Skip navigation
blue_abstract_data_visual.jpg Getty Images

How to Build a Strong Predictive Analytics Process in a Rollercoaster Reality

Without a strong predictive analytics process, organizations lack the ability to draw trends to project business drivers.

Predicting future business performance in a volatile world gives organizations the upper hand. However, predictive analytics is often challenged by turn-twisting macro-economic (supply-chain disruption, inflation), and geopolitical (war, elections, etc.) factors. In addition, the absence of a strong historical data base contributes to greater forecasting complexity. The truth is that without a strong process, organizations lack the ability to draw trends to project business drivers.

Similarly, roller coasters spark the same uncertainty that one experiences in building predictive analytics models. In fact, it was in a recent amusement park ride where a I had a flashback moment. In previous lives, I often had to build forward-looking dynamic tools that had to adapt to “quick turning” external factors. A roller coaster ride of emotions indeed.

If you ask me if we always hit the target, the short answer is no! The art of predictive analytics is in closing the gap between commercial objectives and forecasts. In the latter, we succeeded. Here are the best practices that made our forecasts highly accurate and trusted by senior leaders:

Start With a Pragmatic Approach

Being pragmatic requires the use of a practical approach to solving a problem. The 7-steps below set the foundation for a simple yet highly efficient predictive analytics process:

  1. Begin by running an analysis on the quality and quantity of historical commercial data at your disposal. Note: Focus on co-relating patterns in commercial drivers (i.e. shifts in prices linked to shifts in demand).
  2. Draw a connection between all commercial objectives, and predictive analytics metrics that are part of your sales scorecards. This will showcase the value and relevance of your forecast.
  3. Run an inventory of existing predictive analytics tools (i.e. machine learning, linear regression modeling) to decide on the model that best fits the organization’s needs.
  4. Identify a use case (i.e. pipeline management & forecasting) and run a predictive analytics pilot to prove its use and scalability.
  5. Set predictive analytics benchmarks (i.e. 2-3% Margin of error between target and forecast) that help track the success of your predictive analytics model.
  6. Build “easy to consume” predictive metrics by creating score cards that synthetize KPIs insights, easily explaining deltas between targets, forecasts, and actual data trends.
  7. Implement a 3D view scenario that includes a pessimistic, realistic, and optimistic predictive view. This will give a sense of how aggressive commercial targets are compared to each forecast view, and help justify future incremental resource (i.e. product or account re-alignment strategies) asks that can be used to help close the gap.

 

Build Strong Cross-Functional Alignment

It is evident now that uncertainty is at the center of predictive analytics models. Particularly in challenging times. Therefore, building strong business partnerships is key to break down internal forecasting misalignments caused by sudden shifts in business drivers. This is key to let predictive analytics run its course. Below a few best practices to build alignment:

  • Create strong partnerships through ongoing taskforces including data (i.e. business intelligence, finance) and business experts (sales, products, and commercial leaders). This will allow sales leaders to control misaligned data insights that could create internal noise.
  • Communicate variations in business and data drivers through formal channels (i.e. monthly pipeline reviews, ELT presentations) that may result in overachieving/underachieving a commercial target. Always ensure that deltas in forecasts vs. targets are accompanied by strong business context and clear corrective actions.
  • Establish clear roles and responsibilities to execute the actions derived from predictive analytics models. Leveraging the RACI matrix (responsible, accountable, consulted, informed) could set structure and stability to ensure that predictive metrics result in actionable insights.

The combination of practicality and strong business relationships allows organizations to “squeeze” the value from forward looking data, resulting in maximized business potential. Like a roller coaster ride, keeping calm and focusing on the facts makes the journey much more manageable.

This article originally appeared on the Gartner Blog Network

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish