Alexei ChernobrovovConsultant on Analytics and Data Monetization

Data analytics: what, who, how, why and how much

Today, when big data is surfing the cloud blockchain, every organization is striving to become data-driven. Leaders see Data Science as a panacea for all strategic and operational challenges. And the data specialist is considered a universal soldier with silver bullets. Why this is not true, why we need different Data Professional’s, and whether it is possible to save on expensive analytics, I will tell in this article.

Whom, when and why the data analytics is needed

Determine the most profitable areas, identify the sources of hidden costs, evaluate the effectiveness of marketing strategies, clarify the portrait of the target audience, make a plan to increase conversion - this is not a complete list of typical tasks of a data analyst. Having collected disparate information from various sources (Yandex.Metrica web services and / or Google Analytics, email distribution systems, corporate data warehouses, 1C, etc.), Data Analyst (analyst) will make short and understandable reports so that the manager clearly saw all the most important indicators of his business and could make timely necessary management decisions. For example, increase investment in those advertising channels that bring more visitors interested in products to the website of the online store, or introduce a recommendation system to increase the average check through cross-selling [1].

Data analysis will be needed when entering new markets, developing a new product or expanding the geography of sales. A detailed study of potential buyers will help to form the most effective marketing campaigns based on the real needs and capabilities of consumers. Also, management will receive fairly accurate answers to questions about the feasibility of opening a new branch and forecasts of achieving self-sufficiency.

A complete picture of historical data (expenditure statistics, sales dynamics, target audience growth, the impact of seasonality and the political situation on business, as well as other external and internal factors) contributes to a systematic vision and identification of new growth and development opportunities. Therefore, without exaggeration, we can say that today, every data manager needs any data manager.

Who analyzes data and how much does it cost

It is believed (and not without reason) that data analytics is an expensive pleasure, and an experienced Data Professional can be worth as the TOP manager of a large corporation. In particular, the salary of Big Data experts in 2018 was at the level of 200 thousand rubles - almost 2 times more than the average salary in IT at that time [2]. HR studies of the My Circle social network for the first half of 2018 show that the monthly cost of analysts of various specializations varies from 50 to 300 thousand (Fig. 1) [3].

Зарплаты аналитиков за 1-ое полугодие 2018 года [3]
Fig. 1. Analyst salaries for the first half of 2018 [3]


In the second half of 2018, according to the study of the social network My Circle, the salaries of Data Professionals slightly changed: the work of a scientist according to the data (Data Scientist) is estimated slightly more expensive, and the cost of the services of analysts of other profiles remained at about the same level (Fig. 2) [4].

Зарплаты аналитиков за 2-ое полугодие 2018 года [4]
Fig. 2. Analyst salaries for the second half of 2018 [4]


In 2019, the situation has not changed too much. For example, a survey of the IT job market, compiled by Yandex research service according to Yandex.Practicum and HeadHunter analytical service, shows the Data Scientist’s median salary of 115 thousand rubles per month. Moreover, there is a significant increase in demand for these specialists compared to last year [5]. A similar study of the foreign IT portal Stack OverFlow names the following data on the wages of analysts [6]:

  • a data scientist and machine learning specialist earn about 61 thousand dollars a year, which is more than 300 thousand rubles a month;
  • data analyst and BI specialist receive 59 thousand dollars a year, i.e. almost 300 thousand rubles a month.

As a rule, the cost of a specialist’s services directly correlates with his experience (Fig. 3) [6]. At the same time, taking into account the increased demand for Data Professional’s in the conditions of limited supply, the salaries of a scientist and engineer according to the data remain quite high, even insufficiently long experience in practical work. This observation is confirmed by Yandex research: the share of vacancies for candidates with less than a year of experience in data analysis and machine learning is 25% higher than the market as a whole [5].

Зарплаты ИТ-специалистов в 2019 году [6]

Fig. 3. Salaries of IT professionals in 2019 [6]

However, big data is not always a big expense. For example, it is not advisable for startups and small companies to hire an expensive specialist and independently deploy their own Hadoop cluster for storing and processing information. It is better to temporarily attract a freelance consultant and use cloud services for computing and application tasks. In particular, many CRM systems and Task managers (AmoCRM, 1C, Megaplan, etc.) contain built-in dashboards - visual storefronts with graphical displays of all business metrics. Free Business Inteligence (BI) platforms, such as Pentaho, Microsoft Power BI, etc., allow you to integrate data from several sources (CRM, local table files and databases) and quickly view them on any device (Fig. 4) [7] . Of course, they are inferior in quality to paid versions, but even Tableau can be a rather expensive solution for a startup.

Пример дэшбордов BI-системы Tableau [7]
Fig. 4. An example of dashboards of BI-system Tableau [7]

Sometimes BI specialists are called BI analysts. And they perform tasks in both BI and analytics. Despite the fairly broad capabilities of ready-made BI-systems, their tools are not enough to solve the problems of medium and large businesses. In particular, a data analyst will be able to formulate a nontrivial hypothesis, for example, about an increase in demand for construction products in the area of ​​new buildings and verify it by independently compiling a dataset from open sources [8]. Also, most likely, a deep integration of the BI system with other components of the corporate IT infrastructure will be needed, which cannot be solved through simple file sharing or APIs. In addition, you should always remember the nature and extent of the analyzed data. For example, if the information comes continuously from various sensors of production equipment, you should seriously think about specialized dashboards for Big Data, such as InetSoft's Style Intelligence and other analogues.

From the point of view of human resources, for the organization of medium sizes and turnovers, it will be enough to have one specialist according to a wide profile who has competencies in information analysis and marketing, understands the subject area and statistics, and also knows how to work with specialized BI-systems (Fig. 5 )

Базовые компетенции специалиста по данным
Fig. 5. Basic competencies of a data specialist

Since the volume of the analyzed data in a large company is even greater, and the IT infrastructure is richer, it will require Data Professional’s with specialized separation [1]:

  • Data Architect is responsible for the design of a new system for collecting, storing and processing data, including the features of all current and future data sources and models, their integration and presentation, as well as technical means of implementation;
  • Data Analyst hypothesizes and extracts business-friendly information from “raw” data arrays, clearing them of incorrect values ​​and outliers, and also selecting variables necessary for modeling - machine learning;
  • An ETL specialist works with dashboards and structured repositories (Extract - Transform - Load, ETL), creating analytical reports.
    data engineer (Data Engineer) creates and maintains the infrastructure of Big Data project, providing the collection, storage and management of data flows in real time;
  • ML-specialist develops models and algorithms for machine learning, and is also responsible for their implementation in software applications;
  • A Data Scientist is engaged in the analysis of information, and also develops models and algorithms of machine learning that test or refute the hypotheses of analysts;
  • Director of Data (Chief Data Officer, CDO) manages the data life cycle so that every corporate client (user, information system or cloud service) on time receives the necessary information in a suitable form and acceptable quality [9]. CDO also controls the work of all Data Professional’s: architect, analyst, engineer and researcher. I’ll talk more about the responsibilities and competencies of CDO in the next article.

In practice, one person can own several competencies and combine related roles. From an organizational point of view, analysts can report to both the general manager and the top manager of the direction, for example, the director of marketing or finance [1], because So far, not even every large company allocates a separate CDO position. Therefore, very often the analyst has to work at the intersection of BI and Data Science, which, of course, increases the burden on the specialist, but also gives a synergistic effect, allowing you to see the digital picture of the entire enterprise as a whole.


На практике один человек может владеть несколькими компетенциями и совмещать смежные роли. С организационной точки зрения аналитики могут подчиняться как генеральному руководителю, так и топ-менеджеру направления, например, директору по маркетингу или финансам [1], т.к. пока далеко не каждая даже крупная компания выделяет отдельную должность CDO. Поэтому очень часто аналитику приходится работать на стыке BI и Data Science, что, разумеется, усиливает нагрузку на специалиста, но и дает синергетический эффект, позволяя видеть цифровую картину всего предприятия в целом.

How to start: first steps in data analytics

Before looking for experienced analysts or urgently training your own employees, it is worth answering the following questions as accurately as possible:

  • why do you need data analytics: what kind of business benefit do you want to extract from the data arrays? Set clear, measurable and achievable goals, for example, to increase sales conversions by 20% or reduce financial and time costs for logistics between production sites.
  • what data you have and how much: describe the sources and nature of the presentation of the information that you are going to analyze. For example, a local ERP system, cloud-based CRM, web analytics services (Yandex.Metrica and Google.Analytics), 1C and the "magic" Excel file of the chief accountant.
  • how much your business could be managing: whether management and operational processes and procedures have been established, are their frequency and repeatability clearly observed, and important metrics and indicators are kept account of. As a rule, to determine the level of maturity of corporate governance, the methodology for assessing CMMI business processes - Capability Maturity Model Integration (Fig. 6) [10] is used.
Уровни зрелости корпоративного управления по модели CMMI
Fig. 6. CMMI corporate governance maturity levels


By identifying your goals, objects and processes, you will be able to determine the tools - systems and specialists that will make your data work, extracting from them real business benefits. Aiming at your goal, do not forget about the surrounding reality: the amount of data and the nature of the activity. For example, if a company has not yet reached at least CMMI level 3, it’s too early to talk about data analytics - you should first put the processes in order. Then, with a gradual advancement in CMMI steps, it is advisable to expand the area of ​​analyzed data, increasing the capacity of corporate repositories and the competence of data analysts.

Where to find and how to hire Data Analytics

As a rule, a good professional is not looking for work - it finds him or her themself. Therefore, you should not expect a quick and high-quality result by simply posting an ad on the job site. It’s best to “open the hunt” for analytics on their own where they live - on data analysis, Data Science, Big Data and Machine Learning topics. For example, the following online resources are very popular: Data Science Central, Kdnuggets, SmartData Collective, Cross Validated section on Stack Overflow, as well as competitive platforms Kaggle and Boosters. It will also be useful to attend offline events for analysts: conferences, festivals, championships, which are held quite often. In particular, DataFest, mitaps from Yandex,, Avito and other data-driven companies. There you will not only meet many potential candidates, but you will be able to immediately assess their experience by listening to reports and evaluating answers to questions, as well as socializing in an informal setting.

As for the necessary competencies of the analyst, it is difficult to give any universal recommendations, since it depends on your tasks, the level of maturity of corporate governance and the nature of the data. For example, for startups and small enterprises, a business analyst with marketing skills and experience working with free BI solutions, as well as Yandex.Metrica web services, Google Analytics, etc. will be useful. Advanced Excel, CRM systems, knowledge will be useful statistics and several programming languages, so that the analyst himself can write an elementary script to process information or form a query to the database.

In addition to the BI analyst, it makes sense for the medium-sized business to hire an ETL specialist who will configure the infrastructure for corporate data warehouses and integrate them with convenient dashboards. A large company will most likely need a full set of expensive Data Professional’s: from architect to CDO, including various analysts, Big Data engineers, data researchers and ML specialists (Fig. 7). In this case, your HR and IT directors will have to work hard to find free or entice experienced personnel from competing firms [11].

Потребности в аналитике данных в зависимости от размера бизнеса [11]

Fig. 7. The need for data analytics depending on the size of the business [11]


Finally, for a business of any scale, it will be very useful for the candidate to have relevant experience in a particular industry: Internet marketing, restaurant industry, retail, etc. Practical knowledge and skills mean much more than those words that are written in the diploma of your potential employee. Remember that specialized education is only an additional bonus to the real competencies of the candidate, and not a strict selection criterion.


Instead of the conclusion

Given the widespread digitalization, we can conclude that data analytics is becoming an integral part of a modern enterprise, regardless of its scale and specifics of activity. Information is the lifeblood of any business, but in order not to drown in its flows, you need to be able to single out the most important among many indicators. This will help experts in data analysis and related software products. Taking into account their rather high cost, consider the cost of data analytics as an investment that, with the proper setting of goals and objectives, will bring you much greater dividends. Hire Data Professional’s and implement software solutions gradually, depending on the level of maturity of business processes, the nature of the analyzed information and the IT infrastructure of your enterprise. Small businesses have enough low-cost cloud services, and large companies need solutions more complex and accurate, and, therefore, more “expensive” specialists. However, these investments will help businesses save time and money by solving a variety of strategic and operational tasks with the help of analytics, from marketing to production.



  1. Data Analysts: When to Hire and What Tasks to Delegate
  2. Salaries in AI: where there is more money and who they are looking for in Russia
  3. Mid-2018 IT salaries
  4. Salaries in IT in the second half of 2018
  5. IT Job Market Overview
  6. Developer Survey Results 2019 
  7. Self-service analytics in the cloud with Tableau Online 
  8. Factory analytics. 5 Steps to Implement Business Intelligence in Production
  9. CDO without AI - money down the drain?
  10. CMMI
  11. Carl Anderson Analytical culture. From data collection to business results. - M .: Mann, Ivanov and Ferber, 2017 .-- 336 p.