August 25, 2016

Data Preparation, Predictive Analytics, Predictive Analytics Demystified, Salesforce Reporting, Tableau

Using Salesforce, ERP Systems, Tableau & R

Salesforce and integrated ERP systems collect tons of valuable information, but the challenge is figuring out how to use that data to predict future sales. Predictive modeling has promised to deliver tomorrow’s numbers, but remains under-used due to complexity or inaccuracy.

In this webinar recording, we share effective, pragmatic sales forecasting approaches. We describe best practices for organizing and preparing your Salesforce and ERP systems for predictive modeling, and reveal how to best use Tableau and R to visualize the future.

Effective methods for sales forecasting discussed include:

  • How to best organize, prepare and integrate your sales pipeline and ERP data for predictive modeling
  • Data preparation techniques specific to Salesforce
  • Three of the R-based algorithms available for sales forecasting
  • How to setup a continuous feedback framework so that your algorithms will improve as data volume grows
  • When it makes sense to use simple algorithms with less accuracy instead of complex ones that are more accurate
  • How to leverage the built-in integration between Tableau and R to visualize your historical, forecasted, and confidence-level data
  • How the techniques can be applied to environments that do not include Salesforce or Tableau

Tableau, Salesforce, ERP Systems, R


Asa Levi
Senior Consultant
Senturus, Inc.

Asa joins Senturus with over 10 years of experience in business intelligence and predictive analytics, having spent most of his career at Pearson Education. While there, Asa used multiple ERP and CRM systems to unify insight across all platforms to gain more understanding of issues, such as cost-to-serve and pricing model selection. He also led development of a combined SSRS/Tableau/Salesforce BI solution to allow CRM, ERP, customer support, and product master data be used in a unified environment to allow for simpler access to data and enable more thorough, complex analyses. Asa began his career in the Microsoft BI stack, working with SQL Server and SSRS on ROI analysis for product customization.

Greg Herrera
President and Co-Founder
Senturus, Inc.

Greg founded Senturus in 2001 and today leads the demand creation side of the company. Under his leadership, the company has been on Inc. Magazine’s annual list of America’s Fastest Growing Private Companies for three years running. In addition, Senturus was inducted into the San Francisco Business Times “Fast 100” Hall of Fame. Greg has 13 years of hands-on experience with Salesforce.com and the application’s underlying tables.


  • Predictive Modeling Methodology
  • Sales Forecast Algorithms Require Historical Snapshots
  • Sales Pipelines are Current Snapshots
  • Techniques for Maintaining Historical Pipeline
  • Salesforce Opportunity History Table
  • Salesforce Tracks History Functionality 
  • Goal: Merge Predictions for Actionable Insights
  • Quantitative vs. Qualitative
  • How to Measure Accuracy
  • Regression KPIs
  • Custom KPIs
  • The Common Refrain: Why can I predict that our net sales will be within 5% of last year, but we are 30% off at the product level
  • The Central Limit Theorem
  • Granularity Up, Accuracy Down
  • Accuracy By Industry
  • Choosing a Model
  • Investigating Predictor Relationships
  • Removing Outliers
  • Setting Up A Regression
  • Occam’s Razor Applies
  • Results
  • Learn and Iterate
  • Insights
  • Clustering to Find Insight
  • Conclusion