Chapter 18
IN THIS CHAPTER
Highlighting business goals
Forecasting business needs and managing inventory
Allocating company assets and resources effectively
Making better and informed decisions
Exploiting new opportunities
Navigating hidden risks
Increasing returns on investment
Business competition is fierce and global, which leads companies to employ everything in their power to survive and thrive. In such an environment, companies are seeking to increase their revenues while keeping the operating cost to a minimum.
Predictive analytics is a business enabler that helps accomplish the very essence of increasing companies' return on investments. If implemented successfully, it will drive your company's profits higher. It will help you turn your data into valuable information you can capitalize on, and it will give you the competitive edge you need to outperform your competition.
Predictive analytics enables companies to use the massive data they've accumulated from their operations as a refined resource to advance their businesses.
Predictive analytics can help you solve many business problems more effectively, but the crucial step toward that advantage is to identify specific business goals so you can choose the appropriate analytical tools to achieve them.
The process is twofold: The business goals you define help you build your predictive analytics project, and the completed model helps you measure how successfully your company is moving toward the goals. At the heart of this process is the development of your predictive model by running real data — both historical and incoming — against it.
Set project goals that you can accomplish in a relatively short time frame and can measure — concrete steps toward the company's general vision and overall mission. Depending on your current strategies and which phase of market development your business is in, your business goals may change — which is why it's important to evaluate your project goals periodically and adjust them accordingly. Here are examples:
If you seek to identify fraud, then the model should be designed to rapidly and accurately evaluate whether a transaction is fraudulent — and react accordingly.
Industries that already use predictive analytics to detect fraud include healthcare, finance, and insurance.
The quality of your predictive models is dependent on the quality and relevance of your data. Predictive analytics explores your data through well-defined algorithms and techniques, mines the data, searches for hidden patterns or trends you were unaware of, and uncovers valuable facts about your business. But you have to feed good data to the model to make that happen.
Data can be a big driver of predictive analytics. To ensure that your data is the best resource possible, your organization needs to know
Poor data quality — generally incomplete or irrelevant data — can skew your predictions' results and lead to much wasted energy and work by your experts to distinguish between real patterns and false signals and noise in the data.
The model helps you improve the quality of the data; the better the data, the more useful the predictions. The better the predictions, the better you know your data. One feeds into the other.
Knowing your data will lead to better predictive analytics. Fortunately, establishing a predictive analytics project will lead to better knowledge of your data. In the course of that process, the output of your model helps you uncover relationships within your data that you were unaware of — giving you a much firmer foundation from which to embark on the predictive analytics journey.
Transforming your data into knowledge — and transforming that knowledge into actionable decisions — is the core promise of predictive analytics.
Raw data will need to undergo extraction, cleaning, and transformation before the model can use it to create useful predictions.
In addition to historical data, however, your predictive model needs some up-to-date input, and you may need to derive data from existing data points. To analyze and evaluate a stock's performance, for example, you might calculate a 200-day moving average of the stock price.
Organizing your data is more of a challenge when the data comes from multiple sources. It's essential to closely examine each data source for quality and relevance.
Global competition drives companies to lower prices to attract new customers. Companies strive to please their customers and gain new ones; customers increasingly demand high-quality products at cheaper prices. In response to these pressures, businesses strive to deliver the right balance of quality and price, at the right time, through the right medium, to the people most likely to buy.
Customers' experience of a product, if publicized through the power of the Internet — essentially a vast medium of communications — can make or break a business.
The Internet is a two-way street; companies have been gathering valuable information about their customers through transactional data that includes
To complete the picture, other sources of information come from business operations — for example, the amount of time customers spend on the company websites and the customers' browsing histories.
All that data can be combined and analyzed to answer some important questions:
Any information that can shed light about how customers think and feel can bring insight to understanding them and anticipating their needs.
Predictive analytics can also help improve your customers' experience by enhancing such processes as these:
Predictive analytics can also help you focus on identifying significant segments within your customer base — and make accurate predictions about their future behaviors. It can allow you offer relevant products to them with competitive prices, build targeted marketing campaigns, and enable you develop strategies to retain your customers and attract new ones.
Many online businesses use predictive analytics to manage customer relations. They harvest information about their customers, identify attributes that are the best predictors of customer behaviors, and use that information to make recommendations in real time. The result is that their customers get specialized and personalized service and attention, from marketing and cross-selling to customer retention.
Predictive analytics is an effective tool for more than customer management. It can help you reduce cost in many ways and at different levels of the organization — planning resources, increasing customer retention, managing inventories — and that's just for openers.
Predictive analytics is especially useful for reducing operational costs in these areas:
Predictive models can reduce operational costs by helping you decide when to make new orders, when to increase your marketing campaigns, as well as how to correctly price your products, manage inventories, and obtain a clear view (and a solid grasp) of your supply-and-demand chains.
By making more accurate decisions that correctly anticipate your business needs, you gain an advantage over businesses that manage their operations as a guessing game.
Predictive analytics can help you increase return on investment (ROI) through
By implementing predictive analytics, companies can accurately assess the present state of the business, optimize their operations, and compete more effectively in gaining market share. By scoring the predictive outcomes of future events, and using that information to their advantage, companies can improve their revenue and enhance the business performance as a whole. Deploying successful predictive models can help companies minimize risk and increase revenue across the board.
Increasing the number of successful business decisions always leads to better performance. Informed decisions — backed up with accurate predictive scores — increase the confidence of management in decisions that resulted from deploying predictive analytics models. Better decision-making, on the basis of more accurate information, is the core of predictive analytics.
Predictive analytics enables business to make smarter decisions, some of which take place in real time. It allows businesses to improve all aspects of decision-making — including confidence in decisions based on insights derived from the in-depth analysis of trusted information.
Predictive analytics helps your organization predict future events with confidence and make optimal decisions to improve business outcomes. It automatically gives your organization an edge over competitors who are still using guesswork to manage their operations.
Predictive analytics models can evaluate complex scenarios and make real-time decisions quickly — and more accurately — in anticipating your business needs. The more such results accumulate, the more you boost confidence in the model and its predictive capability.
The automation of this decision-making, backed by thorough testing and refined by feedback from operational deployment, allows for greater consistency of those decisions that require precision. The effect is to avoid the subjective analysis or emotional attachment that can often lead to biased decisions.
Deriving actionable information from the use of analytics promotes faster response to rapidly changing business environments and external conditions. This process empowers your business with the agility to better position itself, taking advantage of emergent business opportunities while managing risks and reducing costs.
Companies that adopt predictive analytics undergo a cultural change that affects every area of the business. Making more informed decisions provides them with the confidence necessary to adopt those decisions and develop better strategic plans. The larger result is that they become more efficient, increase their profitability, and position themselves to play a more competitive role in the marketplace.
Predictive analytics, properly developed and applied, turns your data into key insights, and enables you to take action by making informed decisions about many areas of your business — based on extensive analysis of your data. Greater accuracy in predicting future events is an advantage unto itself — in part because it can be applied to so many areas.
Sometimes the ultimate objective of a predictive model is the automation of certain business decisions. An example is an automated trading system that places real-time trades on your behalf, manages your portfolio (money and assets) and any financial leverage you may have. The goal is to make the best decision as quickly as possible — automatically — taking into consideration the many complex factors that affect money management in response to existing market dynamics.
Models' outputs can help a company make decisions affecting many aspects of the business, from supply-chain management to identifying opportunities and budgeting.
Companies use predictive analytics models to identify effective strategies that are effective and optimized to handle future events automatically, their actions guided by strategies based on knowledge acquired from thorough and exhaustive analysis.
A functioning predictive model can lead to making informed decisions guided by data analysis. If the model does its job well, its results are reinforced through testing — and validated by the feedback generated in response to its deployment, then when faced with new events, the business can rely on models that were built to handle them — especially if the events are unprecedented and unfolding in real time.
Making decisions with confidence based on predictive analytics models can provide your business with an edge over the competition by enabling you to
Predictive analytics can also play an important role in your planning, giving you the agility you need to stay ahead of your competitors. You can no longer afford to rely only on past experiences and executives' intuitions to run your business. Instead, predictive analytics can help you turn your wealth of data into a wealth of actionable insights and informed decisions (some of them made in real time).
Predictive analytics can free your company from taking a scattershot approach to improving performance by boosting various areas of the business:
Because predictive analytics models can also improve the reliability of business data and the agility of responses to emergent conditions, companies see a range of improvements: