The Five Steps To Getting Wise With Data
Step 1 – Value your data as an asset
Most organizations consider their main assets to be their people, products, customers, IP / trade secrets, etc. Rarely do they consider the vast amounts of untapped potential in the data they have collected over the years. With the right techniques and methods, that data can make you more responsive, efficient, competitive, and profitable. These techniques and methods are broadly referred to as Business Analytics, and are leveraged by many of the largest organizations around the world.
Step 2 – Understand the role of Business Analytics
Business Analytics is broadly the application of internal and external data to solve business problems, improve service, or gain a competitive advantage. It involves using a set of methodologies to extract operational and external data, clean, integrate, transform and apply iterative data exploration and statistical methods. Insights derived from business analytics can be descriptive (current / past state), predictive (future state) or prescriptive (actions for desired outcome) in nature. It also involves planning for the acquisition of additional data resources, such as surveys, focus groups or external data. Those insights must be actionable to have value. Frequently insights such as predictive model scores are incorporated into automated decisioning systems.
Descriptive, Predictive, and Prescriptive Analytics
Analytics can be thought of as one of 3 distinct types:
- Descriptive Analytics is rooted in past or current state. Examples include sales performance, inventory measures, correlation maps, link analysis (graph theory), KPI’s, and dashboard reporting. This is the foundation of wisdom, as you must understand exactly where you’ve been and where you are before we can improve your future.
- Predictive Analytics, also called machine learning, uses historical patterns to predict a future state at a macro level (forecast model), a detailed level (predictive model), or grouping level (segmentation / classification model). Predictive analytics can also be used to explain how and why things are the way they are, such as separating out factors that impact your sales in a particular month.
- Forecast models are used to provide insights into the future by predicting likely aggregate outcomes at points in time, such as total sales by store, call center volumes, or service level demands for the next few months. By providing likely outcomes along with confidence intervals, organizations can plan and staff appropriately, saving money and improving customer service.
- Predictive models provide specific estimates for each entity, customer, item, or event. Examples include identifying customer churn or credit risk, likeliness to respond to an marketing offer, likeliness of item to fail, and predicting a customers age or income. Insights into these areas can help make personalized or customized decisions about what treatment, if any should be taken.
- Segmentation / classification models help to separate entities, such as customers into a number of different segments that behave in similar ways. It is a technique often used for marketing purposes, to better align product offerings to customers or to provide recommendations to a customer.
- Prescriptive Analytics uses the insights from the data in order to actually prescribe the optimal solution to a given problem. There are two main types of prescriptive analytics, operations research and test and learn strategies.
- Operations Research (OR) is a way of optimizing the way in which resources are allocated to complete a task. It looks how operations are conducted, including understanding constraints such as materials, human resources, and time, and identifies the optimal way in which to assign resources to achieve a goal, such as minimize costs. Solutions that incorporate OR concepts are often associated with Optimization.
- Often the highest form of data wisdom, and often the most difficult to achieve; test and learn strategies involve running controlled tests or trials to obtain information that can lead to the optimal outcome of the goal. Using controlled trails allows for an unparalleled way of obtaining knowledge about your business. For example, with a single structured test, you could understand what is the optimal packaging layout (font, shape, colour, messaging, etc.), and pricing for your product to drive the most sales and profit.
Business Analytics vs Business Intelligence
In contrast to Business Analytics, Business Intelligence (BI) as is commonly termed, is a sub-component of Business Analytics that focuses on reporting of current or historical state of business (descriptive analytics). Typically BI is delivered through tabular, graphical or dashboard style reports.
Step 3 – Understand the role of Data Scientists and other analytics resources
The term Data Scientist has often been misused, causing considerable confusion in the market. A true Data Scientist is person who has the capabilities to understand a business problem, build a strategy to solve it, and execute on that strategy. They need to be both business and technically savvy, understanding the entire analytics life-cycle; problem definition, strategy, data acquisition, data quality and cleansing, data integration and transformation, analytical model development, presentation/reporting, etc. Data scientist need to be creative problem solvers, as data science is as much an art as a science. Resources that possess this broad level of skill are extremely rare, but their input in analytics projects is often critical to achieving the best outcomes.
In addition to data scientists, there a number of specialist in the some of the most demanding areas of Business Analytics such as:
- Project Managers
- Business Requirement Analysts
- Data quality and cleansing specialists
- Data Integration and Transformation (also known as ETL) specialists
- Data warehouse and data modelling specialists
- Analytical model development specialists (also known as applied statisticians or machine learning experts)
- Reporting / BI specialists
At WiseWithData we have data scientists, like our founder Ian J. Ghent who lead and coordinate with a number of specialists we work with. Depending on the size and nature of the project, we can align the right resources, be it within WWD, external talent (carefully screened by WWD), or internal to your organization.
Step 4 – Being ready for the analytics journey
Becoming WiseWithData is not an overnight process, it takes months or years of careful planning and execution to get the most out of your data. It involves aligning people, processes, and technologies to truly leverage the power of your data. That being said, WWD has extensive experience at identifying opportunities to get quick additional value out of your data, regardless of how far along your organization is on the analytics journey. Those that incorporate analytics into the core of thier business, have seen profound positive changes across thier organizations.
If your just starting out on your journey, the insights and results that WWD can bring about in your data will be eye opening. For those further along the journey, WWD can show you how to accelerate and get to the next level, achieving great business results along the way. The following chart is a rough guide meant to help you understand where your organization is in the on the journey.
Step 5 – Plan and execute
At WiseWithData, we offer comprehensive expertise to help you develop the right plan and strategy for your organization. Whether you just need some advice on getting started, or would like us to provide resources from start to finish, we offer a full range of services to get you the most from your data.