a need for data scientists or a lot of manual work to analyze conversations content and provide exact business ideas about actions required.
What is the technology stack you would name as the most efficient one for contact centers and CX boost?
Modern and powerful Contact Center Solution, then Workforce Management, Knowledge Management, Quality Assurance, CRM, and Speech Analytics seamlessly integrated.
Since Ender Turing is the fastest-to-value speech analytics, could we talk a little about it? What industries or businesses does it suit the best?
We have great happy costumes among internal contact centers of
- Public institutions
- Financial Services
What about the company's economic effect of AI solution implementation? What is a profit? How do we calculate it?
It is a many-parameters formula where we take: time savings in QA processes and training, coaching, then improvements in Customer Satisfaction Score. For Sales departments, this also in conversion rates, the sales-to-service ratio in mixed contact centers, and others.
To make the C-level of big corporation believe that CX boost product can change their Contact Center performance, they would require some trial period. Does Ender Turing provide such an option?
We provide the client with a demo tour of the solution. If a client is interested, we are going into a pilot project. The main goal is to show the company how easy and user-friendly modern speech analytics solution is and what opportunities it gives Contact Center agents. Usually, the responsible person for CX can see the results of our cooperation in 3-4 weeks.
What would you recommend to the audience of our webinar?
If you feel there is an area of improvement in CX or Call Center operations in the company you work for, do not hesitate to ask for a demo tour of Ender Turing solution or any other available on the market. Get familiar with this technology. It's good to have additional information and be aware of existing tools. And it helps with a career.
Just in case you want to check the recording of the webinar, check it here.