Business Intelligence, BI, Data Lake, Artificial Intelligence, Big Data, Smart Data, Machine Learning … are terms that are in all the roadmaps of companies, large, medium or small. It is a trend, it is present and it is future.
But do digital marketing professionals really understand the depth of the term?
If we Google the words Business Intelligence or Business Intelligence and read the Wikipedia definition we will find some dissonant concepts in digital marketing such as:
Some simple concepts and other complex concepts that make it difficult to really understand what BI or Business Intelligence is, its dimension and why we must apply it in the company, in this case the digital marketing area being the main driver.
Because do not think that it is a modern concept or a passing trend.
BI was born in 1958 from the IBM researcher Hans Peter Luhn, who was the first human being to use the words Business Intelligence. Then, it is true that the concept has evolved from the DSS (decision support system) of computers (see that one feels old when saying this word), to the data warehouse and OLAP cubes of the 90s to platforms current cloud and open source such as Power BI, Tableau or Microstrategy …
BI stands for business intelligence or applying intelligence to business.
Intelligence is learning, understanding and reasoning to solve complex problems and make decisions.
But not just any decision but the best possible decision, with a specific meaning and in a specific context. We would really be planning what we believe will be the best solution to the problem posed.
Therefore, BI seeks through data analysis to find patterns that help us to provide a solution to the problem solved. Does this sound like something to you? Of course, because in digital marketing we monitor everything that happens, design dashboards in real time and make decisions based on the data, even if it is only historical data.
Whenever we talk about Business Intelligence we have to be very clear about these concepts in order to see the scope of action.
The forecast is based on facts and assumptions about past and current performance. It is what all companies have been doing for many years. Make decisions based on historical data.
For example, from my client portfolio, those who have bought this product or services with this average price in this period of time, I will analyze them to launch a personalized promotional campaign.
Refers to a calculation or estimate that uses data from past events, combined with recent trends to obtain a result of a future event. In this sense, we could take the data of the same customers from the previous example, but adding variables such as trends in the purchase of similar products in order to launch a cross selling action.
It is the act of indicating that something will happen in the future with or without prior information. Here things are complicated but for good. In this case, hypothetical action scenarios would be designed for certain clients who meet certain requirements. That is, for customers who will leave the mark for the next three months, have pre-designed actions to retain them before they leave.
BI tries to scale decisions from forecasting, forecasting and forecasting
When we talk about Business Intelligence, concepts such as Big data, big data or data intelligence are quickly associated. True or not?. The concept of Big Data is a term that was born in the 90s and seems to be from the hand of John Mashey. Although what we understand today as Big Data accelerates exponentially from 2012. The reason, the enormous volumes of digital information, going from megabyte to gigabyte to terabytes, petabytes, exabytes, zettabytes and yottabytes.
I don’t know if any of you will remember the 1028 bytes, we have already raised them to the eighth power!
If we talk about Big Data we get into the mud of artificial intelligence, machine learning and the use of complex algorithms. Big data allows you to store, classify and analyze large volumes of data. Data that must comply with the “6 V” rule:
If they do not meet 6 we would not be talking about Big Data, that is why many companies still do not really do Big Data.
Does the case sound like the 2 Boing 737 Max that were injured in Egypt and Indonesia? Or the IBM Watson megacomputer? Two example of use and application Big Data of book.
When we talk about BI we must know that there can be Big Data or Smart Data. In both cases we can and must apply BI methodologies.
The key question to answer would be, really, who makes the decisions? Normally people are the ones who make the decisions, every day, in these decision wars there are people who make good decisions and people who make bad decisions.
If we are tying the strings, intelligence then has to do with decision making, and decisions have to do with people. Because decisions are the answer to a good solution.
So far everything is correct, but how do we build a good solution? Here is the magic formula:
Intelligence + people + information. This is the basis of BI.
Information helps us find the answers to the questions given, and the answers help us make decisions.
In this sense we can group the information into 2 types of data: structured (Word documents, Excel sheets, database …) and unstructured (videos, social networks, email, post …).
If we transfer this to the business world and digital marketing, to be able to answer business questions such as:
What product will give us more profitability next year?
What is the digital audience that will convert the best?
How can I improve the conversion rate of my PPC campaigns to get more ROI?
Companies have to do Business Data, that is, work on business data. The first step to do this would be to collect and store all the data on products, customers, sales, margins, buyer people, web visits… at this point the digital world has it easy.
The second step would be to dive into the data, analyze, to convert the data into information. In this analysis we must find insight.
Insight, that magic word in BI. An Insight is the sudden understanding of a complex problem that arises from observation, analysis, and data searching.
BI tries to detect “patterns” that help us solve a problem. It does this through Data Discovery.
We could say that BI would consist of having access to business data + data management to increase sales. In other words, analysis and reporting. How many of you are currently in this phase?
This is called BI 1.0 and was born in 1989 !!!!
With the digital takeoff (the famous www) the volumes of information grow exponentially and the BI needed to be fast and adapt. In this context BI 2.0 emerges
– Online and ofline access
– Custom, cloud, private and open source solutions
– Market consolidation ( Gartner 2019 magic quadrant offers us more than 50 private BI tools and opensource).
– More intuitive and easily accessible for any user
– Less technique and less dependence on IT.
– More visual, more intuitive and better results
If we talk about professional methodology, a BI project needs at least 5 steps or implementation phases:
Collect the data and that they are reliable. For this we must analyze the data sources, the type of data, the veracity of the data and the format of the data.
Transform data into information, adding context and meaning. For example, if we talk about total sessions, user quality, conversion rates … these are data that need to be put in context and meaning. If we add that they are the data from the last Google Ads campaign for the Mirando al Mar property development in Malaga, we begin to understand and interpret them.
Information in knowledge. That data with context and meaning we have to transform into knowledge. That is, analyze them and give them relevance to be able to make complex decisions with them.
Knowledge in insigths or patterns. This is the most complex part but there are already complex algorithms and tools that help us. How we discovered that pattern that will give us the solution. Imagine in the previous SEM campaign that we detected that whenever Google Ads gives a specific combination of keywords, in English, on certain days of the week, in some cities in the UK, the conversion data is better and the highest quality leads generated.
The insigths in profitable business decisions. Measure again. What would you do in the previous situation ?, at least invest more (double or triple the investment? And / or apply modifications to the campaigns and digital strategist to be more efficient. Here is the true value of BI.
These 5 steps are the essence of any BI project applied to digital marketing.
The Bi puts the focus on people as we have already seen; And digital marketing focuses its strategy on people, as we have reiterated on many occasions. It is what we know in digital marketing as Customer Centric
For this, every brand that invests in digital marketing must start the strategic path of Data Driven, that is, turn its current model to put the data within the business strategy.
BI in digital marketing is nothing more than analyzing the right data at the right time to the right people to make the right decision on the issues at hand.
Where will the difference be in looking for the quality and reliability of the data, in the depth and efficiency of the analysis and in the interpretation of the data to launch digital actions, measure again and start again, with a focus on business profitability digital.
In short, it is about monitoring and analyzing all:
And apply the magic formula for digital marketing: Data + insigth + Action
With the application of this formula we will go from designing forecasting models to prediction and prescription. Or mixed. Having a deeper knowledge of all online business data, really knowing our client and their circumstances, to be brands that connect with people.
Let’s imagine that we can cross all the data of the Social Ads Campaigns of our brand, with all the data of the portfolio of clients who have bought the products of the promotion of the last 2 years. What kind of customers have bought, when they bought, where they bought, how was their relationship with the brand once they purchased the product, behavior of online versus offline customers, average geolocation purchases, problems found in the process of purchase and shipping …
With all these data, Social Ads campaigns could be further refined at the level of audiences and micro-publics, copies, products, locations, interests … to re-measure and assess results. Reducing investment and maximizing sales.
To end the post, I want to delve into the darkest part of all this explosion of data, BI and Big Data, machine learning and artificial intelligence (AI).
It is true that the use of real-time BI for decision-making and the design of predictive models is increasingly advanced. Complex algorithm training helps find solutions to increasingly complex problems. So far phenomenal, but we cannot ignore the risks that must be controlled.
– Whoever designs the algorithm (if it is black box especially) is the one that conditions the final answers. It is possible to manipulate the process from the beginning with good or bad ends. For this reason, white box algorithms are advocated where the entire process is transparent.
No to black box algorithms and yes to white box algorithms.
– There is no single version of the truth. Whenever working with data, this statement is universal. You can see the data one way (to do evil) and another way (to do good). But remember the recent case of the US elections.
– To date, machines cannot make relevant and complex decisions at the neural level, where not only data but intuition, experience, improvisation come into play … but to see the recent cases of accidents by Tesla and Boeing, where artificial intelligence took the final decision over the person and ended in two fatal accidents.
– And what about ethics, protection and good data governance. This debate is on the rise especially on a digital level, to the extent that the integrity of people and their private lives is at stake. As digital marketing professionals, we must advocate setting limits, putting the person really at the center, and putting the data custodian to good use.