Starting with data analytics also means being confronted with industry jargon - often complex terms that deserve some explanation. The three main ones are Business Intelligence (BI), Business Analytics (BA) and Data Science (DS). In essence, they are different approaches that have more in common with each other than they differ.
BI, Business Analytics and Data Science
What is it, what are the differences and similarities, and what impact does the choice have on your organisation? The difference between Business Intelligence (BI), Business Analytics and Data Science clearly explained.
Business Intelligence
Business Intelligence (BI) is the use of descriptive statistics for business questions. Think of questions such as: What is the average wait time at our customer service? At what point in the purchasing process do customers drop off?
- People: primarily IT professionals
- Products: personalised dashboards and dynamic reports
- Advantages: quick overview of the key figures of your company
- Challenges: depth (often) limited
- Example: a dashboard giving you quick insight into the status of your production line
Business Analytics
Business Analytics (BA) goes a step further and also investigates the why question. For example: Why does a production process stall? Why do customers drop off?
- People: analysts with extensive domain knowledge
- Products: explanatory and predictive analyses
- Advantages: powerful analyses for most business questions
- Challenges: requires much domain knowledge; with large volumes of data sometimes less suitable
- Example: forecasting revenue based on many characteristics of your customers
Data Science & AI
Data Science (DS) is the use of complex mathematical models to discover patterns in data. An example is mapping which combination of factors leads to disruption in a production process.
- People: specifically (mathematically) trained individuals
- Products: complex (predictive) models
- Advantages: capable of discovering useful patterns in large volumes of data ('big data')
- Challenges: often difficult to explain why a statistical model does or does not work
- Example: automated scanning of large volumes of old files for patterns
Which type suits your organisation: BI, Business Analytics or Data Science?
Characteristics of Business Intelligence (BI)
For managing any organisation, key figures - internal data - serve as the starting point. You look back in time and know how your organisation has performed. BI tools are aids for retrieving those key figures 'from the organisation'. Such tooling provides much information including KPIs (key performance indicators). The data is visually translated into easily readable conclusions - often dashboard-quantified results - that are moreover automatically updated. A powerful tool, therefore, for keeping sight of your organisation's performance.
Strengths of Business Intelligence
BI is highly suitable for answering the what questions: what is the revenue, what is the number of customers. For deeper questions such as why and where to, BI is not suitable. If you want to know why customers are dropping off or which direction revenue is moving over a certain period, you will need to work with a more advanced form of data analysis - so-called Business Analytics. A step further still is the use of Data Science solutions, for example via machine learning or artificial intelligence (AI).
Weaknesses of Business Intelligence
BI solutions are often picked up by the IT department and therefore by IT professionals. They generally have much understanding of technologies and less of the business side of things. In practice this means that technical dashboards are often created that do not connect well with the daily practice of subject-matter staff or managers. Dependency on the IT agenda also arises - and especially the lack of space within it. Therefore, if the organisation is large enough, it can be worthwhile setting up a 'data department' (a data lab).
Analyse your processes and predict the future with Business Analytics
Characteristics of Business Analytics
Knowing why something happens goes further and is considerably more complex, because multiple variables play a role. Business Analytics (BA) is focused on analysing data to gain insights into the processes to which that data relates - answering the why question, therefore. Sometimes you also need data from outside your organisation - this tooling is also extremely suitable for that.
Business Analytics helps connect conclusions, recognise relationships and trends, and make predictions. This way, your organisation lays the foundation for new knowledge, insights and strategies.
Predictive value
With Business Analytics you gain deep insights about your own organisation. For example, a detailed picture of the behaviour or wishes of your contacts. This allows you to make accurate predictions about their behaviour in the near future - so-called predictive analysis - and Business Analytics shows where new opportunities arise. Business Analytics thus provides you with predictive values based on internal (and sometimes also external) data.
Advantages of Business Analytics
- Fast and current: Better insight into performance so you can adjust course more quickly.
- Well-founded: Substantiation of decisions through better information.
- Reliable: A step ahead of your competitors, because you take decisions based on well-founded information.
- Competitive edge: You have better internal insight into your own processes and external insight into the market, so your organisation achieves both tactical and strategic advantages.
Automate decisions through Machine Learning & AI
Machine Learning and Artificial Intelligence (AI) are methods in which computers perform tasks for which humans use their intelligence. This concerns matters such as interacting with the environment, analysing and reasoning, problem solving and making predictions. Two things are necessary for this: a programmed algorithm - a series of instructions that must lead to a certain result - and data.
Data is the raw material that makes the algorithm work. By analysing ever more data, the algorithm also functions ever better. Its performance is only as good as the quality of the algorithm and the input - the data - with which it is fed. The algorithm takes decisions based on what it has learned. It can be programmed to continuously collect and process new data, making it 'self-learning'. This makes AI very useful for organisations handling large volumes of data, where gaining insight into relationships, patterns and labelling does not happen 'by itself'.
Points of attention
For all data-driven solutions: garbage in = garbage out. If the input data is unreliable, unreliable conclusions will also emerge. This also applies to the reliability of the algorithms themselves. Understanding the algorithm and insight into data quality are essential.