Unifying Analytics for Hyper-local Marketing

  • October 15, 2018
  • Case Studies
  • No Comments

Rebalancing FSI’s for a major retailer to grow more confidently



  • A leading US apparel retailer was looking to rationalize its investments in certain marketing channels. The retailer had substantial evidence from its own internal and vendor-provided Marketing Mix Models (MMM) that free standing inserts (FSI’s) were not performing well, despite being the largest area of investment.
  • The problem was that although the MMM identified this low ROI, it did not provide specific detail on how to reduce investments. That, paired with a large organizational heritage in FSI’s, meant the retailer was caught in a marketing trap that was preventing change.
  • Fortunately, the retailer had data and lots of it. Sales could be identified to the zip code level and mapped to customer loyalty segments by sales channel. Digital data was available at the log level with unified ID’s across all paid channels. (These data sets alone exceeded 2 terabytes!)
  • Marketscience was tasked to turn this hyper-local data into a solution that would not only identify where to cut FSI investment but also how to redeploy the savings in order to drive results in strategically identified customer segments and sales channels.


  • Marketscience developed an integrated MMM & multi-touch attribution (MTA) approach to address this challenge. The solution was designed to be executed at multiple hierarchical geographic levels in order to support variations across media.
    • TV operates at national and DMA levels, promotions and offers at store level, print and FSI’s at zip-code, and digital at the individual level.
  • Leveraging a combination of state-of-the-art methodologies, Marketscience looked at sales data cut both by geography and customer segment. Over 100,000 models were built and integrated to inform the investment optimization, with recommendations down to the specific publication, day-part, message and segment.
  • Finally, to address the risk and organizational concerns, a test and learn program was designed, allowing the company to implement these recommendations progressively in order to establish confidence through live, in-market changes.


  • Marketscience identified over $100 million in marketing expense savings, dropping half to the bottom line and redirecting the other half to digital and outdoor channels.
  • The client saw significant increases in new strategic segments (millennials and new customers) and in online store sales, while also preserving the stability of sales to core loyalty segments.
  • Overall sales improved by 2% with a more profitable mix of customer and sales channel.
  • Marketscience’s analytical solution provided this retailer with a new and effective lens to pursue even more strategic planning options with greater confidence. Today, with a new culture of experimentation, the retailer is exploring more opportunities in personalization and localization. “Test learn and deploy” is the marketing team’s new mantra.

What we did

  • Integrated MMM & multi-touch attribution (MTA)

Subscribe to the Series

Share Now:
Posted in:

Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Scroll to Top