If you have an analytic and inquisitive mind, perhaps you have wondered what lies beneath an efficient and well-run system such as our Affiliate and Advertising platform. The part that you see is just the tip of the iceberg, and there are many mechanisms running perfectly underneath the surface, making sure you get the best user experience possible. One of these mechanisms is Business Intelligence, so we asked our expert, Igor Bucec, to answer a few questions and introduce us to the intricate world of BI.

Igor, as BI expert you partake in every development project for AdCumulus products. Can you explain in detail how business intelligence works with internet and mobile marketing systems?

“When it comes to internet and mobile advertising, the focus is on users. All marketing activities from campaign planning, user targeting and campaign tracking to campaign success analysis, strategy improvements and ad delivery optimization are oriented towards users. For that reason, we need to record as many information about users as possible. This includes information like which pages they visit, which ads interest them, from which device and location they accesses them, at what time, and so on.

In the context of billions banner impressions and clicks, we can see that the amount of recorded user data is immense. However,
all that collected data means nothing on its own. We need to know how to organize it, arrange it in a structured way, group it, implement semantics, enrich it with added value and as such, analyze it and use the results to make decisions. In other words, it is not our end game to analyze the data, but to get results, which we can use to perform smarter actions.

For that reason, to avoid wild guesses, we need a business intelligent system that contains, among other things, all necessary processes for data management, a data base structure that is most suitable for data analysis, mechanisms for timely data retrieval and analysis focused on dimensions of interest (geolocation, device, campaign, site, month, etc.), as well as various types of algorithms (optimization, prediction, classification, segmentation, etc.).”

Very interesting. Why is timely and detailed insight into data important for online advertising?

“Let us take a moment to focus on a single segment of our business: the marketing campaign. The goal of every advertiser is to reach a desired target group, and he is willing to invest money to achieve that. If the information does not reach the appropriate user at the appropriate time, the advertiser will have wasted his money. That kind of system would be worthless, because the advertiser could not reach the right target group, nor did the end user get potentially interesting information. That’s why it is important to deliver the ad at the right time to the appropriate target group.

Additionally, quick and timely analysis of collected data allows us to make better, educated decisions for information delivery, enriched with added values. Data analysis shows us whether the campaign is running as we expected, or we need to change the way we create campaigns. The timelier the analysis, the quicker our reaction can be to optimize the campaign.

Yet another example is a timely comparison of campaigns to determine which has more perspective. This can be determined based on real events in campaigns initiated by users. Detailed insight into data also provides a reliable way to detect issues with the system’s operation, and so on.”

It’s becoming more clear why business intelligence is so important in this business area. Technical principles of business intelligence are mostly independent of the industry to which they are applied. Are there any specific challenges when it comes to online advertising?

“Each system is specific, some more than others. This is especially true for BI systems, which are similar only in principle. Every BI system follows previously detected business processes. Better said, it follows what the business needs. This is where various specificities are revealed, from various specific types of complex logics for data analysis, to various types of algorithms specific exclusively to a certain business process.”

How do you approach such challenges? Is there a secret formula that makes your journey towards innovative and efficient solutions easier?

“Every new challenge has its own set of problems that need solving. The fact that we hadn’t encountered this challenge before does not mean somebody else hadn’t. That is why the first step is always to research what has already been done. We often come upon different ideas for a certain challenge, and we select the most interesting ones to test for highest potential within our system. Once we find the most suitable idea, we implement various specificities that relate exclusively to our system.

However, if we don’t find any interesting ideas, we delegate different guidelines within team members. Each of us then takes time to explore the subject and different possibilities of solving the issue based on our own perspective and knowledge. Finally, we join all our ideas and reach a decision how best to resolve the challenge. This is usually the best formula for a durable, quality solution. The “secret formula” I use is: A good feeling of forging a new value creates and builds inspiration, which finally becomes a pillar in the creation of this new value.”

Describe the atmosphere in the development team.

“The atmosphere in the team is amazing. Everyone is responsible for certain products, but when we encounter a challenge that surpasses the limits of our current environment, and they come often, we think together, share information and find a solution. Everyone in the team is enthusiastic and passionate about BI and endless opportunities it brings, with an emphasis on the respect towards the entire field.”

What exactly is your role? What does your typical workday look like?

“The role of a BI expert is diverse. I am included in every discussion regarding software architecture and design of the data model. Based on my knowledge of the source base and certain aspects of the application logic, I develop ETL processes that will finally enable filling the data storage.

Once the data is stored, I divert my attention to the presentation layer, or the user who will be performing the analysis over the database. For this purpose, I develop a Meta model with pre-aggregated information (so-called cube) which serves as a foundation for developing API methods. These methods are later used by the developer to implement attractive analytical dashboards into the application. As you can see, there is a lot going on, many different segments have to be considered and in doing so, I communicate with a large number of people.

The educational aspect is evident in presentations and workshops, which I hold for project managers, developers and other colleagues. The goal is to transfer general knowledge about BI and give insight into its possibilities and strengths. On many occasions, I serve as consultant to product managers in collecting demands and using them to model the data storage. The storage has to be filled with information, so I work with developers to consider the technical aspect of the application realisation for which we are developing the BI solution.

Due to the extensive number of areas we cover, what other words can I use to describe a day in the life of a BI expert than challenging, dynamic and interesting. Without exaggeration, there are no “typical days” for me, because that would mean that I enter a new day following a pattern of a previous day. The only routine I cherish is coming to work at a certain time, making coffee, attending the morning scrum and spending time with colleagues in rare moments of down time. :)”

What do you think is the next goal of business intelligence in the field of online advertising? Do you see a big turning point coming our way?

“The technology changes daily. Every day presents new, innovative solutions with new possibilities. This is yet another thing that endlessly motivates us to nourish our knowledge and find meaning in our field based on the principle of lifelong learning. That is why it is very important to set aside time for exploring new technologies and opportunities in the world of IT.

Considering we use Microsoft technology in developing BI systems, I must say that Microsoft invests a lot of resources into innovation and development of advanced tools that help us build more advanced BI and other systems. These systems increase our creativity and open up an endless set of opportunities.

As for big turning points, we are facing real time data processing on one side and real time big data analysis on the other side. Other things to look forward to are the advancement of the optimization algorithm and potentially crossover from OLAP to Tabular model, which will enable significantly quicker data fetching over a large dataset. Talking about hardware, we are planning to acquire an even faster server with drastically better performances than the one we currently use.”

Thanks to Igor for taking time out of his, obviously busy day, to answer our questions. We hope that now BI is not just an abstract notion, but that you gained just enough insight into the field to understand how important the unseen parts actually are.