Offloading of data to big data platforms
“In the beginning, expectations were extremely high, and led to disillusionment. Now, thanks mainly to new technology, the market is catching up with those original expectations.”
International software and consulting company Adastra Group saw revenues exceeding two billion Czech koruny (US$92 million) for the first time last year when, alongside building data warehouses and conducting business intelligence analyses, it began to direct its attention to big data.
“[Even though it’s a hot topic at the moment], I don’t think big data will replace data warehousing or business intelligence in the future,” Pavel Kysela told Lupa. Kysela joined Adastra thirteen years ago, and is now Managing Director of its Czech branch. The Czech Republic accounts for roughly one quarter of the company’s total turnover.
According to Kysela, a number of companies have sprung up on the Czech big data market, and are making an effort to assert themselves.
“We’re now at a stage where the buzz around big data has prompted a lot of companies to feel compelled to move into this sector and try to establish themselves within it,” says Kysela, adding that the first wave of consolidation is probably approaching.
Adastra operates in the business-intelligence sector, but has recently begun to concentrate on big data. What’s your primary focus now?
Adastra grew up on integrated business intelligence, but we’ve gradually expanded into other areas that are, to a greater or lesser degree, connected to that original focus. So, from the very outset, we’ve concentrated on data and all its related fields. The traditional spheres of operation are widening steadily.
Into which markets?
We’ve added data integration as well as solutions that use specific data for specific business purposes. The latter includes, for example, solutions for campaign management, fraud detection in telecommunications, insurance or banking, and approval systems in the banking industry.
For instance, when you apply for a loan, the bank uses historical data, behavioral patterns, and so on, to assess the likelihood of the loan being repaid.
It’s about a data-warehousing superstructure that, at the beginning, consolidates, cleans, and readies the data for future use. Then, we build additional “boxes” around this superstructure to support previously specified business processes.
So you’re also interested in data analysis, not just in data warehousing?
We’re able to provide everything I mentioned earlier. In other words, we not only carry out the analysis and implementation for data warehouses but also prepare them for use. We have pre-built systems that we can deploy and implement in individual cases, where, of course, we tailor them to the client’s requirements.
Over time, you’ve logically arrived at big data. Do you feel there’s a lot of potential there?
Essentially, yes. We see big data as a field that, to a certain degree, both overlaps with and extends what we’re already doing. I don’t think that big data will replace data warehousing or business intelligence in the future; it’s simply an additional, complementary sector that can extend the use of information systems.
The fact remains that every big company has its primary systems that sustain the organization’s fundamental agenda, for which using and processing data will continue to be necessary.
Here, the traditional practices still make sense. Big data is completely new. It’s one more area that naturally relates to and extends our portfolio. But that doesn’t mean that it would replace what we’ve been doing so far.
But it’s still a new market with great potential.
As we see it, big data basically has two main directions for further development and use. One of these, of course, is the increasing emergence of new technologies that are capable of processing unstructured data or data from sources that weren’t used in the past. This means it will be possible to enhance your information-management system or business-intelligence solution with additional resources and, in fact, make it more intelligent.
The second area for development is the technology emerging in parallel with big data that essentially enables hardware to achieve the performance and reliability currently offered by traditional solutions. In other words, you’ll be able to process a lot more data for the same amount of money.
Which gives companies certain advantages.
Yes, the first of which is that you’ll be able to use this new technology with some traditional solutions to significantly reduce your company’s operating costs. This means that you’ll be able to achieve, for a reasonable price, things that weren’t previously economically viable.
Can you give a specific example?
For example, we conducted an analysis for a car manufacturer, where it was necessary to process massive amounts of data obtained from the control units of vehicles on the market. Statistics showed that, under the mobility guarantee, 40% of breakdowns are repaired on the spot, while the rest are all fixed in service centers.
This German carmaker wanted to reverse this statistic, and have 60% of breakdowns repaired at the site of the incident. Not only would this dramatically lower the company’s costs, it would also give the customer a positive experience and increase his or her satisfaction with the carmaker.
How was the analysis carried out?
Our model was based on analyzing data from the control units of vehicles that had broken down, and of similar vehicles. This assessment increases the likelihood that, if a problem does arise, the mechanics will be able to repair the car on the spot, as they’ll know what’s probably causing the issue and how exactly to fix it.
If this model had been built using traditional technologies, it would have been so expensive that no one would have tried it.
You’ve spoken about banking and now about the car industry. Which sector are most of your clients from?
Historically, we’ve worked for companies that have large amounts of data but haven’t called it “big data”. That is to say banks, insurance companies, telecommunications operators, sizeable retailers, energy companies, and, to a lesser extent, state institutions.
I’ve mentioned a car manufacturer, but we don’t have too many clients from the automotive industry yet. In general, these are all sectors in which we’ve been established for a long time or which we’re now trying to penetrate, for example, the energy industry.
All are companies where big data bears some significance, and all of them are involved in it, to a greater or a lesser extent. In fact, even though big data has been a hot topic for a while now, there aren’t all that many companies that have a clear idea of how they want to use it.
And corporate demand is rising?
Unlike many other fields, this is an area that has a lot of buzz around it, and companies are willing to invest in it. But in many cases, they don’t know how. As a result, a few different approaches can be seen on the market.
One is that a company buys and implements technology that can handle big data, but only figures out afterwards how big data might actually be helpful. A second strategy is that the company first defines how exactly it wants to use the big data, and then chooses and implements a relevant system. And the third approach has come from the fact that big data’s recent popularity boom has given rise to many companies that want to take advantage of this phenomenon, even though they’re not offering anything new or revolutionary. So they simply stick the “big data” label on traditional products and sell them for a higher price.
In this respect, however, the firms’ maturity differs. There are companies that are ahead of the game and are able to derive concrete business results from big data. Some companies are also capable of selling big data. And now I don’t mean an individual solution, but rather their product, which they sell to others. And then there are companies that are just fumbling.
But the trend is clear.
In my opinion, which is corroborated by a few foreign analysts, this market will develop in the shape of a “W”. In the beginning, expectations were extremely high, which led to disillusionment when companies realized they didn’t know how exactly big data could help them.
Now, thanks mainly to new technology, the market is catching up to those original expectations. There are concepts that work, and now the real phase is beginning where companies are starting to use them. So, in the long run, those expectations may be met.
It’s a domain that can really move the world forward. But it will take some time before the technology is more advanced, before the market understands the true scale of opportunity, and so on.
How do Czech companies differ from foreign ones in their approach to big data?
If we compare the markets as such, there isn’t really a big difference; I wouldn’t say the American market is many years ahead of us. But it is true that there are some big players over there who are far ahead of even the most progressive here. They were able to respond quickly and with courage, and invested relatively heavily. At the same time, there are companies that, because of the nature of their business, work with categorically different volumes of data and different information sources than Czech companies. But then we’re talking about Amazon, Google, and similar corporations.
Are small businesses interesting for you as well?
That isn’t our target group; we’re not able to offer them suitable solutions as yet.
You mentioned that you’re not very active in the public sector. Why is that?
Private companies are built on certain principles that don’t work too well in state institutions.
There has been fierce competition on the big data market recently. What potential is there on the Czech market? In other words, is there still space for more players?
That’s really hard to quantify because the whole market is booming. We’re all wondering how big it will grow.
Regarding the competition, as my grandmother used to say, “many are called, but few are chosen.” A large number of companies are pursuing big data. But, on the other hand, a considerably smaller number of them have any significant practical experience in the field. In this respect, the market is still in the process of establishing itself.
Has the time come for a certain consolidation of the market, or aren’t there enough companies yet?
We’re now at a stage where the buzz around big data has prompted a lot of companies to feel compelled to move into this sector and try to establish themselves within it. But I think that, over time, the market will consolidate and clear, and will show who is actually able to deliver something real and who can succeed in offering the client something new, something advantageous.
At this point, it’s difficult for clients to make decisions based on previous experience because hardly anybody has any. So they’re currently making selections based on other criteria.
If we speak again in a year or two, there’ll already be a number of companies that will have delivered real projects, which will obviously strengthen their hand in future tenders. In my opinion, this market is only just starting to gain shape and definition.
Will the market also branch out?
The market may split in the sense that some companies will start to specialize in specific areas, perhaps focusing only on the technology, or on consulting, or on data analysis, and so on.
We often hear technology companies talking about how hard it is to find qualified people. Is that true for you as well?
Yes, our team is really overloaded at the moment; it’s hard to find technically educated people on the Czech market. Plus, the situation may only get worse in the next few years, from a demographic perspective, as the number of people completing higher education is decreasing. There’ll be an almost 25% decline in the number of graduates from the last academic year to 2017/2018. On the one hand, demand is rising, while on the other, the pool of capable people is diminishing.
How many people do you have now, and how many do you need?
There are now about 220 of us in the Czech Republic, and if we added 30 to 50 employees in the next twelve months, I would consider that a success.
Now a slightly philosophical question: isn’t there a danger that our decision-making processes will become increasingly dependent on output from information systems capable of handling big data, which the human brain inherently doesn’t have the capacity to grasp?
I’m not sure if that would really constitute a fundamental break [from current norms]. Big companies already base an overwhelming majority of their decisions on the data gathered from information systems.
The fact that we’re now going to use slightly different technology, that we’re going to add more data, doesn’t change anything with regard to the principles underlying current operations. Big companies’ information systems are already so complex that, to a certain extent, you have to rely on the IT infrastructure. The margin for error depends on the quality of the system, the processes, and the people who are operating the systems and processes.
Now, it’s another layer of complexity, another reason to look at the quality of the data and the way it’s used. But, fundamentally, it’s just an extension of a system that’s already in place.