- Sergey, how long have you been working with Big Data technologies, and how do they help your customers?
- Approximately 5 years. The last two years we have been working more in BDSMM (Big Data Social Media Marketing). In case of the Kazakhstani market, we are mainly turned for some know-how. The largest customer of Big Data in Kazakhstan is the state, because it is an expensive technology to ordinary companies, it is easier for them to buy "small" software with a short payback period.
The biggest problem we have is the integration of all services in the company. In practice, there were self-written ERP-systems in which there was no possibility of data transfer and integration with them. One of our achievements is work with the client's portrait. The client base is expanded to 75 parameters. For example, an insurance company has a million people in its database, but every year at any company at least 20% of the base die, due to lack of communications or misunderstanding with clients. Customers do not buy for various reasons: the manager was rude to them; they were not called back, competitors intercepted and so on. We solve this problem as follows: we divide the client's portrait into micro-sections by collecting information from open sources. Let's say you support "Liverpool", we automate it all and send you information: "Hey, "Liver" is playing today, what's the score?" People often leave their data about themselves; they talk about family members, share information about where they have been and so on. If all this is collected and processed, powerful results can be achieved. Of course, everyone wants to increase sales or reduce the number of employees by removing the human factor with the help of Big Data.
- What inquiries were addressed personally to you and your colleagues?
- There are a lot of customers. We are now conducting projects with the government; so far I cannot talk about them. The state wants to come to some kind of transparency. Examples: the program "Digital Kazakhstan" or "Smart City". The latter is a large amount of data. Unfortunately, we still do not know how to process data in our country.
The second con: people do not trust local companies. There are 4 international franchises; these companies outsource our state bodies on about $4-5 million minimum. Then the contractors do something themselves and find subcontractors here, such as us. Thus, the margin is reduced by 80-90%. For us, a check for 300 thousand dollars is good money in Kazakhstan. But when you find out the actual amount of the order, the margin... And this is one of the main problems in the market. Though they say about the need for Kazakhstani specialists as a priority, in practice this is not done. I'm talking about major projects.
- What data are collected by your client companies? What data are taken from social networks and so on?
- There are many monitoring systems that collect information from different parts of the Internet by keywords. For example, if you wrote the word "Loan" somewhere in your comment, the systems "monitor" and collect that. It is possible to get a good feedback, mood, or involvement. And now with the help of Big Data, many companies no longer walk the streets, but set the right keywords and collect information in a very large amount. Here is an example: Cisco Company has been using an unusual way of communicating with customers for a month and collected more than 800 useful ideas for its product development. During that period a specially created community in the SS Social Media Handbook has been visited by 1,200,000 customers. The second good example: Dell Company with the help of social media is actively restoring its own reputation and has already increased its portfolio by 6.5 million dollars.
By your posts, it is easy to determine what type you are, what you "repost", who you "like" or “comment” on who ”likes” you and so on. Based on that, you can build chains of people: whom you are friends with and so on. In Kazakhstan we like to boaster, for example, if we go somewhere, we show geo-tags, which are one of the best parameters for collecting information. By these parameters you can roughly understand the average person's spend; how open he is to the new; whether some products or services can be offered to him or not, and so on. Let us see, what only one parameter can give us?
Recently, the British company has developed an algorithm that will be able to predict your location for all 24 hours. The program uses the data of your smartphone and correlates them with information from the gadgets of people who are nearby. The discrepancy is no more than 20 meters. Having carefully studied the habits of the target audience, marketing specialists are able to plan places for advertising and even change the content on outdoor billboards at a time when potential customers of a certain target group are passing by.
Now in the post-Soviet space, the main problem is not the collection, but the processing of data. Big Data is constantly updated, there are more and more of data accumulated. Analysts cannot cope with that. Therefore, there are many orders on the market, and the competition is small, there is enough work for everyone.
- How many competitors do you have in Kazakhstan, and who of them have a good knowledge of Big Data?
- 3-4 companies. If we talk about orders, then sometimes they are taken by some indistinct people, and the Big Data market is small, everyone knows each other. When such an unknown contractor appears, it becomes clear that this is some kind of recoil tender. Then this order is redirected to either Indians or Ukrainians.
Another problem in the market is support. With Big Data, you are forever attached to such technical support: something needs to be constantly improved, repaired and so on.
- In your articles and posts on Instagram you wrote about how the beginners can start working in the Big Data filed. How difficult is it to learn, and why the number of good specialists in this field in Kazakhstan is so low?
- Because the Big Data technology is very young. It became more or less popular worldwide in the 2000s, and at the end of 2016 in our country. They say a lot about Big Data, but this information is not always correct. Some say it is a large database, others call it blocking and so on. And very few people understand that Big Data is not just the data, but the methods of their processing. There are structured and unstructured data. Structured data are the data that come in a specific format, divided into cells. Unstructured data are, for instance, when you need to digitize the call center records. Managers talk, you using a neural network convert all conversations into content, understand it and give the result, as a picture or video. All customers have a problem with this.
As for the Big Data training, the problem is that there is not a decent higher education in this area yet. There are courses that, in my opinion, are far from practice, everything is on the level of theory. There are 2 directions in Big Data: engineering and analytics. If you have a good knowledge in higher mathematics, you can go to Data Analytics; and if you just have programming, you can go to Data Engineering. We still cannot grow even Data-engineers, as there is no appropriate environment for practice. Each new order is individual, there are no templates, a good background on programming is needed, it is important to understand ERP, CRM-systems, cloud technologies, and various services. It is a long time talk.
- What did you want to say with your posts on Big Data: just show the direction or teach yourself?
- I just want to show how you can start this way of learning Big Data, because I have not seen a learning methodology anywhere, a lot of waffle, and no real step-by-step technique. Having gone through these stages of training from general practice, a person, of course, does not immediately become an excellent specialist, but at least he can get a job as a fresher. In Big Data, a lot of people outsourced. It is possible for a small salary, purely for practice, to go to work in Ukraine, to India for six months, but there is no point in going in Big Data without knowledge of English.
In fact, the shortage of personnel in this field is observed not only here, but also in major global corporations, for example, Google, Yandex, Mail.ru, Facebook and others. They started to create such free courses themselves. If upon finishing these courses you successfully pass the exams, these companies can hire you. We could also train, but, again, we do not have an environment for practice in our country.
My task for today is to at least slightly popularize Big Data, because this technology has great prospects. For example, in Kazakhstan there are no large deployments of Big Data yet, although there is a need for it. Until now, no one in Kazakhstan, neither the state, nor business can acknowledge a successful implementation of Big Data.
I think the first major implementation of Big Data will be in the Ministry of Finance, because we have a big economy market, including numerous bazaars, and it is difficult to control. And it impacts the country's GDP. And there will be enough work for everyone, since great software will be accumulated, and a common task will be to collect different services within the Ministry of Finance into one lake, as for 25 years a lot of data have been gathered and their number are increasing every year. And analysts simply will physically not be able to handle them. This is just one example. Another example is the Ministry of Health, which also introduces digitalization, for instance, electronic cards. Here are the huge prospects for Big Data, as this technology can understand and predict a person's disease, how to treat and so on, if all the necessary parameters are set. And there will be a lot of parameters, since all the cards are converted into electronic form, unless there is an error made by the doctor while filling the card. And it will be immediately apparent from the data, who is a good doctor, and who is not, why people die and so on.
Big implementation of Big Data will also be in the Ministry of Education.
So if we introduce Big Data on such a great scale, we will make the market more transparent, which means that the opportunities for corruption will be much less.
Another problem is that specialists in Kazakhstan go to narrow markets, where there is a lot of competition. For example, Ukrainians and Indians are strong in the development of mobile applications. And in Big Data, you can achieve great results: for example, by going to India for six months, spending about $3,000 on a trip, returning as a hard-liner and getting a salary of 500-600 thousand tenge for at least 10 years. I often go to America and on my return I am amazed how our people do not understand the global trends.
- What solution do you see?
- Start giving orders on Big Data within the country. So far, unfortunately, we do not see trust in local specialists. Good experts on Big Data will not work as consultants for a small salary in a state company; they will earn more on freelancing.
I would not say that the situation is that bad; the programs "Digital Kazakhstan" and "Smart City" are good. I believe we will be able to achieve some results in this area in our country within 5 years.
To be continued...