Predictive analytics is no longer a fictional technology best suited for sci-fi movies. What was formerly only potential is now possible. But with the proliferation of technology companies selling space-age, overly complicated solutions, organizations are often challenged about where to sensibly start.
It turns out that most (or all) of the data for making meaningful predictions more than likely exists in your organization. And getting to ROI is easier than you might think.
In this webinar recording, you’ll take a journey into the attainable. We walk you through an actual client use case study to show how using predictive modeling saved them over $1million in their first quarter alone. And we talk about different applications for predictive analytics and what’s needed to start your own project.
You may also be interested in learning how to use built-in functions and third-party tools to gain deeper insights in this Advanced Analytics in Tableau webinar recording.
Arik T. Killion
Advanced Analytics Architect
Arik has 16-years of experience in advanced and predictive analytics. Before joining Senturus, Erik was a consultant working at IBM as a data scientist/ technical professional for the North American Analytics Channel Development Team, Prior to IBM, Arik was the director of analytics at a large national marketing agency.
Arik has developed deep experience using a variety of methodologies from simple quantitative statistics to predictive modeling techniques and text/sentiment analysis to produce actionable insights, guide strategic management decisions and acquire meaningful business intelligence. He has created solutions for clients such as Chrysler Group, Kraft Foods, Allergan, Lowe’s, Verizon Wireless, JD Power, Nielsen Ratings Group, Sony Entertainment and many others.Read more
If This Is The Future, Where’s My Jetpack?
- The relevancy problem with predictive analytics
- Advances in technology seems to have made things more difficult
Marketing Case Study – The $1Million BI Column
- Major flooring manufacturer in the U.S.
- $8.1 billion in revenue (2015)
- Distribute to ~18,000 independent retailers in North America
- Floor samples units were a major investment
- 1 factory plant to produce year round
- ~$3k-$4k per unit (materials, labor, shipping, setup)
- Placement guided by relationships with company sales personnel
- $12 million/year marketing line-item
- Compile historical sales, RFM Analysis on prior unit placement sites
- Compile profile data for each store
- Gather demographics for each store
- Outcome bands:
- Green = Exceeded margin cost of unit in 1 year for the related products
- Yellow = Broke even or came close
- Orange = Fell short
- Red = Didn’t even look like they tried
- Created predictive models for each band within 3 brands (12 models)
- Automated scoring of every retailer on a monthly basis with a voting mechanism to select the best outcome prediction
- BI report for ordering sample unit placements
- Column for likely predicted margin outcome
- Saved $1.1million in the first quarter of use
- How to start
- Start in a problem area that will have great impact
- Realize you probably already have all the data you need
- Find a trusted adviser to help navigate
- Most predictive project ROIs are between 3 & 8 months (particularly in marketing)
- Understand systems can be automated
- Marketing is the tip of the iceberg
- Employee growth and satisfaction
- Accurate forecasts, demand planning, assortment planning
- Root cause analysis