London Fintech Podcast
00:16:07 · Nik Storonsky, Tony Clark
[00:16:07] S0: Yeah. I mean, for example, if you look at Amazon, like, I don't know, twenty years ago, they just started from selling books, right, online. And then they later expanded into tons of more. So what is Amazon? [00:16:16] S2: It's like a shop, right? Yes. And interestingly, shows I'm wrong on everything. I thought no, no, Amazon can't expand out of books because they do books. Mean how can they do other things? Now I buy most of my Christmas shopping through them. So from that perspective then you don't seem to find the brand development challenging from a sort of descriptive perspective because actually your customers are asking you for these things. So from a different perspective, on the software side, one day you've built this nice simple little app that does one thing and then you add another thing on the side and you add another thing on the side and another side and then before you know it your nice simple clean software has grown into a huge bloatware with thousands of modules and 58 home pages and all that kind of stuff and it's almost like building an inverted pyramid. You started with something small and you put something big on it. So how do you cope with it from a software perspective about re engineering the development? At some point presumably you go, okay, we built all this based on that, but maybe we should just re engineer the whole thing?
00:17:14 · Nik Storonsky
[00:17:14] S0: Yeah, mean, there is a whole science, you know, how to get it done, we're kind of very rigid that following certain principles. So first of all, everything that we build is based on modular principles, right? So every piece that we're building should be like separate, like separate block, separate modular, separate code, right? And Whenever we want to build something bigger, we simply are not building it from scratch. We're just taking all these small blocks, we're combining them together. And here is a feature, right, that allows us to iterate much faster, build things much faster without any bugs. Same applies in app interface as well. So if you look at the whole app interface, it consists of very, very small, unique blocks, right, of design guidelines. So we apply the same blocks for all the products that we have: credit, premium, insurance, wealth, money transfer, FX. It's all the same, underlying is the same, We're just constructing something different from the same blocks. That's the key.
00:18:13 · Tony Clark
[00:18:13] S2: It's interesting because I was thinking about a million years ago, about thirty five years ago, when I had to do a dissertation at college. My degree and my dissertation was a nightmare. I got completely dropped in the deep end and I ended up working all of Easter which was amazing in those days. We were recording this the day after Easter now. And my dissertation was porting a Modular two compiler from the VAX to the IBM three seventy, which was big main fame at the time. Modular two was the first modular language. It was a development out of Pascal. This is in the early '80s. It's the kind of thing that people wouldn't do these days. I had a printout which was one inch thick. There was not one comment in the whole code, there was not one description and it was just a nightmare to sort of read through it. So I have direct experience myself of what this modular code stuff is like which is in the old days amounted to on my room in college. Over Easter I just couldn't understand this so I just ripped up all the printout into the modules and I put them on the floor and I could see like the 50 modules spread out over the floor so I understand that. Mean the technical challenge there of course is that it makes it easier as you say but you can end up having trouble on the interfaces. So if you've got interfaces, know for the sake of argument defined on 15 parameters or exporting or importing or however it works these days and you've got a whole idea of products. But then as you say, mentioned insurance, one day you introduce insurance which has got different parameters and a different kind of perspective of the world.
00:19:34 · Nik Storonsky, Tony Clark
[00:19:34] S0: I mean, you'll be surprised. I mean, you can still build insurance on existing blocks that we have, right, which we are doing at the moment. Have you seen, you know, for example, WeChat in China? [00:19:44] S2: Yes. WeChat? Yes. I haven't used it myself. I've heard of it. [00:19:47] S0: Well, you know how many services you can have, like Facebook, Twitter, you can also pay with WeChat, you can invest money with WeChat, you can buy yourself insurance with WeChat, right? And that's just huge. But the whole point why everyone is using it, because the interface is so simple, right? So they managed to create interface which effectively covers everything, right? So it is doable, it is possible.
00:19:47 · Nik Storonsky, Tony Clark
[00:19:47] S0: Well, you know how many services you can have, like Facebook, Twitter, you can also pay with WeChat, you can invest money with WeChat, you can buy yourself insurance with WeChat, right? And that's just huge. But the whole point why everyone is using it, because the interface is so simple, right? So they managed to create interface which effectively covers everything, right? So it is doable, it is possible. [00:20:07] S2: So the software problems, surprisingly enough, have been sorted out over the last thirty five years that we're still in the teething stage at the early stage. And I think there's more art than maybe what you're implying and more skill than maybe what you're implying in terms of keeping a nice simple user interface. Just take a simple example, if you go back to the earliest spreadsheets or actually not the earliest word processors the earliest word processors were awful like VI but go back to early versions of Excel it's actually relatively straightforward and then as time goes by and more and more features are added it gets more and more complicated and occasionally Microsoft does sort of rejig the design and you know my children sort of laugh at me and say oh you're using an old fashioned version, it's just simpler to use than their version. So there is quite a lot of art and quite a lot of skill in keeping it nice and usable especially on quite a small real estate which is a phone screen.
00:20:07 · Tony Clark, Nik Storonsky
[00:20:07] S2: So the software problems, surprisingly enough, have been sorted out over the last thirty five years that we're still in the teething stage at the early stage. And I think there's more art than maybe what you're implying and more skill than maybe what you're implying in terms of keeping a nice simple user interface. Just take a simple example, if you go back to the earliest spreadsheets or actually not the earliest word processors the earliest word processors were awful like VI but go back to early versions of Excel it's actually relatively straightforward and then as time goes by and more and more features are added it gets more and more complicated and occasionally Microsoft does sort of rejig the design and you know my children sort of laugh at me and say oh you're using an old fashioned version, it's just simpler to use than their version. So there is quite a lot of art and quite a lot of skill in keeping it nice and usable especially on quite a small real estate which is a phone screen. [00:20:56] S0: We for example have six designers, right, you know, working on this problem, how to make sure we package, you know, all these features, all these products, and still, like, you know, very simple interface. It's a difficult problem, right, but I mean, it's solvable.