Nowadays, we hear more and more terms like AI, data science or data driven decision making or business analytics in a business world where the one with the competitive advantage is the one who can implement solutions with this type of technology. So what should we take into account to be able to implement it?
For this it is necessary to have data and every time we talk about these, we intrinsically talk about their preparation and maturity, it is not enough to have the data stored with the best engines or tools to store large volumes of information, it is necessary to understand how these Data represents a business domain, one face within a cube with a lot of information within a business that has many edges and needs to be understood holistically.
The tools to store the information come from something called data strategy, but prior to this we must have a north of how we want to exploit the data to turn it into relevant information for a business.
For this we cannot ignore that data and strategies for how to store it must be matured, and for this there are levels of maturity, the idea is to identify in which rung my company is in order to be able to identify what capabilities it has and what is needed to to be able to climb one more step:Level 0, without analytics: Most companies are on this rung since they have biases, it is often thought that having applications with relational databases is enough, but more than the same data, it is to be clear about how they are made up, to understand why they are organized that way and how they represent my business. To start climbing from this step, the ideal would be to be clear about the MER of the different databases, to understand what data is held in the most relevant entities and why it is stored, if it is only to have reports reactively and if it is it is capable of building domain-oriented models to be able to create dashboards or do some graphical analysis.
Level 1, with localized analytics: Being at this level is very important since we have descriptive analysis capabilities, that is, we can answer business questions about what is happening at the moment in my organization. The capabilities of this type. level is that we can separate data by business domain and understand through BI that my business has different edges that can be concatenated with basic statistics such as counts, averages and generation of KPI’s. To begin to evolve the data, doing data wralling exercises on the data, understanding and characterizing the data both in bulk and individually, and being clear about the statistics that I can generate will give way to continue to the next level.
Tier 2, an aspirational analytics organization: At this point we already have our information stored and characterized, always thinking of a clear business objective, in a domain to which my information is going to contribute, now with the historical information we will be able to explain what has happened, the statistical analysis begins to evolve from being only descriptive to also having stochastic values that can vary over time even with similar conditions, that is, we are able to see the first mathematical models and be able to convert them into even dispersed Machine Learning models that have some component of prediction or of unstructured data analysis that were previously unthinkable to have. In order to evolve at this point, it is necessary to add to the data strategy a very important concept that began to be built with DataWraling and that we know as data lineage, since at this point our data has many origins and depending on my domain. I will have applied some type of transformation to them, either to do graphical analysis, present a report or create a data set to model it, the data lineage will give me a clear north where my data comes from, and how I have transformed it to convert it into information that answers a business question.
Level 3, analytical company: Being at this level is something that very few companies have fully achieved. At this point, having strategies to enrich the data with open sources of information or with data from social networks in order to build predictive models supported by AI would be what these companies at this level already have in a mature and democratized way, that is, for When members of the organization have business questions, they know that there are these models and ways of analyzing the data to propose solutions where predicting what can happen is common. To be at this point, there are already business translators on the team, people who understand the details of a process and who are capable of analyzing it from statistics and working with technology people to create advanced models.
Level 4, analytical competitor: This level is the maximum point, it is the one desired by all and although it seems utopian, this level can be reached by following the steps and recommendations of the previous levels, but always thinking that each person in the company understands at least at a basic level , which is a model and how it could help you to take advantage of and improve your capabilities in the required business domain. Unlike the previous level, at this point there is experimentation, we no longer just want to analyze the data, we want to change what that data tells us, we want to be prescriptive in decision-making, and for this we experiment with data, with models and with new technology trends is an important piece. In order to have disruptive models, it is also necessary to understand that something productive cannot always come out of experimentation, many times inputs for future models come out or we simply find ways to answer unusual questions, to be disruptive it is necessary to understand that they will not be small changes , but long and medium-term experiments that will give us strong competitive capabilities and not only follow what the market is demanding.
Companies in general do not go beyond the first two levels because in many cases finding this data strategy does not depend only on business knowledge, but on having a strategic ally, at Heinsohn our clients see us as that ally that is strong in technology and who understands that a consultancy or building a solution is not enough, we understand that in order to move forward we have to be hand in hand learning about the business and guiding strategic decisions regarding technology.