Podcast | Designing advanced solutions for connected experiences – by the numbers, not science fictions.

Jean
Shin

Director, Strategy & Content | Podcast Host of Mobile Interactions Now

23min Podcast
Episode 38 Guest: Anders Weijnitz, Valtech, Director of Connected Technologies

In this Part 2 of our conversation with Anders Weijnitz, Valtech’s Director of Connected Technologies, we delve deeper into recent developments and know-hows in designing solutions for connected experiences, from how to think about the enterprise complexity and operating costs to integrating emerging engagement tools.

Podcast transcript

Jean:

Anders, welcome to the show. In our previous episode, we had so much fun talking about the broad concept of connected experience, but now I would love us to delve deeper into the backend side of it and get even deeper into where we are at in terms of what's already available to impact business outcomes. So with that, I want to pick up from the previous episode, where we were talking about the importance of data, needing the right set of. So when you work with your clients, how do you usually advise them to think about connected experience, the overall approach before choosing which technology, which use case to start with? Is there more of a helpful guide in terms of how to think about it in general terms?


Anders:

Yes. First of all, thanks for having me back. I would say there are general recommendations. Modern technology, whatever that is, enables us to do many modern and amazing things, but for a customer coming in new to this field, it becomes a little bit odd mix of expecting science fiction because they know there are things out there that are amazing, but missing the really advanced stuff that is around us everywhere, like in every mobile phone, I mentioned in the previous episode, fingerprint sensors, there is more or less a whole weather station, fingerprint sensor, camera filming device.

So it has a lot of capabilities, but the customers are often so used to that that they don't see, that's your enablers. They are maybe in another state of mind where they think, "Oh." They saw some showcase from a much bigger firm or like Google or they saw walking robots, whatever, and it's in the end and more with connected technologies than anything, it needs to be grounded in two things. One is the customer experience. Don't try to wow the customer, don't try to surprise the customer. It needs to be not stalking the customer. So don't use technology for effect. Use it to support a business case and always start with the business case. What's the ideal, most comfortable, best customer experience and the journey to support that and find out how connected technologies, connected experience can help enable that.


Jean:

Is there a more methodical way for a company looking at their customer journey to identify the most opportune moment to help their customers?


Anders:

The method we use is through a systematic, what we call a foundation or discovery phase, in Valtech, where we work with the customer across the organization to identify, first of all, the demographics. Does process and organization align across the board with the intended services that they want to use? The technology discussion of course is happening in parallel. Connected experience starts and ends with organization process alignment and the customer focus. You cannot just look at the KPIs. Classic is that different departments have different KPIs for their touch points with a customer, and if they're not aligned to, work towards one goal of helping the customer and customer satisfaction, no amount of technology is going to help that.


Jean:

It’s such a desirable state where the alignment is already there, and you just have to move some of the pieces of internal processes or how the workflow goes to deliver that desired customer experience. I actually have a hard time finding a solid example where this goes without huge technological plumbing work, because any large organization, they have this whole layers and layers of other systems that they've been using. I was intrigued by what Microsoft said…because all businesses use emails and they’re thinking of using it not only for data mining but in terms of process mining in a way…mining the email requests to figure out the steps required. Actually using machine learning and finding out, basically, charting the flow… say, people are usually asking about this before they get to the next one and so on. Thinking about the data and systems required to connect the process, I'm like, "Oh, the amount of integration that needs to happen in background to make that, it’s not as straightforward as it may sound." Are we actually making progress in terms of making that kind of transformation work without 20 year rollout?


Anders:

The thing with complexity is that it remains constant. So to me, it's a little bit like energy. You cannot destroy energy, you can redirect it in and see it in different forms. To me, enterprise complexity has similar properties. We can only try to direct it or deal with it using different techniques and technologies, and some will be better fit than others. Microservices and Cloud come to mind where you can get a certain level of enterprise agility, or if you want to call it that, by switching over from running the systems to subscribing to a service for database or for Message Plus, and that gives you at least a reduced amount of investment in system knowledge and system... The expertise that you have to build up and you can focus more on your building the application part, which is your business logic, and you can reduce the amount of effort that you would have to put into the plumbing, because the plumbing plumbing is provided as a Cloud service.

It doesn't mean it's a magic wand or anything, but that's something where at least my experience, the promise of the Cloud and part of these microservices actually delivers. There's so much hype around microservices and what it can do, and if you stare too much at it, of course it's not a magic wand, and if you have a lot of microservices, you have a new problem. You need to coordinate and orchestrate your microservices, which you didn't have to do it in a monolitical landscape. It's like, "Is it up? Or is it down?" That has some real benefits as well. So you shouldn't just explode your systems in thousands of, or hundreds of services without needing to do so. It needs to have a purpose, of course.

That's, I think where, as a solution architect, working with big customers, we can bring a point of view there and help our customers, help them compose a good solution that's based on the actual need and not just microservices or machine learning, or... It's all there in the Cloud and you can use it, but doesn't mean you should use it all the time and all at once.


Jean:

I can tell you an example of when Uber, before they started implementing some of the microservices for their notifications, each department had notifications going to drivers or customers in different ways. Thinking back at some of the examples whether it's microservices or new communications platforms, anything like that, anything in recent years that really made you change how you think about designing a solution?


Anders:

Well, in watershed moments or... Containerization comes to mind. So Docker…that's one of those things that... A huge enabler, and still has potential. It's not so flashy maybe, but it's really having an impact on everything from ramping up developers and having consistent developer environments to your project promotion path, how you deliver systems, if you can move images between environments, and I hate to say it, but Kubernetes is also going in that direction although there are mixed feelings and there's a little bit of a backlash right now, I feel. But it's the extension of that.

So it's a mini Cloud almost that I can have on my laptop, and it's almost like a meta container, more than just hype. Also machine, we have said machine learning too much in this podcast, but it works, and it enables you to do things that you couldn't do before. You couldn't do it. So your example was a good one, reading emails or just finding out patterns in large amounts of data, this is very difficult, and I would say almost impossible or unfeasible to write a program that does the same thing. We use that as a large customer in Germany for mundane thing, but making sense of reporting data, so this is part of a payment workflow or revenue workflow, and the data quality is not that great.

It has a lot of legacy in it and different sources and we use machine learning, and one of the process steps to match that, let's say you get 10 different statements, and they are actually only relating to three of your customers, but there are spelling errors or variations due to legacy and things. That's a perfect case for machine learning, where you can train it on data and you can show the machine examples of what it should do, and refine it over time. So machine learning, although it's a buzz word, it works.


Jean:

Do you explain this to the business side of your clients, and if you do, what's the secret? How do you do it? Because it's not the easiest topic to make them understand the value of it.


Anders:

Definitely not easy, and often when it comes up, the expectation than reality, machine learning is good at doing one thing, or matching these statements from a data flow, doing specific things where you can train it and have examples to show it the universal thinking machine, or it's not automatically a great recommendation engine. The example explaining it to customers is often by, I would say by example, avoiding theory and hand wavy things and avoiding too much jargon, and just showing cases.


Jean:

I'm just wondering, because when you are dealing with something that could be as abstract as a connected experience and stuff like this, is it more of re-imagining what the experience will be, the flow of it? Is that good enough to get an okay on it and move to implementing it? Or do you guys actually run the benefit analysis in a way?


Anders:

12:40 For larger projects, you have to run the number and it's extra important with Cloud because what does Cloud mean, or [inaudible 00:12:56] from a solution architect perspective? It's the numbers. So previously you could ignore it, not ignore it, of course, but estimation of the project effort was the major part maybe for selling. If you're a solution architect, your task often in pitching a project or selling projects, like, "Okay, how long is it going to take? What are the major risks?" But now that's extended in a Cloud context to, what does it cost to run?

Previously, the customer would have to buy a license and then they had the license. Now, depending on the choices you do for the solution architecture, the services you recommend, they cost as a subscription. So they will cost depending on running consumption. So you don't have to forecast based on data volume or traffic and the way you as a solution architect compost the architecture, because there's never only one, as always, there's not one answer, can have a huge impact on the running cost. So there's much more... A stronger impact of running the numbers, as you say, as part of the convincing the customer that this is the right way to go, and you really have to do your homework there to see, "Okay, what if we go with Lambda and serverless functions vis a vis running it in an always-on Docker container or a container which is running constantly on a VM?” Those are two different cases. One is on the go of per request billing, and one is more or less per hour billing, and that can have a huge impact on-


Jean:

Any trend you are seeing?


Anders:

Seeing the trend that the Cloud providers, of course, in the end, will get all the money. So they are also... So that's a part of the challenge, Jean, when they are changing the pricing, of course, and it's not as easy as saying, "Oh, this always on VM, always running." That will, of course be more expensive in your mind than something that we have just rarely someone is calling this service, and serverless function. No, but if you run a container all the time, you will get the bulk rebate from Kubernetic.

It can be actually more expensive to shut it down, let's say on non-business hours, that's often something that think, "Oh, but okay, then we shut it down at night because this is only ever needed when people are at work," but that can turn out to be more expensive than having it running all the time. So it's not a fixed set menu. They are also tuning their prices as they are launching or upgrading their services. Yeah, that's what I see. The trend, it's a landscape that's finding itself a little bit in terms of pricing. You can make mistakes, but you shouldn't look at the final cent.


Jean:

As a solution architect, is there something that’s emerging that gets you a little excited and try to put it into every proposal that is going out the door?


Anders:

I hate to say it. So of course, machine learning on edge, every proposal, but I wouldn't say that's emerging. That's maybe trending. This sounds like a cliche, but I really fascinated, I follow the explosion in NFTs and digital collectibles. That's not something you can really put in every proposal because it has a certain... But beyond the high, beyond Bitcoin and Ethereum and just looking at the possibilities... Blockchain is tired maybe, but the smart contracts are wired. That's still hot and digital collectibles and the things you can do around that, and for me, the creator economy and this digital collectibles and NFTs smart contracts, it's currently a huge hype around that, and I follow it closely and there is so much happening in that space. I wish I could include it in every proposal.


Jean:

Thou shalt not be fundable. That is-


Anders:

There is also exciting things around that smart contract, and-


Jean:

Are you really... I have to actually pick on this because here's a reason why, and I'm sharing with a little bit of a hesitation. Three years ago now with some developers work at tyntec we did some smart contract demos and I wrote blogs about that, and we created a smart contract that connects the blockchain world with the real world to trigger some kind of an authentication moment. Then it was buried actually, and then the blog piece was collecting dust, and then all of a sudden... a couple weeks ago, we are getting emails about that, and now you’re mentioning that it's getting hot, and I'm like, "Are you serious?"


Anders:

So this is a hyper trend right now, and you could see it's going in a little bit mainstream. It's over-hyped again, of course, but it's going mainstream, and what's my observation around that is look at NBA top shots. So NBA, National Basketball Association in the US, they have launched a website where you can buy sports moments. So basically a digital version of these collector cards that you had for players, collect all the…baseball legends, whatever. They have made a digital version of that, which is you can collect moments. So it's sports moments, small video clips, and they dropped these in packs, and you can buy the packs. You can showcase them on the site as a fan that I have these rare moments, whatever. What's interesting to me and my observation. So those are NFTs, but it says nowhere on the site, so it's a sport site. It's all blockchain technology under the hood, although it doesn't say that, and if you look at the reaction videos on YouTube, where people watch these sports fans, it's not crypto nerds reacting, it's a sports sites for sports fans for people who like collecting.

To me, it's a sign that it's going mainstream and it's found a use, and it's a simple use case translating something that worked in the real world online, because you can through the smart contracts and the non fungible tokens guarantee scarcity. You can create these things and they have sold in the last 40 days, I remember I checked at the beginning of this week, the traded volume, because you can collect them, but you can also sell them to the other fans. So I buy packs. I have duplicates. I have a couple of these moments. I can sell them on on the marketplace. They have traded more than $300 million on that site in the last 40 days and $500 million since it launched. So it's definitely going beyond the crypto community and the nerd community, let's say, and there's interesting aspects in there.


Jean:

And finally enter the funny money territory…


Anders:

Yes. Well, it's not fun... Well, it's dollars. It's actual money, but they made the onboarding really easy. I bought one just to see what's the customer experience like, how can I buy one of these cards, and you know what. I don't have that kind of crypto. They don't use Ethereum. They use another, and they really made it smooth and the players get kickback. So there are a lot of interesting, through the smart contracts, the players themselves that's featured in the moments get a kickback automatically as it's sold, and of course each transaction [crosstalk 00:21:15] contract. You can track the sales and also get shares on the sales back. So the players are incentivized to promote this site on Instagram or whatever they are to drive traffic. So it has a lot of interesting aspects in there. Yeah, I would love to do something around that.


Jean:

Yes. I would expect you to put that in your next proposal, but Anders, I think that you just did it. I think that you are able to take connected experiences into a totally different realm. Now we are talking about connecting the blockchain world with the real world, and I think that was the point of this. It really is about not necessarily what happens in one physical store and then going to a website. It is about wherever it's taking place and connecting all of this. So I think that we can be talking about this for the next two hours, but I think I need to take this as a wrap, and that was absolutely brilliant. Thank you for joining the show.


Anders:

Thank you.