Privacy Policy

Episode 46 of The Andy Show

Episode 46 of The Andy Show

NICKY PENNYCOOK [00:00:35] Hello and welcome to today’s Andy Show. I’m your host today, Nicky Pennycook, and today is the 24th of June. It’s a gorgeous day and today we’re going to be talking around analytics. And our guest today is Dan Somers, who is the CEO at Warwick Analytics. And let’s say welcome to him now. Hi, Dan.

DAN SOMERS [00:00:59] Thank you Nicky.

NICKY PENNYCOOK [00:01:02] Thank you. How are you doing today?

DAN SOMERS [00:01:06] Yeah, very good. It’s a sunny day where I am as well.

NICKY PENNYCOOK [00:01:11] Lovely. Got a few questions for you. But I think it would be nice just if you introduce yourself and talk to us a little bit about Warwick Analytics.

DAN SOMERS [00:01:23] Yeah, sure. So, I’m the CEO of the company and I help to spin that out the company. From the University of Warwick. A few years ago, my background is I’ve run another business before, so I guess serial entrepreneur is the word. Used there, and before that I was a management consultant, a trained engineer. Warwick Analytics, it’s all around making very difficult data science very easy and quick. Difficult data science, we specialize in analyzing conversations particularly with contact center. To be able to produce actions for people to improve the customer experience.

NICKY PENNYCOOK [00:02:01] So that leaves me very nicely into your first question that I got ahead. So obviously, things have massively changed recently. So how are you seeing customer experience changing? So, from it was before and how it is now?

DAN SOMERS [00:02:17] Yeah, I think we like to see in our everyday lives over the past few years, people’s tolerance for things going wrong has become lower. People becoming more impatient, brands becoming more competitive. And when something goes wrong, the expectations to be fixed very quickly are very high. So, brands are really focusing not just on their product or service. But also, in how that customer feels if or when things go slightly arrived or when the customers try to set things up. The whole point of the customer journey from start to finish is really important because you also have social media where unhappy people, angry people can make a lot of noise very quickly. And the converse is true. we are delighted we can also express themselves and the industry is sort of trying to figure out so far. How do you measure these kinds of things? They’re all measures for it, but there’s a lot of noise around this in this industry. So I think it’s a sort of a ballooning issue that is still quite immature. And there are a lot of people trying to wrestle with it, too. They know it’s linked to profits somehow. But making happier customers is at the core of it, but it’s trying to navigate how happy do you have to make them. As to which ones once focus on what you have to do when there’s a lot of challenges in cross-sector, and of course the organization itself doesn’t usually fall within one person’s remit. It touches everyone in the organization.

NICKY PENNYCOOK [00:03:52] Absolutely. And so obviously, like you say, customers, for instance, are really important to brands. And it has change all over the years because especially now we’ve got the social media. And it’s a lot easier to get messages out there, instantly. Whether that’s good or bad. And what can brands do about this to make sure that everything is as positive as possible and obviously, when things go wrong, they obviously want to make sure that the good stuff is out there. So what can I do about that?

DAN SOMERS [00:04:27] Yeah, so there’s a lot of things that can do and there are other things that some brands do and they are both really around listening to the customer. That’s really, really important. And you can do this in the form of surveys, which are what we will get bombarded with all the time. They do have it merits. They also have their drawbacks. I think people survey what’s now called survey critique, where you’ve just had interaction with the customer, a company. As a customer, you are now being asked, how do you feel? What did you do? How would you rate us? And surprise, surprise. You only get really to people who have something incredibly emotive to say, positive make to actually respond. So you don’t get the full sample and you’re only going to get a handful of people, a few percent generally actually come back. The other things you can do is listen to all of the conversations with all of your customers all of the time. And that actually sounds like a big task. It’s actually made a lot easier these days to the fact that there’s a lot of different data flying around. You can access it. You can marshal that. And clearly, there are lots of technical challenges on the way and data privacy issues, the rest of it. But it’s therefore up for grabs. And the really question is, how do you analyze all of those conversations to get that rich insight and make it actionable rather than sending out surveys and getting at least your finger on some kind of pulse, but not really being able to move the needle in terms of figuring out how to prevent churn? How to delight customers to spend more? and how to find out the things that people like or don’t like about your product? and how to make them that the customer journey as delightful as possible.

NICKY PENNYCOOK [00:06:08] Do you think at the moment as well that these things are probably more important than they ever have been and taking online experience, obviously, people has been stuck at home they’ve had to do order online. And having seen that change in any way or the way that things is becoming more important, maybe not quite so important?

DAN SOMERS [00:06:28] Yeah, it’s a really good question. And we’ve been looking at this in some depth actually. It’s not clear what the answer is, though. It’s where experience is actually enhanced because there’s a sort of empathy during COVID between the normally angry customer and the very tired agent at the other end who’s trying to sort something out and everyone knows they’re at work. So there’s a sort of firm, some kind of two entity, which sometimes mean that’s customers, the more forgiving now than we have been. On the other hand, a lot of chain parts of the customer experience are broken. And that has ramifications. Just high volumes of demands on contact centers, skew toward certain queries. Agents, a lot of the time working from home, working the shifts, not having access to systems, sometimes not IBM to use voice and having to digital media. We’ve seen companies that are speeding up chat, Web chat more and more to try and cope with the demand for the contacts themselves. Trying to move them to a low touch channel, so to speak. And also trying to avoid the occasional dog barking or doorbell ringing. Which is, you know, the sort of status quo. And that comes with its own suite of challenges moving to any medium, particularly one voice interaction is the best medium to solve complex queries. But it’s also the most expensive to serve and also customers. Prefer to deal with things, as you said earlier, as quickly as possible. We’ll see you on their terms and voices. You’ve got to hold for a while and you’ve got to get through security and some other things. You might even be forwarded once or twice for music then, there are other things you probably would rather be doing. Let’s put it that way. Chats it’s not exactly, without its challenges, but a lot of customers find that much easier. Chatbots, sell the dream but, you know, again, that’s a two-edged sword because chatbot are demand for chatbot’s and its performance. Let’s put it this not quietly the manufacturers promises, marketing spiel and the chatbot are not so quite quickly to serve some percentage of all of the types of queries that can be made by customers. And of course, it cause problems itself. So there are definitely some sort of asymmetries and some opportunities in the current situation which weren’t there before. A lot of them are on the hook. Just kick-started the ongoing trend, some of it are very specific COVID. It is really interesting to see which ones will be a lot more people more familiar with video, for example, and they may have been and you know, time will tell to see if some of these things look afterward. But for sure, there are opportunities created for them, for best brands to seize the moment and to differentiate themselves in a very challenging time.

NICKY PENNYCOOK [00:09:52] And so we went to the next question, I’ll go ahead. You kind of cover some of that. And then you mentioned data privacy as well a little bit earlier. Are there any other barriers? And when it comes, this technology?


DAN SOMERS [00:10:08] Yeah, well, ask any CTO and certainly, so many barriers to technology and the answer is, you know, where do I start? Right. You’ve got a whole world of issues and other priority and projects. And it really depends on how your business infrastructure is also set up to take on new technologies, which some companies, if they don’t have the Cloud, fully enabled. And then security is one thing, but they may not have enabled yet Cloud infrastructure to be able to adopt rapidly new technologies. So there are definitely some challenges more for privacy key one, really goes human beings involved, and particularly for A.I. for technology. People might be surprised, but chatbots are run by people. And all the technology that goes into all of this stuff is trained and run by people. Machine Learning, if you’ve heard that phrase, it’s learning from humans and that can be a very laborious task. People assume it will work out the box. And if it doesn’t, because every conversation is specific to the company, every sector is specific. And if you have a generic solution, then you are not matching up the training set with the operational data set. If you follow clift you may not get what you put in an event. You’ve tried to use a chatbot, and you got some hilarious results. You all like everybody else on this planet, because that is really where it becomes glaringly obvious. So there are challenges that if you are in technical and data science in particular to the adoption.

NICKY PENNYCOOK [00:12:05] So obviously, you mentioned A.I. towards the end there and things like chatbot like that. Do you think they’re overcoming these barriers and also, do you think people trust them? Because I imagine some people probably go weary of things like A.I. and they don’t really understand it. What are your thoughts on that?

DAN SOMERS [00:12:26] Yes. So that’s a very good point. I think A.I. works best when you get really know it’s there, or you know it’s there but you kind of don’t care because it’s kind of doing what’s expected “when it’s not to do”. And that doesn’t mean that it’s pretending to be a human. It’s like know that when you’re flirting with a robot, actually, the robot is a robot, but it still needed you and you get on your day in and you just use it for what it is. The challenge is when it doesn’t really meet expectations. And that really kind of flows into the first question, which is how you overcome some of these challenges. So there are technologies appearing which make a lot of these heavy lifting. Let’s call it that way. Hugely easier and not necessary by data scientists. So the technology, for example, which we spun out for more at university, is around putting the power of a data scientist in the hands of somebody who understands business concepts but is not really there to plug in the A.I. per se. And the A.I., if you like behind the scenes, interact with those humans, these are people. There in the contact centers or managing or analysts or business people. The A.I., if you like, asks a human to train it and has a smart feedback loop. So it understands where it’s not sure. And it will try to ask the human to educate it, or to supervise it to use the technical term where help is need because something needs a new signal is detected all because of something ambiguous. Maybe someone said where’s my credit card? And someone else’s says if I need to block my credit card and if it gets confused between that, it might come back to the human side, can you help me out? May not be in real-time. The query that happens may get wrongly channeled but very quickly. So there are a bunch of technologies like this, like ours to not specifically like ours, but trying to address this problem, to massively lower the barriers to entry, to adoption and to scale up and to really unleash the power of the promise of A.I. because the A.I. has been with us for decades. It’s really now that the power of these barriers are being lowered, that they can be applied to these massive datasets. One thing also to bear in mind is there are billions and billions of conversations happening and analyzing all of them at scale in real-time is, with all of those other challenges, something which those a lot of brands and these technologies, these new technologies are there to make it very easy.

NICKY PENNYCOOK [00:15:15] Is there any particular examples of perhaps personal experience that really stands out to you that this worked really well. Like you said, with the A.I. elements sometimes you can just leave it to get on and do its thing. And, yes, anything that really stands out to you thinks as a really great example and show how this can work effectively?

DAN SOMERS [00:15:36] Yeah. So there’s several examples that I can think of, the first example is really if you analyze a conversation for three main models of the topic intense, so what is the call about is about? Is it about credit cards? And exactly what about the credit card itself that the customer is trying to do? Then you got the sentiment. So how the customer felt and how that changed throughout the interaction. Now you’ve also got things like emotions, slightly different sentiment, the emotional intend where customer might express that had to go back before they’ve been on hold for a long time. They began to leave the brand because they’re really annoyed or actually, they’re not the brand or something specific. And then you got other types of signals. You could pick up phone signals and other things that you could imagine having all of that switching sides that can be used to train the agents to do better. It can be used to improve the product and the service that you offer. It can be used to actually in the moment to help automate some of those groups, particularly come in for a chat and email and also can be used to help supervisors chatbots there in A.I. and there’s a lot of fantastic companies doing that. Some of these use cases that we get involved with is actually supervising chatbot. Providing, if you like, a side by side model. We say that it doesn’t mark its own homework. “If you follow and that would be really good because all the chatbot always think that they’re doing really well. Now, that is half battle is finding out where they need some help. So just to put this spring into life, when we get involved, usually a few weeks later, that’s all it takes to get all of these models built because it’s so quick and easy. The success go up by 20% at least, and cost savings, 20, 25% to take out all of the people calling back. All of the unnecessary queries that can be pushed through a chatbot. These are metrics that we would expect. And also, they measure things being reduced. Again, that can be 20% plus category. It’s very typical. And so, this is really the power of putting up solutions. A huge, huge business benefits. It really moves the needle for these companies and allows them to serve the right customer with the right channel for the right enquiry.

NICKY PENNYCOOK [00:18:03] And it sounds like a really important part of the whole customer experience from actually ordering something to receiving that product, to having and use it and if you need support or anything like that. It sounds really interesting. And obviously, things have changed a lot recently, we’re referring to the pandemic that probably no one see that happening 6, 7 months ago. Now that this has happened and perhaps before this happened, what does the future look for you? Look like for you now? And has it changed much or do you think things are going to go relatively back to normal?

DAN SOMERS [00:18:42] Yeah. So it’s a big, huge question. And, you know, all that we have at the moment is this sketchy data on the consumer business. But with the data that we’ve seen and certainly the view that we are forming is that some of these things are here to stay. And so is that because they do practices and they are efficient, they help drive a customer experience. And for example, using chat medium. Updating some of the way that a customer that the brand interacts with its customers. And also, most the expression, when the tide comes out, you can see who’s not wearing shorts and that of course is really there to say when times are tough. It’s the better companies and businesses with the care most about their customers or respond more reactively to them and give them what they need to succeed. Many great companies have come out of the sessions and other problematic economic situations. So, I think you’ll see brands doubling down. A lot of brands that we work with are doubling down on this rather than cutting back on their investment in this area. The doubling, particularly on these use case as it pays for themselves. I mean, if you’re automating people, knocking on your website and send you an email, you know that 3-hour e-mail process is just a little journey in this period. People are putting an automated solution. Using us, using other solutions there. And that’s it to stay, the chatbots so under immense strain. People again, some doubled down and improve those. Those improvements are here to stay. People have moved away from voice or they’ve optimized their voice. It’s only used for the queries that it should be. And again, that’s here to stay now, you know, some of these volumes in the mix, mixes will change up to COVID for sure. But all those efficiencies as we process is under due rest. But, you know, in many situations and even more time, unfortunately, to use that analogy, you know, you have great inventions and innovations and new ways of doing things. And during these times, some brands have really prospered. And I’m not allowed to name them. But this is ensuring that we work with has really doubled down on its customer experience and is making a big thing out of it. And it’s really succeeding where others are struggling. And it will learn all of those learnings after this pandemic. It will be business as usual as a brand with a bigger brand equity and a more loyal customer base.

NICKY PENNYCOOK [00:21:33] Thank you for that Dan. So just to finish up, obviously, customer experience is very important to you. You’re working with it, day in, day out and obviously develop different ways and how to make it better. Personally, to you, you don’t have to name names. Is there anything that’s really stood out to you as absolutely amazing or anything that’s really stood out to you as, oh no, why would you do that type thing? Just interested to see from your past, in your point of view.

DAN SOMERS [00:21:59] That’s a really good question. So, every time I did a survey, I do shake my head. There’s not to say that there’s no room for service. Most of the surveys are on things that really they should know about. So, they should be targeted towards the things that you don’t know about. So, you know, when I get that survey and write things from 1 to 10, you know, in other words makes my eyes roll. But it also makes me smile because it makes me realize that the industry we’re in, there’s a lot to go up to, a lot of opportunity. And I usually tell someone in my sales team to go and call them because there’s potentially a lead. The other thing is really around chatbots as well. You know, it’s difficult. People ask the same question in many different ways. And was inventively come up with new questions to ask. It’s chatbot always on the curve? So again, my personal experiences of a chatbot and sometimes I do where I see that they are not understanding the content and just see to try and break them and unfortunately, I think that that is also human nature. I am reluctant to name the brands. But there’s one very large telco that their chatbot is interesting. And there’s a certain sequence of things that if you ask about the account, it will send it into a real test. But then on the other hand, as I said, some of the other people in the industry are really shiny and you know, when you have spoken to a brand. You know, it comes down to a lot of the time you got through the automation. This is the irony really, is that it’s humans that really make the difference. And you don’t notice humans because they work hard to make your experience great. When something really goes wrong, we really need help. And you talk to that person to the contact center. To try and help you when you really get that white glass service. You remember it, and you remember it with that brand. You know, that person is at the front line and whatever the CEO of that company is doing. However well or badly, they’re running the company. Your relationship with that brand has come down to essentially one conversation with one human being who’s trying to do the best job they can on that day. And if they succeed or even with the can’t succeed but they still make you feel like you matter, then you remember that we’ve all got our experiences of those brands every day, as well as the ones that are not so good.

NICKY PENNYCOOK [00:24:42] Absolutely. I know I’ve had some past issues with some brands recently and largely don’t sort things out. And I’m sure everyone has something that springs to mind when they think of good or bad customer experience. And thank you very much for joining me today again. And that’s been really interesting. It’s been good to learn a bit more around A.I. and the customer experience and hopefully we can catch up soon.

DAN SOMERS [00:25:05] Thank you very much, Nicky.

NICKY PENNYCOOK [00:25:11] That was Dan Somers, he is the CEO at Work Analytics and talking all-around customer experience, how it’s changed and what brands are doing. And also how A.I come into that. And so it’s been another great show today. We’ll be back tomorrow and if you want to be a guest in the show, you can check out our website and get in contact on and or contact us through Facebook, LinkedIn or Twitter and that’s just Disruptive Live. Thank you. See you soon.