We have already written about how *instinctools conducts the Techtime project and gathers experts from different companies to discuss trends in the IT industry. Today it’s the turn to talk about Business Intelligence. As it turned out, this topic is still hotly discussed not only in IT companies, but also in large corporations and businesses. Gradually, Business Intelligence is moving from the “fashionable” options category to an indispensable competitive policy component. But at the same time, the issues with choosing a tool and the BI specialists effectiveness remain. We have collected the experts statements on the modern BI role that were raised in the Techtime project discussion in this article.

The discussion was attended by Data Engineer, Microsoft Gaming Dmitry Anoshin, Head of BI, Business Intelligence expert Roman Bunin, BI Engineer, Currency.com Anastasia Kotova, BI Analyst, *instinctools Arthur Vysotsky.

“BI hits a productivity plateau”

It seems to me that now BI is reaching a productivity plateau, no company can live without BI, everyone understands the importance of this area very well. I have noticed a trend: before BI was a fashionable toy, and now this trend is not only in large, but also in small companies on the market. Any company that employs more than a thousand people has some kind of BI tool.

Roman Bunin:
Photo: wordpress.com

For me, BI is an important attribute of any company. Absolutely everyone understands the analytics value today. And BI is the only convenient tool, it allows you to build a bridge between the data that is generated in the company and business users. But here another question arises: how to get people to use this tool?
Even if companies implement expensive tools, they still use this tool to upload data to Excel. For preventing this from happening, people need to be trained and encouraged to use BI tools correctly. For example, in Canada, I came across a Tableau with interactive dashboards implementation that specialists did not want to use.
They just said that the skills to use these tools were not required at the time of the employment contract conclusion, the employer did not register BI tools in the job description. And that was a big problem. But today it is difficult to work effectively without BI tools, especially speaking about analytics and data storage. And the BI tools introduction should take place before we start working with Data Science or ML. Let’s just make dashboards and teach people how to use them.

Dmitry Anoshin:

BI is becoming more popular, companies are looking at their competitors and perceive data analysis tools as an opportunity to improve their position in the market.

Anastasia Kotova:

Amazon and Microsoft don’t use mobile BI apps

I have been working with BI since 2010. I have already known about SAP Business Objects since 2011, that is a popular BI tool that is still used in many banks. The last version of this tool came out 5 years ago, but the solution is still in use today. It even has a mobile client for smartphones. This is a cool story if you imagine that the data will be available at any time. But this rarely happens, we always face restrictions due to security requirements. As long as I have been at Amazon, Microsoft, and other companies, I’ve never seen people who have actually used mobile BI apps.

Dmitry Anoshin:
Photo: labsnews.com

So the story, although not new, has not “landed” yet. Perhaps it would be interesting if there would be some vendor with a focus only on mobile applications. Or complement existing vendors with a mobile app. For example, the same MicroStrategy company has always focused on mobile analytics. But they have switched to the custom applications side. I have never worked with their products myself, but I know that they have fully customizable e-commerce offerings.

Photo: YouTube

ML needs to be made available to everyone

I see two ways to make ML accessible to everyone. The first is the ML use by the vendor to make it easier to work with the tool. Nowadays there is a high demand for developers for adding some algorithms to the product itself. The simplest example is when users build a data model in Tableau and drag tables, and links are built using the ML algorithm. The AI ​​will scan the tables, figure out what type of data is being used, and build connections. And the second way is the ML tool usage by the users themselves. The most popular example is a forecast. In Tableau, we can take historical data, drag a time series onto a time series, and overlay a forecast model. And ML will build us a sales forecast for the next year. At Amazon, this was a great decision, management required everyone to have a forecast for a year, half a year or a quarter. Of course, it is possible to use Python and R inside the BI tools. The ML use may differ. For example, ML may have already been built into a BI tool, but there is also an option when the analysts themselves can add it.

Dmitry Anoshin:

“Using complex tools in BI leads to complex infrastructure”

The complex tools use in BI leads to a complex infrastructure. In my opinion, it is more correct and efficient to use BI only as a frontend tool. Thus, we render what we have in the database. As soon as we start moving a lot of logic into BI, we lose control over it. I like the approach that turns BI into a coding tool. You can “feel” and describe the code, see the latest changes. When there is an external part, a visual one, you can still commit the code and be sure of the changes. It seems to me that while there is no normal change management on the BI side, it will be dangerous to store business logic in it.

Roman Bunin:
Image by Pexels from Pixabay

“BI must be one of the components in the business chain for bring money”

It is foolish to expect that if you install a BI tool for yourself, then the money will flow like a river. It’s just a dashboard that helps you understand where you’re going, but it doesn’t guide or manage the business. Anyway, decisions are made by a specific person. If we put a lot of devices in the car, it will not go faster. BI should be linked to company performance metrics. I know an example of companies where projects aimed at building efficiency have always started with a dashboard. We have a process, we first make a dashboard for it, only after that we optimize the process. And it is good practice. BI must be one of the components in the business chain for bringing money. My favorite question when I make dashboards is what business decision will you make or what will you do when the number changes? If a person does not know the answer, then most likely the dashboard will not help him.

Roman Bunin:

There are cases when BI solutions do not help business. BI does not need to shift the responsibility for the business. BI solutions are just a help in understanding the big picture at different levels of detail. In addition, you need to understand that BI can be implemented poorly, the data warehouse is built incorrectly, the data loading into reports is poorly organized. Then there is the visualization question, how information is presented in reports, how well the employees are prepared and can analyze the report correctly.

Anastasia Kotova:

If you are a “terry” enterprise, then you should pay attention to the upper right Gartner quadrant

When we say which BI tool is best to use, it all depends on the goals and objectives. If you are a “terry” enterprise, you have a lot of money, you are ready to invest in tools, then you should pay attention to the upper right Gartner quadrant. It’s well-established, has support, understandable tools, and access to most databases. Speaking about new tools, trendy BI systems, it seems that they are well suited for some specific cases. There are now many niche BI systems, like ThoughtSpot, that are popping up in small niches. They are tailored to specific tasks. Recently, a colleague shared an interesting Narrative Science tool with me. It is created by a team that makes BI for storytelling. It is a narrow direction of small BI that can work for those who need to tell stories in the form of long reads. It turns out that instead of BI, we are reading an interesting article in the NYT. If your task as a company is to use an enterprise solution, you need to go up to the right in the Gartner quadrant. If you have narrow goals and you can find BI to help solve them, then it is better to look at the bottom of the quadrant. A startup can focus on open source tools. They all work great and solve most problems.

Roman Bunin:

I’ve always liked Gartner, I’ve always followed it. For me, it was the only tool where I could see all the solutions available on the market. And if I wanted to understand their pros and cons, categories, it was always easier to turn to Gartner. For your understanding, I divide BI tools into three categories. The first is Enterprise BI (e.g. Oracle). These are very heavy tools that get the job done. And if you have 10 thousand users, then you can restrain their “onslaught” very well, distributing them by access level. But such solutions work slowly and ugly. The second category is Tableau, Power BI. For me it is best to use Tableau. And it’s not even about functionality, I just enjoy working with Tableau, but do not enjoy with Power BI – that is a matter of taste. And the third BI tools category is cloud-based. These are Looker and others that are 100% cloud-based. And these tools have their advantages.

Dmitry Anoshin:

Gartner provides a good rating for BI tools. And its advantage over similar ratings is that it is not seen in lobbying the tools holders commercial interests. No one sponsors this rating, like ratings, for example, sponsored by Microsoft, where it solves the problem of increasing interest in its products. And Gartner is a company that hasn’t been seen in an overt advertising lobby. And this will allow us to talk about confidence in the agency ratings.

Anastasia Kotova:

One of the problems in IT is the HR positions names for BI

I have divided BI specialists into two categories. There are BI analysts – these are those who prepare the data and make visualizations. They work at the professions intersection , to some extent, such people belong to the data engineer specialization , as they know how to prepare data well. To some extent, this specialty is closer to frontend development, as good visualization is created. In my opinion, nowadays most of the vacancies are just about that. But since the mentioned specializations are different, it is quite difficult to combine them. There are very few cool specialists who can both prepare data and visualize well. However, now there is a separate, dedicated specialization – a BI developer. Within this specialization, people are engaged only in data visualization. This area is new and interesting, as it proves that the business has understood the dashboard tool complexity. Dashboard is one of the most complex applications, there are a lot of charts, colors, it is difficult to make everything clear and fast at the same time. And the latest trend in the labor market is the vacancies for data visualization emergence.

Roman Bunin:
Photo: blinkux.com

One of the IT problems is the positions for BI naming . From the job description, you can understand that BI is required only by mentioning specific tools. And so vacancies are aimed at finding analysts. An analyst is generally a universal position that can suit a BI developer or an ML engineer. For some reason, when a job description for a BI is published, they are very similar to each other, they constantly lose their essence and meaning. For a BI developer, the main function is to understand what the business wants, and with the help of BI tools, give the customer the possibility to obtain the necessary data independently. At the same time, installing and updating BI tools is a BI developers function.
Just tinkering with a server that is constantly breaking down may be interesting, but such specialists’ value is not growing. A BI developer must be able not only create dashboards and reports, but also understand business requirements. Its task is to act as a bridge between IT and business. That’s the BI work beauty, you’re not very deep in technical details and not very deep in business. If we know one BI tool well, then it is easy for us to master another tool, since the their work principles are similar. For me personally, the stopper was the requirement for BI developers to give business recommendations. There is already work with data analysis, you need not just build a dashboard and look at the indicators, but write to the business: they say, this product category is not for sale with us, let’s remove it.
It has always been difficult for me, especially when I worked as a marketing analyst and BI developer. I was given the task to analyze why a page on the site does not work well. Here are all the data and metrics, but it was very difficult to rewire the brain and understand what customers are doing on the site.

Dmitry Anoshin:

“Business is not happy that people jump from place to place”

Now in BI you need to be very proactive. It doesn’t matter what you do, you need to be as active as possible and try to take on everything. After all, from the very beginning it is not clear what works and what does not. You were hired as a BI developer and asked to build a simple report. And there are two ways: ok, I have built a report, this is where my work ends, and the second – if we use Power BI, then I will sign up for Power BI courses, download a book on this tool, and so on. My biggest piece of advice for anyone starting out with a new tool is that you should learn 120% about the the tool functionality.

Dmitry Anoshin:
Photo: imgflip.com

I liked projects where you had to implement a BI tool from scratch. Typically, such projects take a year or two, after which you can go to work in another company.

It is unpleasant for business that people jump from place to place. But here you need to understand that if we want to increase salaries, then every 2-3 years we need to change jobs. It doesn’t work otherwise. Everywhere there are corporate rules that prohibit salary increases by more than 10-15%. If you were a junior BI developer for 70 thousand rubles, then it will take you 3-4 years to crawl up to 120 thousand Russian rubles. But if you quit after a year, you can get a new job with a salary of 150 thousand.

But it is necessary to have a beautiful story upon dismissal, so that we have figured out the product, got certified. It’s good to have several positive cases at hand according to the “before and after” formula. And having this, a BI specialist can safely enter the labor market.

Volha Abukhouskaya
Volha Abukhouskaya Recruiter
We will be glad to see you in our team!

    * Full Name

    * E-mail

    * Job Vacancy

    Additional Information (LinkedIn, Github, etc.)