Site Search
September 30th, 2019

Intel: 2019 Genius Awards Finalist (Data Analytics Innovation)

This is the eleventh of 11 posts Neustar will be running to highlight the achievements of the 2019 ANA Genius Awards finalists. Winners will be announced during ANA Masters of Marketing Week, Oct 2-5.



The problem with black box algorithms is that no one trusts them when their results differ from what human intuition can predict. How did Intel create a credible Marketing Mix Model (MMM)?

The Challenge:

In the semiconductor industry, the brand is paramount. It takes years to build working relationships with OEMs, and the brand identity arguably does most of the heavy lifting. Intel had been working on a way to quantify that lift for several years. 

Specifically, Intel needed to create an MMM that was both credible and defensible to all stakeholders. They also needed a Return On Advertising Spend (ROAS) model with the same properties. Previous attempts using black-box algorithms had failed to gain widespread adoption.

The Process:

Intel know that their next approach needed unimpeachable credibility. They partnered with an advanced customer modeling firm that was chosen for their academic background and which had published several papers. 

Yerjanat Khabai, Marketing Data Science & Analytics Manager at Intel, used this background to gain leverage with the finance department. Once the finance stakeholders read the research, they realized that the organization could not only create credible models, but also lower costs by allowing Intel to optimize its marketing spend.

The Results:

Working with their modeling firm, Intel created an advanced Dynamic Mix Model to measure ROAS. This model provided both credible estimates of short-term impact as well as a complete measurement of brand marketing, and it was based on a complex error-correction model that looked at long-term relationships between variable.

As a result of the implementation, Intel was able to identify key performance drivers and understand which of their marketing efforts was able to bring in revenue versus those that lost money. Within the data, they found some interesting conclusions.

For example, Intel’s marketing efforts have a long tail – ROAS increased when the model incorporated the long-term effects of a marketing push. Intel was able to increase sales by capitalizing on this affect. In addition, brand marketing generated up to 20% more sales than originally thought. 

At the end of the day, Intel was able to cancel underperforming initiatives and double down on successful efforts, with resulted in a roughly 20% increase in marketing productivity.

Final Thoughts: 

This isn’t just a story about marketing – it’s a story about collaboration. It can be very hard to gain traction for data science in an organization where these projects have previously failed. Intel showed the right way to lay the groundwork for an accountable data science program in a way that drew marketing, finance, and data science together.



Previous: Pittsburgh Cultural Trust I First: Adobe

For more about all the 2019 Genius Awards Finalists, click here.

To learn how Neustar helps enterprise marketers like you build real, actionable insights that help you truly understand the impact of your marketing investment, click here.

Let's Connect

Learn How Your Company Can Benefit from the Power of Trusted Connections.

Contact Us   Give us a call 1-855-898-0036