The Value in the Machine
Release Date:
There’s so much talk about artificial intelligence these days. With stories of how it is changing the world on so many fronts, customer experience professionals can’t help but explore its impact on their organization’s CX program. And when we think about another common struggle with CX – linking the work of customer experience back to revenue – a little bit of help from machine learning might offer a solution. Host Pat Gibbons welcomes Nikhil Nadiminti, senior program manager ISG for worldwide customer experience at Lenovo, for a discussion on proving CX’s value through machine learning.
Nikhil Nadiminti
Lenovo
Connect with Nikhil
Highlights
Ask the right questions
“So we might visit our operational data and probably understand how we can do better. But it’s really important to understand or rather request our respondents to explain the feedback in as detailed as possible, you know, in an as detailed manner as they can. So one of the ways we can do that is by framing our questions appropriately. We ask many of our customers and channel partners across the world what are some of their business challenges that they face today, and how can Lenovo help? It’s very prescriptive.”
Intent is Very Important
“…intent is very important. You know, systems are supposed to pick up intent. They’re supposed to pick up the tonality of the text. Some of them might be sarcastic, where they might be stating the exact obvious. So that’s when the system is supposed to determine that it’s actually negative when it, you know, although the system, although the sentence might read something as positive. So there are a lot of variables that go into determining this.”
Transcript
The CX Leader Podcast: "The Value in the Machine": Audio automatically transcribed by Sonix
Download the “The CX Leader Podcast: "The Value in the Machine" audio file directly.
The CX Leader Podcast: "The Value in the Machine": this wav audio file was automatically transcribed by Sonix with the best speech-to-text algorithms. This transcript may contain errors.
Pat:
A problem for many customer experience professionals is analyzing the vast amounts of data that can be collected and then showing the proof that it's actually working.
Nikhil:
I would say CX is in an evolution stage right now. You know, whenever I've asked any leader, can you rate the importance of CX on a scale of 0 to 10, not a single leader would even say nine. I think it's really important to show the impact in a tangible way. That's when you can realize the importance of CX.
Pat:
Let's look at how text analytics via machine learning can help you tell your CX story. On this episode of The CX Leader Podcast.
Announcer:
The CX Leader Podcast is produced by Walker, an experience management firm that helps our clients accelerate their XM success. You can find out more at walkerinfo.com.
Pat:
Hello everyone, I'm Pat Gibbons, today's host of The CX Leader Podcast, and thank you for listening. You know, we like to say it's never been a better time to be a CX leader. And we explore the topics and themes to help leaders like you develop great programs and develop amazing experiences for your customers. There's so much talk today about artificial intelligence, and with the stories about how it's changing the world, customer experience professionals can't help but to explore its impact on their organizations CX programs. And when we think about another common struggle with CX: linking the work of customer experience back to revenue, well, a little bit of help from machine learning might offer a solution. My guest on this episode is Nikhil Nadiminti, Senior Program Manager ISG for Worldwide Customer Experience at Lenovo, a global technology company. Nikhil, welcome to The CX Leader Podcast.
Nikhil:
Thanks for having me back. Really appreciate it.
Pat:
Yeah, well it's great to have you. And I want to start by congratulating you because you were recently honored as a recipient of the 2023 Emerging Leader Award from the CXPA, the Customer Experience Professionals Association. That's a big honor. And actually, we had another award winner on last week. So this is this is fantastic to have such leaders on our program. But you know, there were only five recognized globally. So you must feel pretty good about that right?
Nikhil:
I feel very overwhelmed. Thank. Thank you. Uh, really didn't expect it, but think I should thank my manager for, you know, the initial nomination and encouraging me to kind of apply for it. And it's still sinking in. And day by day, I think I'm kind of realizing the value of this award. And, you know, being invited to this podcast itself was like, you know, one of the indicators. So I'm really thankful.
Pat:
Yeah. I'm curious, do you think there were certain things on your application for the award that that kind of stood out? You know, that may have helped CXPA say, here's somebody that's making a difference.
Nikhil:
I would say CX is so it's in an evolution stage right now. I've been in the world of CX for the last eight and a half, nine years of my career. And during my tenure I've, you know, whenever I've asked any leader, can you rate the importance of CX, on a scale of 0 to 10, ten being the most valuable? Not a single leader would even say nine. Everyone says ten, right. But then how do you actually, you know, show the materiality of it? That's, you know, one of the things which I addressed in my application, I think it's really important to show the impact in a tangible way. And that's when you can realize the importance of CX in any industry, in any organization. So I'd say that's, you know what, how I applied.
Pat:
Well, I can see how that would stand out. I think we've been big supporters of CXPA here at Walker for many, many years, and we love the way that they've really focused on the impact on the business. So so that's great. Well, tell us a little a little bit about yourself and how you got into the CX profession.
Nikhil:
I started my career in sales and even before that I'm an engineer by training. Uh, I realized that, you know, sales isn't really my cup of tea. It requires the tenacity, which I have utmost respect for, unfortunately, you know, wasn't built for it. But then I soon realized that there is a function which enables sales and also provides all the intelligence in the world that sales requires to be successful. It also allows you to learn about every aspect of the business. So, you know, in a very I would also take the liberty of saying, you get to poke your nose into everyone's affairs. When you're in CX, you really get to learn. You get to learn how, you know suppliers work, how you know customers demand their needs. You get to understand how the supply chain system works, how product development works, how tech support, how customer service works. And it's that breadth which really excited me. And that's kind of how I got into CX. I was working for a multinational organization prior to working in Lenovo and had a relatively similar role and learned a lot of, you know, the retail business. And it gave me an opportunity to understand how every single department functions. And that's, you know, why I continued in the world of customer experience.
Pat:
So tell us a little bit about Lenovo and your role. And and just so listeners know, you're actually based in India, I believe in the central southern part of India. Is that correct?
Nikhil:
Yes, that's right. But I'm based in Hyderabad, India. I have a dual role at Lenovo. Before I explain my role, Lenovo, as you know, is a global company. We are present in over 180 plus markets. We sell everything from pocket to cloud solutions for small and large enterprises. My role is dual. I work as a customer experience manager for the Infrastructure Solutions Group in Asia Pacific, supporting all the Asia Pacific countries. And in addition to that, I also work as a program manager for the worldwide CX team, basically working on enhancement opportunities so that we can continue to provide insights and actions better.
Pat:
That's great. And so we mentioned at the top of the show that, you know, there's so much going on with some of the technologies today. And I think you have focused or actually Lenovo really has focused on the use of text analytics and machine learning. And you know, how these processes can really accelerate some of the work in CX. Maybe you can give us kind of a high level view of that, and then we can kind of dig into some of the details to share a little bit about how that's applied and used at Lenovo.
Nikhil:
So we run hundreds of surveys, hundreds of different surveys across the world, and we receive a lot of structured data through the form of ratings, through the form of scores. But we also receive even more unstructured data, like and when I say a lot, I mean thousands and thousands of comments, it becomes humanly impossible to read through each of them and say, this is what the issue is. And that's where text analytics and machine learning comes in. As I mentioned when we were chatting earlier, I'm no expert in this technology, but I can proudly say that the technology, which has been developed by my colleagues and the leadership, I've been able to, you know, use it, understand it, and also proudly facilitate these insights so that whoever is accountable to own actions can do so with rationale, with logic. And, you know, there's a reason why they would take an action.
Pat:
Yeah. So just for clarity, because I always want to be inclusive with all of our listeners. And some people are pretty new to CX. Let's just define some of these, these terms in practical ways. So if you could kind of pretend that you're explaining to a grandmother what text analytics is and what machine learning is, give it a shot, and that will be context for us.
Nikhil:
Absolutely. So again, at a very high level, it allows in categorization of large amounts of data and helps, you know, categorize into smaller categories so that we can have a more focused approach towards, you know, taking actions on each of these issues or any concerns. So imagine that you are running a lemonade stand, right? And in that you you have transacted with a million customers mean you're not supposed to be a lemon stand by now, but you're still a lemonade stand with a million customers.
Pat:
It's a pretty busy lemonade stand, but I'm following and.
Nikhil:
You have a hundred thousand customers who have provided you, you know, different comments, different feedback on how your lemonade stand has been run, like how they value their relationship with your lemonade stand. Now, some of them might say that, you know, the lemonade was too salty or too sour, too sweet. The water and lemon ratio could have been better. Some might comment that the time taken to serve lemonade was longer than anticipated. You know, the guy at the lemonade stand said that he would serve your lemonade within two minutes, but he took five minutes. There could also be concerns that the lemonades, which were, you know, used in December, had a different quality compared to the lemonades that were used in April. Right. Seasonal differences. So imagine that a, you know, a hundred thousand of these customers filled out a survey form and submitted those to you. So you can't sit and read 100,000 of these comments. It's going to take forever and you need to continue running your business also in the meanwhile. Otherwise you'll be stuck in what's that term paralysis by analysis. Right. So that's where right. So that's where text analytics comes in. It reads all these comments. And then it tells you that of the 100,000 comments, 50,000 had to do with the quality of the lemons used or the taste, probably another 20,000 might talk about, you know, the time taken to serve and then 30,000 comments might be about another 40 different topics. So when you read these categories, you understand that your top issue is, you know, the taste followed by your time to serve, followed by many others. So what would you want to address first, Pat?
Pat:
Yeah, would want to address the ones of most concern the taste. And I would look at. Probably the resources that I'd have to be able to correct that.
Nikhil:
Precisely. So that's exactly what we do. We determine what the top issues are, and we also understand the impact of solving those top issues. Like there's always this classic Pareto example, right? If you are to fix 20% of the issues, they solve 80% of your problems. So in this case, if you are able to solve taste related concerns, you're addressing 50% of your feedback, right? So if we were to equate this to Lenovo, what we do is define a taxonomy we go through our experience, we define certain rules and certain categories so that the machine learning algorithms, they can they can learn it. They can understand what these categories are. And then when we feed in vast amounts of unstructured data, they can start categorizing them. And they use this method called numerical vectors. Like if I were to write you have late delivery. It's not like the machine reads, you have late delivery. It actually reads these keywords, converts them into numerical vectors so that it can categorize it. And then the system once again throws back the term delivery to you. Now it's not just enough if you get a very high level categorization, you are also required to get subcategories. So our algorithms share with us what the sub themes are as well, so that we can understand what they are. We can also filter them out by various countries, by various geographies, by various product types, and accordingly take a decision.
Pat:
Yeah. Okay. Well, I think when you got into a couple of those complex terms, you might have lost her, but the lemonade stand I think was the was the best approach. So, you know, I think and you kind of touched on this, it's you know, this is not as easy as setting a system up, pressing a button and boom, all your results. It takes work to… It's a bit of an art to put together the basis for this, so that it is categorizing things in the right way, with the right sentiment, so that it is actually something you can put to use. Correct?
Nikhil:
Absolutely. In fact, intent is very important. You know, systems are supposed to pick up intent. They're supposed to pick up the tonality of the text. Some of them might be sarcastic, where they might be stating the exact obvious. So that's when the system is supposed to determine that it's actually negative when it, you know, although the system, although the sentence might read something as positive. So there are a lot of variables that go into determining this.
Pat:
Right. And it's just continued to get more and more sophisticated and more and more effective over time. I think the various systems that are out there, it's really made a difference in CX and I know I've seen a lot of statistics and and it sounds like Lenovo is a perfect example of yes, we get a lot of structured data, but the volume of unstructured data is massive, and the ability to put it to use is a big step within our industry.
Pat:
So as you gather all that, there's analysis that takes place and obviously distribution to different parts of the organization. Maybe you could explain a little bit of how how that works and how you work with these other groups that are going to be ultimately putting that information to use.
Nikhil:
So let's move from a lemonade stand to a relatively more successful lemonade stand, which now has multiple employees and not just one person taking care of, you know, everything.
Pat:
Have we expanded into franchises yet, or is just still one location?
Nikhil:
Yeah, let's let's keep it at one location. Prime location. Right. And now you have one person who takes care of the preparation of the lemonade. One person takes care of the procurement of the lemons. Another person takes care of the procurement of the salt and sugar. Another person takes care of procurement of water. There's a person who takes your order, someone else prepares it, and someone else delivers it to you. So it's kind of a process driven very efficiently run lemonade stand, right?
Pat:
I have quite an operation going.
Nikhil:
Right. Exactly. So, so now, if we were to take the earlier example where we thought, you know, we saw 50,000 comments around taste and quality. We definitely not want to forward that to the guy who's taking the orders or to the guy who's, you know, fulfilling the orders. But we would want to pass that information to the guy who's procuring and doing the quality testing of these lemons and the sugar and the salt and the water or the soda. Right. So it's all about identifying who the correct recipient of these insights is. You need to share insights which are. As detailed as possible. One of the challenges which we face is sometimes a respondent might not have sufficient time and they just say, oh, you have a quality issue. But then at Lenovo, we sell so many products and so many services, we need to understand if it's a quality issue with the product quality issue with the service, the quality of the service delivery, or if it's, you know, something else altogether. So before sharing such insights, it's really important that we get as detailed as possible. So of these 50,000 comments, if I've got 49,500, that's just a quality issue. And forward it to the, you know, person who's in charge. They say, yeah, I understand. I'll take a look at it myself. But what about quality is the issue.
Pat:
Right.
Nikhil:
Right. So we might visit our operational data and probably understand how we can do better. But it's really important to understand or rather request our respondents to explain the feedback in as detailed as possible, you know, in an as detailed manner as they can. So one of the ways we can do that is by framing our questions appropriately. We ask many of our customers and channel partners across the world what are some of their business challenges that they face today, and how can Lenovo help? It's very prescriptive. You see, they they say that these are certain challenges and this is how you can help me. And when we start categorizing that text, it becomes easier to hand over to the accountable person who can actually impact and take action towards that change.
Pat:
So in for those people that take action on the change, how has it been working with them? So if we find a quality issue that is specifically about the taste or the freshness of the lemons in my lemonade stand, and I take that information and send it over, I don't just, you know, electronically send it and say, good luck. And what happens if my if that person says, I think it's fine. I don't know what they're talking about. There's I guess I'm getting at kind of the cultural issue of being able to have people take impact on the insights that you're delivering. How do you go about that at Lenovo?
Nikhil:
I would say I'm very fortunate that we have leadership who believe in CX. All our executives believe in CX. The reason they do that is because our CX leadership… our CXO initially sought endorsement from the leadership, and it becomes easier to have executives champion your cause. And when they do that, and they encourage their teams to work on customer insights and make it part of their DNA, it becomes the heartbeat of the organization. And that's precisely what we've done. It's very easy to criticize. Right? That's in this particular example that's exactly what I'm saying. I'm saying that your quality is bad because, you know, these many customers said so. And someone who's really putting in efforts to ensure that the quality is good might, you know, very easily get defensive. But then because is part of our culture, we understand, you know, that the voice of the customer is imperative to driving change towards better customer retention, towards better revenue. All our leadership takes that, you know, feedback. And then we understand that there are certain actions and actions take time. And we, you know, back those actions over time. We take those actions. And once that's done, we also communicate that back to our customers through various types of communication. We let them know that this is what you said. This was the concern. We've fixed it. Why don't you come try our lemonade again?
Nikhil:
And yes… And we invite them through various avenues and then, you know, we we let them we show to them that, you know, the change has been made for the better. And then when they try our lemonade, they're happy customers and they refer our lemonade stand to, you know, ten other people.
Pat:
Yeah. I do love the whole lemonade example. So as you do this, obviously when you fix the problem, you want to make sure that it's fixed. How do you measure the impact of those changes that that you're making? And maybe you can weave in an example of something you've done at Lenovo that, you know, where it kind of went through the process, the change was made, and there was a measurable impact that you can point to.
Nikhil:
Yeah, absolutely. I can explain how we presented text analytics or how we've been presenting text analytics for the last few years. Once we categorize the comments and we understand what the issues are, we plot them up against the NPS scale, the Net Promoter score. And for those who new to NPS, it's the likelihood to recommend your business to your peers and your colleagues. It's an 11 point scale from 0 to 10, and the score is statistically between minus one to plus 100. So what we do is take a quadrant okay. Let's imagine a quadrant where the horizontal axis is the net Promoter score scale. And on the vertical axis we have sentiment. So like I said, through text analytics, we can not only categorize comments into various categories or themes, but we can also derive the sentiment, whether it's strongly positive, whether it's positive, whether it's neutral negative or strongly negative. So based on the superlatives used, based on the adjectives used, we, you know, are able to categorize the various sentiments. So far I've spoken about categories, sentiments and NPS. Right. So on the horizontal scale you have NPS. On the vertical scale you have sentiment. So… that there might be a certain theme like probably supply, which those comments might have a high NPS scale, a high NPS score, and also a positive sentiment. So if you've got a strong NPS score and a positive sentiment, it might be in the top right quadrant. There might be another theme which might have negative sentiment and a low NPS. So that might be in the lower left quadrant. So it's really important to move from the lower left to the top right. That's your dream zone. That's where you want all your themes to lie.
Nikhil:
You want your customers to be happy about your quality, your delivery, your account management, your relationship, your website experience, your purchase experience, your code experience, everything you know you want to keep it in that top right quadrant. So to answer your question on how we track it, when we present this information, we show which theme is in which quadrant. And if it's a significant amount, we start working with that department head. If it's a supply issue, we start working with the head of supply chain. If it's a quality issue, we start working with the head of quality. And keep in mind that we can't keep it at a very high level. Like if I'm showing something at an Asia Pacific level that doesn't apply to all the countries in Asia Pacific. Some issues might be more in India, some might be more in Japan. Most of our customers might be very happy about some themes and just have one pet peeve, right? Or it might be the other way around. So it's really important to deep dive to a level where you can start impacting change as early as possible. Some of them might be at a a people related issue. Some might be at a… But we might have to do with technology solutions. Some might have to do with a simple process. So that's how we categorize them and start sharing those insights with those respective leaders. And we have multiple reviews. And over time, as these actions are taken, we start publishing those actions. We start showing the goodness of how most of these teams are moving into that top right quadrant. That's how we track sentiment and themes.
Pat:
Yeah. No, it's a kind of a beautifully simple framework that seems to work well.
Nikhil:
It does. And when you also look at the the goodness of it, you can tie it to the revenue. You can see how. Uh, accounts are organizations you serve complaint about certain issues, maybe two quarters or three quarters ago. How much you know, they had complained. And because of the fix, what's the revenue you are able to generate from them? So there's definitely that aspect as well.
Pat:
So we've come to that time in the podcast where we ask you to provide take home value for our listeners. So, Nikhil, give us your best tip that CX leaders can take away and put to use in their organizations.
Nikhil:
One thing that comes to mind is and this was something which came up a couple of years ago when I was having a discussion with my manager. There's a study conducted by the London School of Economics, and that study revealed, overall, that a 7% increase in NPS correlates with 1% increase in revenue overall. Let me repeat that. A 7% increase in the Net Promoter score correlates with 1% increase in revenue overall. What this statement reveals is that customer experience should align with the strategic objective of any business. You should as a CX function understand what your leadership, what your business objective is from a from a strategic point of view. And you should do your best to align to that. And that's where you receive endorsement. That's where it would be easier to work towards gathering customer insights and move those insights towards taking actions.
Pat:
It's very wise advice, and it really is interesting to hear both the complexity, but all the great stuff that that you're doing at at Lenovo, if someone would want to continue the conversation, would it be okay if they reach out to you on LinkedIn? I noticed you're on LinkedIn. Is that all right?
Nikhil:
Absolutely. Yeah. For sure.
Pat:
Nikhil Nadiminty is a senior program manager ISG for the Worldwide Customer Experience Program at Lenovo. Nikhil, thanks so much for being on The CX Leader Podcast.
Nikhil:
Thank you Pat. Appreciate it.
Pat:
And if you want to talk about anything you heard on this podcast or how Walker can help with your businesses customer experience program, feel free to email us at podcast@walkerinfo.com. And remember, give The CX Leader Podcast a rating through your podcast service and give us a review, because your feedback will help us improve the show and deliver the best possible value to you, our listeners. And check out our website cxleaderpodcast.com. You can follow the show and you will find all of our 292 previous episodes. You'll find podcast series. You'll find a link to our blog, which we update regularly as well as contact information so you can let us know how we're doing. The CX Leader Podcast is a production of Walker. We're an experience management firm that helps companies accelerate their XM success. You can read more about us at walkerinfo.com. Thank you for listening and remember, it's a great time to be a CX leader and we'll see you next time.
Sonix is the world’s most advanced automated transcription, translation, and subtitling platform. Fast, accurate, and affordable.
Automatically convert your wav files to text (txt file), Microsoft Word (docx file), and SubRip Subtitle (srt file) in minutes.
Sonix has many features that you’d love including generate automated summaries powered by AI, automated translation, transcribe multiple languages, world-class support, and easily transcribe your Zoom meetings. Try Sonix for free today.
Tags: AI sentiment Nikhil Nadiminti Lenovo machine learning revenue Pat Gibbons