When should organizations consider predictive analytics?

Lyndsay Wise, president and founder, WiseAnalytics
Lyndsay Wise, president and founder, WiseAnalytics

Many organizations have been leveraging Business Intelligence (BI) for years and are slowly taking advantage of Big Data platforms to gain access to diverse data and address complex analytical needs. But with increased competition, organizations also require the ability to look ahead as well as back to get the sense of how to improve performance, improve efficiencies and gain market share. The use of predictive analytics provides companies with additional insights needed to close the loop of information availability, creating visibility into historical trends, operational efficiencies and strategic planning, while improving their ability to meet performance targets.

Adopting predictive analytics entails some change in the way data is queried, making it essential for organizations to select a solution provider with predictive capabilities. This means that companies need to ensure that they have the right set of solutions to expand their BI use to include predictive analytics. If this is not the case, project sponsors need to find a way to sell the advantages of adding predictive capabilities to their BI deployments, as well as a way to point out the implications of not creating an environment that includes a forward-looking approach to BI.

BI maturity and predictive use

Mature BI use can lead to insights that extend beyond traditional BI access. In many cases, organizations want to forecast their performance and identify trends that might affect their sales and marketing campaigns going forward. Additionally, planning requires the ability to identify potential discrepancies between plan and outcome and develop scenarios that identify how performance might be affected by these discrepancies. Each of these exercises requires the ability to create and apply predictive models within a business intelligence framework.

For many organizations, this need emerges from consistent BI use over time. Organizations gain valuable insights from their analytics, want to look at operational analytics to become more proactive in their operations and planning, and want to supplement these insights with what-if analyses and other predictive algorithms. For instance, if an operational analysis identifies ways to increase productivity, predictive models can be used to identify the most effective ways to improve production and how that approach might potentially affect the supply chain, sales, distribution, etc. This can apply to any organization looking to become more efficient.

The reality, though, is that this type of BI expansion entails either the creation of new business rules, integration of statistical models like R, or the leverage of capabilities provided by the current solution provider. Organizations that adopt BI and don’t consider their predictive needs might be surprised by the level or lack thereof of predictive support. To move forward, organizations require two things:

  1. Project sponsor buy-in
  2. The right toolsets

Building the business case for predictive analytics use

Progression through BI use and maturity to predictive adoption requires gaining business sponsor buy-in and ensuring that the value proposition of predictive capabilities is identified. This means looking at:

  1. Why predictive analytics applies within the organization and what quantitative benefits will be realized?
  2. What type of predictive modeling is required? The level of complexity will affect the type of solution selected, the integration capabilities and the skill sets needed.
  3. How will predictive capabilities be applied and will they be attached to broader BI uses? For instance, is predictive modeling or what-if analysis part of a broader analytics initiative to understand performance more broadly, for financial use to plan and forecast, or for single use by an individual team?
  4. Who will be using predictive analytics and will it be a team-based project or company wide? This can affect how the project is sponsored and the business requirements associated with various feature sets.
  5. What are the overall goals of predictive use? Identifying potential benefits or ROI is one stage in building a business case for adoption. Understanding what longer term business goals are may be a different process as these may include items like the need to increase market share by five percent or to lower costs of production by a certain percentage or dollar amount.

Predictive requirements and the right toolsets

Ensuring the right solutions and capabilities exist mean identifying the following:

  1. Do capabilities already exist and how robust are they? Some vendors provide support for what-if scenarios but do not provide additional capabilities to support other predictive features.
  2. If internal functions do not exist, does the solution provider provide integration with R or other statistical or predictive modeling offerings?
  3. What is the level of expertise internally? Some organizations have statisticians on staff that can develop their own predictive models, or developers who have those skill sets. Before selecting a tool or trying to integrate these capabilities, it is important to make sure that the skills to support it exist internally. Otherwise, solutions selected should provide internal functions that provide easy access to predictive capabilities.
  4. What specific predictive functions are required? Many vendors provide support for what-if scenarios, but not as many provide advanced predictive feature sets. Identifying what business requirements are in advance can help eliminate future challenges.
  5. What partnerships or integration capabilities exist to expand predictive requirements as the need arises?
  6. What types of users will be leveraging predictive capabilities? User expertise may affect the level of interactivity needed – super users and statisticians can create their own models, while general business users generally cannot.

Takeaways

Although predictive capabilities are essential to some businesses, it does not mean that it is a must-have feature set. Organizations need to evaluate the value proposition of adding predictive functionality and identify the ROI associated with its adoption. For companies looking to become more efficient and provide better products and services or increase their market share, the ability to develop what-if scenarios and apply predictive models can help them achieve these goals. Overall, tying goals to product functionality can help determine which feature sets are most important. And understanding what product capabilities are most important to the business during product evaluation can help to ensure that predictive capabilities exist to help eliminate potential business challenges in the future.

 

 

 

 

 

LEAVE A REPLY