In the digital era data analysis is not even a necessity, but an everyday task of any company. The effectiveness and efficiency of decision-making processes has a key influence on whether the organization is successful or fails. The use of advanced analytics in order to obtain the best possible recommendations in business processes has become a new and very effective tool of competition. Starting with data processed by the organization and data available from external sources, the company may gain a unique market advantage. The method that guarantees the achievement of this goal consists in following three main rules for the development of an analytic platform in the enterprise.
Any process of using data analysis to support decision-making is based on the same generic steps. We call them analytics lifecycle.
The journey through the cycle begins with formulating a business hypothesis (ASK) or simply a problem which we would like to describe with data and solve. Having defined the objective, we start with data collection and processing with the use of data cleaning and integration tools (PREPARE DATA). And there the real adventure begins – through exploration and modelling (EXPLORE, MODEL), we find a solution, often coming back to the starting point, and obtaining new data or its different configuration, so that the process of building an analytic model can be more effective and guarantee better results. Having achieved a model i.e. the decision-making formula underlying our business processes, we start the implementation (IMPLEMENT, ACT). Here we set off to use the conclusions drawn from modelling in our actions – we take advantage of the discovered interconnection and decision-making rule in production. At this stage we cannot forget about assessing the quality of the model and the results of the changed enterprise operation. It is crucial to regularly check whether the interconnection discovered in the data is still valid and whether the analytics-driven decisions bring the desired business results. If not, we should come back to formulating a hypothesis and a business problem and start the cycle again. What should we do then to make the application of analytics guarantee that we gain market advantage? We should remember about three key principles of competing on analytics.
Principle 1: Efficiency and speed
Changeability in business is a condition, not a one-time event. We can be sure that a change will come and in the digital era we live in, the changes happen fast and unexpectedly. They often result from the actions of competitors or, simply, changes in our customers’ preferences. The first principle of competing on analytics says that the described above analytical cycle must be executed very efficiently, so that the company operates effectively and is able to quickly adjust to the market changes. If analytics is used to support marketing, this means that the recommendation model defining the customers’ tendency to purchase new products should be updated on an ongoing basis and if a competitor introduces a new offer, we must be able to react within a few hours, not days. Speed guarantees us a competitive advantage.
Principle 2: High granulation
Nothing is the same in our business and generalization does not provide us with the opportunity to obtain higher margins and effectiveness. It is similar for our customers – nobody is the same and, much as acting on customer segments is easier for organization, nowadays the real value is guaranteed by the ability to build analytic models dedicated to individuals. The second principle of competing on analytics indicates that we should be able to build dedicated mathematical models for individual customers (this strategy is also referred to as “hyper-personalization” or “segment-of-one”) or, e.g. for individual product items. If analytics is used for forecasting demand, such information at the level of the whole product group allows us only to foresee the market potential, but once we have build predictive models for each Stock Keeping Unit (SKU) in each distribution center, we will get information which allow us to guarantee high availability of the goods, at the same time optimizing the use of warehouses and reducing the costs. Then, analytics really gives us a market advantage.
Principle 3: Automation and adjustment
In order to address the first and the second principle of competing on analytics, we build IT solutions supporting data processing, comprising advanced algorithms of data analysis and allowing the disclosure of decision-making rules, so that they can be used in business processes and transaction systems. The third principle of competing on analytics indicates that, while building this world, we should understand that an analytic process actually consists of two integrated cycles.
Starting with a business hypothesis, we begin to look for an answer to our question (DISCOVERY). At this stage it is crucial to be able to quickly reformulate the task, change the data used in the analysis and change the approach adopted to solve the problem. This stage is full of experimenting, unsuccessful trials and tests. It will also be crucial to achieve high agility and low time-consumption in the process of data exploration. When a mathematical model or decision-making rule is ready, we proceed to the stage of using the gained knowledge in production (DEPLOYMENT). Here the rules are different – the model is subject to governance, SLA of the decision-making environment is high and there is no room for tests. For example: when we look for new patterns of customer behavior or when we look for new chances for complementary sales of products, we use Big Data environments, combining various items of data, which at first glance are unrelated. In this respect, Data Scientists and analysts exploring various business hypotheses appreciate the possibility of adding new data to the analysis, its mass transformation and experimenting. Once they have discovered the customer behavioral patterns, the rule describing them becomes the basis for recommending new products to these customers. It is used in the online store, call center or self-service website in electronic banking. Everything must work correctly, with a high degree of control. The third principle of competing on analytics says that the key to building a competitive advantage is the automation of the implementation of new analytic models in business processes. The time between building a new analytic model and its implementation in production should be as short as possible; optimally, it should not require the transfer of the model between IT platforms and interference with the code which describes it. This way we become able to automate our actions and, eventually, to automate the model-building process, which translates into the ability to address both the first and second principle of competing on analytics.
SAS and principles of competing on analytics
SAS solutions and tools allow us to address all the described principles of competing on analytics. A wider description can be found here; at the same time we invite you to watch a video which presents in a nutshell the three principles discussed here.
In addition, we have developed a survey that explores theoretical potentials of data coming from the internet of things (IoT). Internet of Things: Visualise the Impact gives you access to the lessons learnt by 75 executives from various industries all across Europe.
Pingback: My top analytics books - 1. general business analytics
Pingback: My top analytics books - 2. How-to guide to business analytics