Sberbank — Technology Innovation and Investing in Startups

This is the translation to English of my interview with Igor Kostylev @igorkostylev published at here.

Q : How long has the innovation center of Sberbank being working, how many people are there ?
:Today there are15 people in the team, apart from me and my assistant. Most of them were hired as project managers, but in fact each one of them is also a technology leader. They must not only manage the project schedule and resource allocation,but also to understand the subject area– know the various players in the market both small and large companies — to understand their technology. Moreover, their task is to create a technological solution for specific use cases in the bank. So, they have toknow business, technology and be able to manage projects.
Each of my employees is assigned to one area. For example, one of them manages big data  initiatives, which is related to data collection, processing, and visualization. Thisperson works with a lot of companies that are involved in developing this technology. But he also works with business departments, which may be potentially interested in big data technology, such as marketing and credit risk.
Q : How long have you been looking for such diverse staff ?
A : All of them arebased in Moscow. It was very difficult – I interviewedmore than a hundred applicants. Most of themare people from IT companies, or from related fields, such as telecommunications. But 15 is not enough, and I hope that next year the numberwill double. There is a rule, and I tested it in practice many times – you cannot hire a hundred people at once. The team has to groworganically, because you need to developa certain culture, and people need time to understand it. These are the laws of human mind.
Q : Sberbank says a lot about innovation. Of course, we see the work of the bank on a new image: new offices and so on. But is there something that can truly be called an innovation in business ? After all, innovation can be of two types: something really new and usage of that was already invented and implemented earlier. In Russia, unfortunately the second option is more popular.
A: There are many misconceptions about what innovation is. There is a clear distinction between invention and innovation. Invention – it is something that has never been done before. Useful or not. You can make something absolutely new, but it’s possible no one will use it. Innovation is something different from what was done before, and it is something that should influence business. It is always judegd in a context. If you decide to exercise, there are many people who are already doing it, but in your life it is innovation. If there is something that has been done ten years ago in Austria and it can be useful in Russia, for Russia it will be innovation. It’s not necessary to be the first, the main thing is to do something different that has an impact.
In the bank there are a lot of areas where you can implement innovation, and improvements are not always visible. You can change the way of management of the infrastructure and IT costs will fall by one hundred million dollars. Not that this has no effect on the business. But, of course, customers can see more innovation in business processes. For example if instead of issuing a credit approval in three days we do it in 30 milliseconds.
Therefore, I believe that the widespread belief that the innovator should be the first, unique, and must invent – this is misleading. All this is good but not necessary. iPhone has become the second most successful product in the world (the first was the “Rubik’s Cube”). Was it the first or unique smartphone? No. It just had some advantages. Apple iPod brought billions of dollars, but when it was just released the Wall Street Journal had an article on the first page saying: it’s a stupid idea; the market is already saturaed with 230 MP3 players. This proves that to be innovative, you do not necessarily should be the first, unique, and invent something.
Q : Sberbank Innovation Center has been around for over a year. What was done during this time ?
A : I’m glad you asked about it (smiles). Here is a list of all projects (shows two sheets of paper, filled with the names of the projects, small print). Not everyone I can talk about, but here, for example, we launched a “pilot” of the technology that allows us to change the website and online bank of Sberbank. It will allow us to transform the site design using components – the so-called widgets. Yes, some banks are already using this technology, but we must understand that Sberbank is a very large bank. It employs 250 thousand people. And we cannot change that easily as a small organization. You cannot run into productive operation without proper testing of something, when you must support tens of millions of customers. Therefore, although my team has completed work on a new technology six months ago, a productiion deployment team is still testing it, and, I hope, in 2013, Sberbank will use this technology. And then we will be the largest bank that implemented this platform.
Also, for example, we have conducted three “pilots” for the use of NFC technology, but it is still not used only because we do not know exactly what will happen with this technology, and hardly anyone in the world knows.
As a result, we do not have to implement the innovation in productive operation, but we have a very important function for the bank – to experiment with many technologies, which significantly reduces the risk for the bank. You cannot just see something and say, “Let’s make a bet on it.” For us, these bets can be costly. If Sberbank introduces something, then because of its volume it may cost millions of dollars.
Q : That is your role – having spent a minimum amount of money and man-hours to create for the bank a business case for using a particular technology ?
A : Exactly. Projects on my list are divided into several categories. First – only investigation without a project being started. There are 17 different technologies on this list that we considered and decided not to pursue them. In the second list – six technologies for which, as we discovered, is just not yet the time to work with them yet. And there are eight that we have recommended for implementation, but our business partners have not seen a business case for them. Six technologies we have investigated, but decided not to do this year, leaving them for 2013 – there are good technologies, which already have business cases. Nine projects are running now, and ten projects we have already completed this year.
Example of implementations completed this year is a tracking system of searching for “Sberbank” on social media. Another successful “pilot” is a system of managing personal finances. A PFM system is good because it motivates people to use electronic transactions because paying cash they cannot keep track of their expenses or will be forced to track them by hand. And again – although the technology has proven applicability, we will launch it only in 2013. Now there is a tender for selecting such a system.
Q : Sberbank also has a venture fund. How tightly you interactwith it ?
A: We work hand in hand, solving one important problem.To survive, a financial institution mustbe innovative. Most innovations, especially technological, come fromstartups. As long as the Big Six” in IT industry are growing, they are constantly buying small companies. I joined IBM myself through such an acquistion in 2003. At the same time, startups on their own are a risky proposition.
To reduce the risks, you should check the technology. That is whywe test them, we try to apply them, and if the project turns out successful – we are ready to make a strategic investments – the company gets a good reference from us, revenue, and an equity investment, to help it survive– $2, 3, 5 million – so they can grow and attractmore customers.
Q : Do you really believe that successful financial startups may appear in Russia ?
A: We have notmade ​​any investments, but we have formed a“funnel” for the selection ofstartups. The selection is not limited only to Russian startups, we are looking for them around the world, in various startupincubators. Now our “funnel” can be called pretty “healthy” – it has 25-30 startups from around the world.

Q : What are these companies ? What a startup should be to attract Sberbank ?
A : It’s easy. First, they should already have the product. We do not work with concepts. The company must have at least a couple of clients, and it does not matter if these clients are not working with the product, but only testing it in the “pilot.” We do not want to be the only customer. We need verificationfrom two, three, five banks.

Q : Can you givean example of such a startup?
A : The same company whose online platform where we are implementing is a good example.

Q : But can it be called a startup ?
A : Absolutely.It’s a small company. I think it is worth to determine what a startup is. This is not necessarily two people who have an idea and 50 thousanddollars. For us a startup is a private company in a growth phase. We do not make seed” investments. Every week, I see a lot of ideas and suggestions. But whenI ask, “Do you have a product?“, I get the response: “No, but I have a great idea.” Fine, but we’re not one of those investors.
Q : You recently attended a SIBOS conference, which is held by the innovation division of SWIFT called Innotribe. What does a seemingly conservative organizations like SWIFT have to do with innovation ?
A : Innotribe includes about 20 of the largest global banks, including me as a representative of Sberbank. In it there are also members of venture capital and consulting companies. The activity of Innotribe is now divided into two major directions. First – working together to solve some of the banking industry common problems (and that is what “old” SWIFT does). One of such projects, for example, is called Digital Asset Grid, and it aims to define standards that allow bank customers to store all of their personal information in one central location with standardized interfaces.
The second initiative of Innotribe – the organization of an event called Startup Challenge, for which Sberbank was a global sponsor in 2012. We had four such events this year. My team and I viewed the presentations from more than 500 companies – start-ups from around the world that are developing technologies for financial services. This gives us a very good idea of ​​what is going on in the world, because one of them was as in Asia, the second – in North America, the third – in Belfast, in Northern Ireland, and the last in Japan.
Q : What do you call digital asset grid, in Russia turns out to be part of UEC, the implementation of which is vested to Sberbank ?
A : Yes, that’s right. Therefore, we do not participate in this initiative of Innotribe, we are just watching it.

Q : You have once said that the main innovation for banks in the near future will be transition to digital, that is to a digital environment. What exactly do you mean ?
A : As simple as getting rid of paper. Electronic documents address only one part of the task. We need to consider and analyze process by process, and if in any of them has paper – need to get rid of it. First of all, get rid of paper in the relationships with the client. We are forced to carry around our passport, make a copy of it in the bank, etc. A lot of things are done on paper.

Q : But how the bank can get rid of paper, if according to the law it needs to be used ?
A : These laws should be brought into line with the realities of the 21st century. Account Management on paper was invented in the 19th century. Since then more than 200 years have passed and the technology has to change. I do not know with which mechanisms we can change the laws, but it is clear that banks, and any large organizations, including government simply cannot continue to work on paper, as long as in a digital form, everything happens much more quickly and flexibly. This is the only way to adapt to the changing environment. I think in 5-10 years the rejection of the paper will be possible.
And, of course, “social”, “mobility” and “cloud” and “big data” – all this will increase the amount of digital data, create the possibility of transition to the digital environment. But organizations need to clearly understand what they want to become.

Q : What do you think about Brett King saying that the main interface will be the mobile phone banking, and the banks will turn in the end to be the “pipe”, conducting commodity transactions ?
A : There is no doubt that the banking business model is changing. However, until now there are banks opening and working on a new branch network – for example, Metro Bank in the UK. There are many studies that show that people’s confidence in a bank which has branches is much higher. For example, I am a customer of the bank ING Direct in Canada, who worked for several years only online. At some point, the bank has opened several offices, to prove that it is “real” and immediately its customer base grew.
Yes, the branches are now transforming, many banks make them look like a cafe, a shop. I saw two branches of different banks on a street corner in New York. One of them had just been made in the form of a cafe, and there were about twenty people who were sitting and talking. I do not know if these people were discussing something related to the bank – I was not listening – but at least they were there. At the same time a nearby bank branch office was empty as a museum on Monday morning.
I respect the efforts of Brett King to look at banking in new ways, and I saw a lot of people, showing a mobile phone with the words : “This is the bank branch of the future!”. I myself did the same. But banking is a tough business. And it needs to be adjusted. When I tell my colleagues in the UK, what some of the Russian financial start-ups are doing, they are surprised – in another country that would be illegal. For these reasons, Sberbank cannot be the first in such initiatives. We are too big to bet on some ideas that may not be correct.
Q : but on the whole, do you agree with the statement that “banks must change or die” ?
A : I think so. Because things are changing around us, whether we like it or not. I have an example and it is not out of the banking industry – a lot of people have heard of 3D-printing technology that allows you to make small objects based on a predetermined model. This will have a significant impact on the world. Many people who know how to operate the manufacturing machines today, will remain out of work. We’ll have to do something about it as a society.
In the banking industry, the idea that the phone can be a means of account management, no doubt, will change a lot of things – the banks will have to provide people with the interface. Otherwise, they will not disappear, but will recede into the background, become a utility. And this is already happening. After all the activity of PayPal, for example, cannot be considered as “payments”. The present payment – is what SWIFT makes, when it transfers money from one bank to another. At the same time, companies such as PayPal, putting very little, take to them the most profitable part of the banking business. Therefore, banks have to learn how to create such interfaces; otherwise they will become only marginally profitable. I already have an e-wallet – this is my bank account, and I do not need a purse mediator. What banks need to do is to create a good interface, and we at Sberbank are also doing that.
Q : Why, for example, Sberbank has not yet offered to small business a mobile POS terminal, which has become well known product on the market ?
A : Like I said – everything takes time. We have to look closely to these devices for a long time; we tested the products of 9 companies already, so that I may know more about them than the banks that are already selling them. Sberbank is the most trusted bank brand in Russia. It takes years to gain the trust of your customers, and you can lose it in a day. So, yes, we will have the mobile POS terminal, this project is on my list, but we need to choose the best of several products and test it.

Q : Many banks consider innovation as toys and prefer not to commit the funds. How large is the innovation budget of Sberbank ?
A : Banks definitely need technology innovation, and that’s why I’m here. Sberbank has a strategy for innovation and a budget for its implementation. It is included in the IT budget, and is less than 1% of it, that is hundreds of millions of rubles. But this budget is for the experiments. As soon as the productuin deployment stage of the project comes, the money will be allocated to it from the “big” IT budget.

Q : How we can stop just talking about innovations and start really benefit from them? How we can build an innovation strategy ?
A : We’re looking at different sources of ideas: idesa from our employees, what is happening on the market, the actions of other banks, startup competitions, offers from our partners. Then we form a “funnel.” If the idea is interesting for us, we will investigate it, we build the concept. I usually give for it no more than two weeks.
There are two main characteristics of the idea. First, it should be a technological innovation, which is determined by matching at least 8 of the 12 criteria that we have adopted. Second, we must see it as a potential benefit for the business. At the level of research we do not do a formal calculation of return for investment – it is a bad practice, because usually kills innovation. ROI is considered only when there are high stakes. So, if we decide that the idea is good enough and it has the potential benefit for the business, we are prepared to put into it, say 50 or 150 thousand dollars. But no more.

Q: Who determines the usefulness of the idea for the business?
A: This is usually the head of the business areas – for example, business card or online banking. All of them are our partners, and we are constantly interacting with them.

Q: And at what point do you ask advice from CIO, Viktor Orlovsky?
A: I go to him after we have a concept. Since this is his budget.

Q: What is his function? He must decide whether the solution integrates into existing IT architecture?
A: No, we discuss the architectural aspects with IT architects even before that. CIO simply needs to know what is happening.

Q: And after that you start the project?
A: Yes, in the project we need to buy some equipment, pay developers to build interfaces and test it all. Earlier I called this stage “pilot”, but then began to avoid this term because it is ambiguous. I like the word “experiment”, but people do not always respond well to it, because in their minds, it is related to chemistry and explosions. You can call it “prepilot”, doesn’t matter. What is really important is that we are working with this technology, not just leaf through the brochure or look through the demonstration. This program, which is installed on the phone, on a server or on a chip, is executing transactions, etc. This stage may last a few months, but not more than six months.

Q : You are testing the technology on employees of Sberbank ?
A : Yes, all employees of Sberbank are our clients. But first of all the technology is tested by ourselves and our families, the staff of our department, to minimize technological risks.

Q : What kind of technology gets into production operation ?
A : At the end of each project my team is writing a recommendation. Of the ten pre-pilots we have “failed” only two. One – because we decided that the technology was not good enough, the other one as we realized that it only solves part of the problem, and we are now looking for a more comprehensive solution. The other eight, we recommended production introduction, four of them have already been implemented.

Q : What part of your projects should be successful for the business not to decide that you’re just wasting money ?
A : I think, based on my experience that a good mark is 50%. Because if all 100% of the projects are successful, it means that we choose technologies that are too safe, not new enough. Utility of new technologies can only be tested in practice, if it is not possible to predict – it is not R & D, not an experiment.

Q : How long before the project should be profitable ?
A : 6 months to two years from the start of operation.

Q : But there are some trends, and if you do not follow them now, in three to five years it will be too late.
A : This is a good argument. There are some technologies, such as big data, the implementation of which takes years. That is why we are starting a project in this area right now – even if we do not know how we will use it, you need to create opportunities to explore technology, to train people to start getting a little benefit from it.

Q : Since Hadoop is free software for experiments with big data you will probably need only people. How many people are you willing to engage in such a project ?
A : Indeed, Hadoop allows us not to spend money on software licenses, and next year eight people on my team will deal with big data projects. We started this year, and with the help of two people we are already building our first Hadoop cluster. But we also recruit the external consultants to speed up the process. It is a leading US company in the use of “big data”. We have already discussed a dozen projects using big data technology, which can be potentially good for business.

Q : For the next year people waiting for bank projects related to social networks no matter what it can be. What will Sberbank do in this direction ?
A : Indeed, the so-called “social” applications can mean many things. Sberbank is planning several projects in this area. For example, using social networks people can find our ATMs and branches. In addition, we will create a crowdsourcing platform in social networks, where you can discuss the work of departments, etc. In the same application ther will be an interface for a conversation with the operator, based on the functionality of Facebook chat. Some banks call it a “virtual office.”
By the way, our new subsidiary, Turkish Deniz Bank, recently received the award in Washington, DC as the best Facebook bank application. This is an online bank on the social network platform. And we are also thinking about implementation of such functionality. In any case, this is not a quick process, because, again, we are talking about tens of millions of customers.