People stopped reporting incidents, and the app sped up by 50%. We point to the exact piece of code that doesn't work
A customer, such as a bank, should have a short-term and long-term goal and choose the appropriate technology solution that aligns with their business needs. "The goal is the most important; without it, we cannot build any technology or architecture well. So I try to stick to the fact that we don't have to do everything at any cost. We have to deliver projects in a way that makes sense," says architect and delivery manager at Adastra Jan Karban in the Adastra podcast.
Ivana Karhanová: He says that technology comes second. The goal always comes first. Without it, no technology and no architecture can be built well. So he can talk the customer out of his original technological wish if it is not in line with his business need. And at the same time, he'll figure it out so that the whole architecture runs like a well-oiled machine. Jan Karban, architect, and delivery manager at Adastra. Hi.
Jan Karban: Hi.
Ivana Karhanová: What was the last requirement you talked to the customer about?
Jan Karban: The last requirement was to use streaming technology with a progression from technology to full-text search, where I talked the technology out of using it, but first, we have to say what we want to use it for.
Ivana Karhanová: And the customer didn't know what they wanted to use it for?
Jan Karban: It usually happens that they don't know what to use it for. He knows that he probably needs full text. He knows that he probably wants to use streaming because that's a modern technology used for many things. It's being used for many things correctly, but without a real use case to do a technological proof-of-concept in corporations, we talk about and don't recommend it.
Ivana Karhanová: You work for the biggest corporations in the country. How is it possible that these customers do not primarily address the business goal but address the use of a specific technology without foundation?
Jan Karban: I don't know why this happens. Maybe it's for a lot of different reasons. One reason might be that they are trying to analyze as much as possible on those projects. They think they can see what technology will be needed. They think that the business objective will be defined later, and they will be ready to meet those business objectives quickly. We know from experience that this is never the case.
Ivana Karhanová: Do you think the lack of experienced people in the market, for example, in large corporate customers, could be behind this?
Jan Karban: Generally, in the market, this premise is valid. Given the development of the IT industry and the boom that has been going on for the last couple of decades, the number of quality people is generally small. It also stems from the fact that it is necessary for non-technology industries that these people have a little bit of overlap with technology ones because otherwise, they cannot define the requirements and don't know what is possible.
Ivana Karhanová: Aren't you considered the bad guy with the customers when you talk them out of their original wishes?
Jan Karban: I am always the bad guy. I try to stick to the fact that we don't have to do everything at any cost. We have to deliver projects in a way that makes sense. And that's exactly why I start these discussions, and if the customer insists that they want it, it's good not to participate if we don't believe it.
Ivana Karhanová: So, what should be the right course of action from the customer's point of view?
Jan Karban: The customer should know what they want in the short term and where they are going in the long term. For example, when he opens a new bank, he should know if he is targeting the retail portfolio. He should know on what horizon they are planning a corporate portfolio. He should know what products they want to focus primarily on and how big they want to get. He should know if they're doing this bank for long-term development or if they want to build it for five or six years and then sell it to somebody as an investment. This would all be good to know at the beginning because it all goes into what needs to be done afterward.
Ivana Karhanová: You said they should know how big they want to be. Who knows at the beginning what the development will be in the next five, six, maybe ten years?
Jan Karban: Well, they don't know like anybody, but they should know if they aim to be the biggest. For example, a retail bank for consumer lending must take into account and behave accordingly.
Ivana Karhanová: Maybe we're putting demands on many scalabilities in the future, on a lot of expansion.
Jan Karban: Scalability and scaling up will probably always come. It's always there, but it's good to say when the architecture is being built that I don't want to sell 50 types of products, but two, three, or four will be my guideline. Then the architecture looks different, and the solution results may be better for the customer.
Ivana Karhanová: What are the current trends? What do customers want from you?
Jan Karban: One of the interesting things happening right now from a corporate perspective is group data warehousing or reporting solutions. It's something that's been going on for the last ten years. We've hardly been involved in it because it's been shown over the last ten years that the architectural approach is not right, and the big, centralized solutions - although I probably shouldn't say this - are being phased out and it's being shown that you can't do it that way.
Ivana Karhanová: That is so we can imagine something concrete under this. Is it that, for example, a multinational banking mother will say they will have one central data warehouse?
Jan Karban: Yes, that is a project that happens quite often. It costs hundreds of millions of euros, and there is no expected result. Gradually the architecture has to be adjusted because centralization is not possible to the extent and in the detail that it was planned or intended at the beginning, so it is undeliverable. It is based on many things, for example, the legal conditions of the countries concerned. Each state has a different level that is intertwined with that architecture, and then you can't connect it even at the level of the individual, detailed data structures.
Ivana Karhanová: Does that mean that each country stores or processes different data about clients in a different form - if we are talking about banks, for example?
Jan Karban: It's always the same to a certain level, but the devil is in the detail. Suppose you want to unify the address, for example. In that case, when we talk about client data across Europe, we find that it is not easy because the structure of addresses in different countries is very different. And it is impossible to have a single view of it that allows full automation because it always turns out that there are intractable problems. And I can imagine that just for addresses and simple data like that, you could still solve it. Still, from the point of view of the banking world, for example, enforcement, legal conditions for marketing, and other things that affect it very much, I think those cannot be unified.
Ivana Karhanová: So, ten years ago, the mothers decided that they wanted unified or centralized data warehouses, and now they are decentralizing again?
Jan Karban: So far, the banks are still holding on to the projects and trying to make the most of what has been invested, but as a result, it turns out that the reporting at the local and national levels of those banks is much more detailed and has to be also from a regulatory perspective to the national banks and the local registries and so on. It has to reflect the legal and regulatory framework of the state, which is always more detailed than group one. If we talk in detail, everything can be stored somehow, but it is not maintainable.
Ivana Karhanová: There would be data lying around.
Jan Karban: Data is never lying fallow, but it's hard to automatically mine it, and it doesn't meet what I expected.
Ivana Karhanová: What do you think will happen to those projects in the end - the really big ones that don't come to fruition?
Jan Karban: Actually, it's not so important what happens. It's more a matter of the company's policy. Either they change, sign off, or declare that they delivered what they wanted to - for some parts, it has already happened. But these projects are very expensive, and even from a market point of view, it is impossible to say that we will throw the whole project away and write it off, even though the benefits and inefficiencies are great in the end.
Ivana Karhanová: What projects are you currently working on with your team?
Jan Karban: I am dealing with the merger of two banks. I have completed one of them and am working on the second one. It means mergers in regulatory reporting and mandatories that have to happen in terms of the legal framework of the mergers or the ones that are going on.
Ivana Karhanová:: It means that the customer needs to consolidate these things. What's the next course of action from your side, then?
Jan Karban: You need to define what is needed, as we discussed at the beginning. There are many things to consider, starting with how long the Czech National Bank has been given an exemption from regulatory reporting and how the bank wants to approach it. That is, whether the merger will result in one entity or two entities, whether it will be a group. You just need to know all that and, based on that, define what needs to be done, plan it, find out how much it will cost, and then deliver it accordingly.
Ivana Karhanová: You know how to work with data not only in the Czech Republic. You have also delivered a data warehouse remotely.
Jan Karban: That was one of the interesting projects. We delivered for a bank in the Balkans where their national bank repeated an audit finding. So we did some mapping there on how to resolve it. And then, based on that, we delivered a data warehouse without ever being there because of the covid.
Ivana Karhanová: You mentioned the national bank finding. What was it based on? Was there erroneous data?
Jan Karban: Data inconsistency.
Ivana Karhanová: What did you do with it then? How do you turn inconsistency into consistency?
Jan Karban: The way you do that is you go to the customer, do an architectural health check of their architecture, and find out the biggest problem. So that everything that the customer has can be used going forward so that we don't come to the customer and tell them: You have to do this all over again. That's not the goal. So those components in the architecture, if they're even a bit good, should bite the bullet to save the fact that it's pointless to re-develop those things. And say where the problems are and propose solutions. And the solutions tend to be long-term. That is, we should define a target architecture, and then we should define a first transition phase into that architecture to solve the most pressing problem of that bank.
Ivana Karhanová: How long will this phase take?
Jan Karban: Because the bank was small, the data warehouse took us between eight and nine months - the delivery. Before that, there was about a two-month solution concept definition phase, which was separate. It happened quite often in banks that when we started doing the solution concept or the roadmap, in the process, when some things were clear, the parallel development started. So it's not a two-month project, but that first delivery might not be years away.
Ivana Karhanová: You said we'll come to the customer and do a health check. You couldn't go anywhere here. What was the difference when you did the whole thing remotely?
Jan Karban: The health check was there at the time before the covid, so we were there, and that's how we knew each other. By the time it came to delivery, everything was fully remote.
Ivana Karhanová: Your whole team, or your work, is very prominent in banking. But you're also currently active in the energy sector. What projects are currently running there?
Jan Karban: At CEZ Prodej, we have delivered or are delivering the integration backbone in CPR so that CEZ Prodej, as far as IT is concerned, is separated from the group so that it is flexible enough to respond to customer requirements. We were delivering an operational data store there. Then the beginning of the data warehouse and the data quality solution, setting up the basic down governance processes for that data quality.
Ivana Karhanová: What is the biggest challenge for you at CEZ Prodej? It's a big company, originally state-owned, so it's going to be quite rigid, I would say.
Jan Karban: The biggest challenge for us is generally not quite the technical stuff. I haven't had any experience in the functioning of the state administration so far, so all the rounds regarding architecture, approval of solutions, and tenders are significantly longer in terms of the Public Procurement Act. It defines the whole process and influences it significantly, so everything is longer and takes more time. So that is a big difference.
Ivana Karhanová: Where do you think big corporations will go next? We've named big projects at the parent level that probably can't quite meet the original expectations, and corporations have had to adjust expectations. Where do you think working with data will go next?
Jan Karban: It's hard to say. It's more likely that more and more worlds will be connected by data, which is already happening. That's a guess. Ten years ago, there was a boom with new banks opening up. We have a phase where banks are merging back together, so we're dealing with these mergers. In general, the retail world will merge more and more with the banking world if the law ever allows it. We already see a lot of concepts for that.
Ivana Karhanová: What do you mean retail and banking - SME or enterprise? Or retail as a completely different segment?
Jan Karban: I mean individuals, where we see a lot of applications in their banking that connect the banking world with the retail world. For example, cards that are cross-bank and things like that. The beginning of the push for these opportunities comes from some directives.
Ivana Karhanová: Banks are opening up their data and making it available to third parties. In the same way, third parties can enrich bank data.
Jan Karban: Yes, and it will be seen more and more. Banks also need to automate more, so we're seeing projects of bank identities and everything else. The more pressure is on the client to not have to go anywhere, to make it as convenient as possible for them to log in, to make it as easy as possible for them to see what they're spending on in banking. There will be a lot of pressure to connect specific shopping carts everywhere.
Ivana Karhanová: Data enrichment and working with third-party data have been discussed for a few years. So, in your opinion, are they finally coming to fruition now?
Jan Karban: I don't know if it's finally happening. It is a continuous process that is going on here, and the question is whether and how the state administration will accelerate it. If the government's promises that the ID card would disappear as plastic, everything would be digitized would come true. That would also have a great benefit and impact on this world. There are many things and incentives that can speed it up or slow it down.
Ivana Karhanová: Says Honza Karban, architect, and delivery manager at Adastra. Thanks for coming into the studio, and see you sometime.
Jan Karban: Thanks.