Sberbank 2.0

I have been interviewed recently by Tatiana Naumova for a Russian investment magazine. Since the interview will be published in Russian translation, I am including here the (slightly edited) English transcript.

4 main directions for innovation in Sberbank this year – in our interview with the Director of Technology Research Center Dr. Mircea Mihaescu.

Author : Tatiana Naumova


Nearly six months ago, Sberbank announced the establishment of its R&D-center. How is this new project of the bank being developed right now?

There are several things that need to be put in place each time you start a new team. First of all, you must have the team. For the team you need to have the hiring policy — what kind of people you want to hire. They have to match to what kind of activities you want them to do. Then, you need to have a strategy for how you are going to move from nothing to having a successful R&D; department. You have to start from zero and grow incrementally. What have we done so far? We have created a strategy for hiring people. We have executed on the strategy — we have most of the team hired. We have created policies, procedures and the strategy that will put everything together on how we are going to do R&D.; We have been able to start some projects. Moreover, we have been able to have a handful of successful projects. We are at the point of executing our 2012 strategy: now we have a proposal of specific projects — over 20 projects to work on this year. Plus, at the end of last year we had a relatively large team in an advisory structure that we put together. We call it a Technology Innovation Council and includes representatives from most areas of the bank.


What projects has R&D; center already realized?

We have looked at modifying the online banking experience. The objective is to move away from a static “one size fits all” web-site to a customizable one. So you can say — “I don’t really want to know about mortgages because, I am sure I will not touch mortgages for the next 5 years. Take that away from what you are showing to me”. At the same time we want the bank saying — “We think that you should be interested in credit cards, because we see that you have an uneven cash flow. And the credit card may be able to “smooth out” your cash flow”.

It is not only the information on those screens that can be customized but also the components. The principle is similar to iGoogle. Those components are called widgets. There are a handful of companies in the world that have developed widgets for banking.

Sberbank has two options: either we go with the generic widget providers, that include large companies, as IBM and Microsoft, or we go with the widget providers for banking. We decided to go with the widget providers for banking. We looked at the main vendors. We picked one of them, Backbase, a Dutch company– we started working in August and by September, at the Sochi Forum we showed the demo. That demo was not connected to the backend system, but still it was showing the principles of how the new Sberbank web-site will work. Then in between September and pretty much middle of January we connected that demo to the transaction system. Now it can show the real transaction data. And for us it was the first successful project, because it moved us from what we are supposed to do in technology innovation all the way to starting the production development process. Because what we do is not just selecting an idea that can help the bank — in 3 to12 months we experiment with a new technology, we pilot it, we modify and integrate it, and finally we make recommendations for the Sberbank IT systems.


What are projects you want to launch this year?

We have 4 major directions for this year. Two of them are very big — it is cloud computing and “big data”. Big data is a huge field that we are now only starting. I just recently read that the world’s leading information technology research and advisory company Gartner predicts that by 2015 60-70% of companies will not take advantage of big data. We want to be part of the 20-30% that will take advantage of it. What we mean by that? We are going to use new technologies (hardware and software) that will let us both collect the enormous amount of data that is being generated in the bank — all the transactions, all that is happening in digital form — put them in one place and run all kind of data science algorithms like data mining and pattern analysis, pretty much any kind of analysis that will help us cut cost and offer better, more customized service to our clients, or help us know more about them. The field is now starting to move from being a scientific field to becoming more client and business-oriented.

We are also looking at social computing — Facebook, Vkontakte, Odnoklassniki and so on. We see millions of people creating a lot of data about themselves on the internet. We don’t map that today with our customers. So, now, we are looking at ways of matching them together to give our customers a better service. Of course, if they want to share it with us. If our customers want to keep to us totally separate from their social networking persona — we are not going to do anything. But if they want — we will build a capability for both them and us to benefit from the amount of data that is being collected totally independent of us.

This is big data. We need to work on it to have a better analysis, for the reduction of costs, to offer better services to customers. And also to understand our business better, so we can optimize it.


What is the second direction?

The second is cloud computing. I have been one of the pioneers in cloud computing. I joined IBM in 2003 at the moment of acquiring a small company where I was the deputy of engineering. Where we build the technology that today is a core component of the IBM cloud computing offering. This year in Sberbank we started a project to move some of our production creation to a private cloud environment. We are looking now at the model where we will deploy the application in a private cloud inside Sberbank. So, we are building something like a cloud, a cloud technology, but inside Sberbank. Because Sberbank is so big that we can really benefit from that. We are now looking at the analysis of all application we have in production. Which ones will benefit of clouds. And usually the one that have variable workload can benefit from cloud technologies. You don’t want to place in your data center a huge computer that cost millions of dollars and most of the time it does nothing. And it does something useful only at the end of the month. So, that is the kind of application that will benefit from cloud. So, cloud is an infrastructure game.


What is the third direction?

There is a book by a bank visionary — his name is Brett King — the book is called Bank 2.0. The main point of the book is that new technologies change the bank a lot. And the top new technologies that will change banking are social computing and mobile. So, our third direction – we call it Bank 2.0 by shamelessly copying book’s name. And the idea is that new technologies are coming all the time to change how we do banking. And banking has to use them to understand them and to apply them because we don’t want to become Borders (bookstore that just went bankrupt because bookselling business has been taking over by the web). The idea is to keep an eye on technologies and to decide what we can use in here. One example — we are launching a project to put a customer service application on Facebook. The Sberbank application on Facebook will help us to share information, some people can comment and some people can say like. The idea is that Facebook or any other social network is a communication platform. And now we are really going to use it to communicate with our customers. Because one view is to say – “they should come to our site and do these things.” But maybe they don’t. Because recent data shows that more people are on Facebook than on any other site in the world!

We are also looking at the customer service. Customer service is a huge problem. It is human intensive and that is why we are looking at the technology called natural language processing (NLP), natural language understanding. Artificial intelligence if you want. It is still a dream — talking to the computer. The technology is not here yet. But there are many other things that we can do in the meantime. For example, work on our frequently asked questions. Right now there are distributed along the website. How to navigate? You don’t know how…We are going to put everything together. And when you type a question on our web site interface we will do natural language processing, try to understand the question and suggest top-3 answers that can match your question anywhere on the site. Seems like obvious — but ordinary computer programs cannot do that that today – it requires NLP. For example IBM has a program that uses natural voice recognition and language processing called Watson. In 2011 it has won the game Jeopardy. But it was very fine tuned for that problem. And it is a very expensive computer. We are not at the level of using that technology in Sberbank, although we work with IBM and we guide them to modify that in order to be used by us.


And we still have the fourth one…

This one is a strategic and tactical one… This is our top priority in finding a solution today — mobile payments. In Russia 95% of transactions are happening in cash for various reasons. So, mobile payments is a specific feature we ned right now. For us it will be both a huge cost reduction, and making our customers’ lives much easier. And the impact on the economy… We don’t talk about that — but we think it is going to be huge,because you reduce the friction in executing transactions. We are looking at this problem in several directions. We have a pilot connected with NFC-technology (contactless payments). The problem we are working on is how can you modify an existing phone to be used as an NFC-payment device? We have two options in the pilot. One of them is a sticker, it is a piece of plastic that you attach to the phone and that’s your payment mechanism. And the other one — is microsd card inside the phone that people use for extension. I don’t know how many will take it away so they will not have games for example — but that is the point in experimenting. And we are talking to a lot of companies — and what we see is that by the summer we will have more than 100 smartphones that will incorporate the NFC-hardware already. To sum it up, this year we hope to come out with mobile payments — person-to-person payment — using NFC-technology.


These four points sound like a huge job. But you also have venture business…

Yes. We are looking for successful start-ups that have passed the early stage, the ones that have passed the idea phase, they have a product that they can sell to a bank and a bank can use it. That is the investment strategy. If we find them useful for us — we first do a pilot. You can think about it from a venture-capital investor point: detailed pilots that we are doing, to see if it works for us, is the best due diligence you can have. So we do the due diligence. If it works out — we are going to make the company one of our suppliers. We are going to give them a huge reference account. We are going to give them validation. And we are going to give them revenue. So they we may estimate they will grow like crazy – and become a good investment opportunity. Our strategy is to set up a venture fund that will invest in any company in the world that is in the following sub-set: it’s in early stage growth, which means a hand full of customers. It makes technologies for financial services that we have looked at and decided that it is something good to use. This venture capital fund, by making an investment, will make sure that the company that works for us or with us will have the capital to survive, and will also benefit from returns since we think that it is a good company.


But still about the process of selection. How do you understand that this is the project that Sberbank needs ?

In venture capital terms it is called the deal flow. How do you get a good deal flow? How do you get 100 companies to look from? And pick the ones to work with? There is the ad hoc way, semi-systematic way, and systematic way. The ad hoc way — someone knocks on your door and says “This is a good company!” I also add to ad hoc whatever you read in articles. The semi ad hoc and this is what the most venture capitalists do — is through networking, personal connections. What we are doing is that we are working in an open manner and in this sense we are relying on others. To give you 3 examples: Gartner is publishing every quarter a report called “Cool technologies to watch”. And for each of technologies the company lists 4-5 start-ups that they think are better than the others. So Gartner is one-way to do systematic search. The second one is sales conferences dedicated to start-ups in financial services. One of them is called Finovate — Financial Innovation. They run Finovate in San Francisco, New-York and London. Last time I attended I talked to all of them — there were 40-50 companies, and I liked 7 or 8 out of them. I distributed for further investigation 2 companies for each one of my project managers. With some of them we are already running pilots. For example, we have a pilot with a unique plastic card with a small button. In the long term plastic cards will disappear, but not in the next 3-5 years. Maybe ever. Inside this card they made a computer so small — that you even don’t know that there is a computer inside. When you push the button the plastic card can become a debit card, a credit card or a points loyalty card. You push the button and change the specialty. By the way, at Finovate this company won the price for the best start-up. We already have a proposal from them and we’re going to do 300-350 cards and see if it works in our system. The third way to search — is start-up competitions. This is where we look for start-ups that need money. I’m a judge for some competitions. Last week I’ve reviewed over 90 start-ups, from which I picked only 2. I talked to them and we picked only one with whom we’re going to work. We may even make an investment. The project we’re doing with this company is related to SWIFT an association of more than 10,000 banks, it’s a mechanism for bank-to-bank transfers. These guys have made a gateway into SWIFT and you can use SWIFT very cheaply to move money across the borders easily. Moreover we can embed this technology into our mobile app. If this will succeed — this can pretty much destroy a business model like Western Union. Putting it all together: Gartner will tell us about the companies at the end of maturity cycle, Finovate and sales conferences will tell us about the “middle period” companies, start-up competitions will tell us about early-staged projects. But we are not looking for ideas, we are not angel investors. We are looking for the ones that passed angel investments stage. Overall I have 1 objective and 2 mandates. The objective is to identify technologies that will be useful to the bank, and mandate is first to try them, and second, maybe, invest inthem.


Starting with investing $100 million in the venture fund, Sberbank intends to bring the volume of investments up to $700 million in cooperation with international partners. In general, how is the process of candidates’ selection organized? Who is the ideal partner for Sberbank in the development of its venture business ?

All venture funds are limited partnerships: with limited partners that will provide money and will approve the strategy and the manager, “the venture capitalist” who actually “runs” the fund. In the first fund that will be $100 million committed, Sberbank will be the only one limited partner, the only investor and Troika-Dialog will be the general partner. Second and third fund will increase the overall volume of venture funds to $700 million total. For the fund where we are the main investor we will be mainly looking for large international banks to become limited partners with a smaller investment level.. But we don’t want to do that from the beginning because we don’t have a long enough track record. After there will be a track record, let’s say in 2-3 years, that will be the time to get some of them to invest with us. Why international institutions are the best? Because the companies we try to invest in will be developing technologies that banks or insurance companies can use. So, if there are 3 investors, that don’t compete with each other (for example Bank of America, Santander, Agricultural Bank of China – neither of them compete with each other or with Sberbank) this will be an ideal situation.


What are other main countries and regions you are looking at right now?

A couple in Asia, 6-7 in Israel, a couple in France. Maybe 4-5 in Holland. A couple in UK. More than 20 in North America, USA and Canada. In Russia we are looking at less than 10 and there are a couple in Ukraine.