Media Mix Modeling Case Study

Client: Department Store


  • Determine effectiveness of each marketing channel
  • Drill down to the true drivers of sales
  • Predict future success
  • Reallocate budget dollars to the most effective channels


Under budget pressure, experiencing flattening sales and with a desire to make their marketing dollars work harder, a national off-price department store chain sought to optimize their marketing spend to maximize sales.


The goal of the project was to determine which marketing channels are currently providing the best returns and to create a strategy for reallocating budget dollars in order to increase success and ultimately, profits.


In order to help this retailer achieve their goal, Lift361 deployed Media Mix Modeling, a proprietary process designed to help clients optimize spend across marketing channels. The process included analyzing marketing channel spend by market, day-by- day, for a 15-month period of time.

This particular client was investing in television, radio, ROP, display ads, inserts, social media, search marketing and retargeting. The analysis projected the sales gains and losses expected with an increased or decreased investment in each channel.

The data collected and analyzed was used to create a model for future spend, showing how changes in investment level across specific channels could impact sales, ideally without increasing overall marketing budget.


A Full Media Mix Model gave the client an extremely detailed analysis of marketing ROI. Some of the highlights of this particular analysis include:

  • Unique Customer Attributes: This particular client has an affluent, older customer base whose shopping behavior is heavily influenced by the stock market. When the market is down, so are sales, which means there will be periods of time when this external force will have a significant impact.
  • Isolating Sales Drivers: The analysis identified a 2% increase comp sales increase for a specific 10-week period with the simple reallocation of spend by channel. Additionally, the analysis isolated the factors that drove that lift. Of the 2% gain, 1.29% was due to an increase in customer base, .56% was driven by TV, .12% was driven by direct mail.
  • Reducing Ineffective Channels: The retailer invested significantly in search marketing in the analysis time frame, and the analysis determined that a reduction in spend here would not negatively impact on sales.
  • Decay Rate: For this retailer, inserts showed significant decay – nearly 100%. This means that inserts sent out the first week of the month do nothing to drive sales by the end of the month. Conversely, TV ads for this retailer showed a decay rate of only 33%, meaning customers remembered them far longer and had a better chance of driving behavior even weeks after the ads were seen.

Through this detailed process, we were able to help this client develop a plan to reallocate existing marketing resources across the most effective channels, with an expected incremental increase of $13 million, a 1.7% lift.

Media Mix Modeling is extremely detailed and does more than simply analyze what happened in the past. The “modeling” portion of the Media Mix Model allows customers to answer “what if” questions. “What if we spent an additional $1M on TV ads? What if we reduced our radio budget and moved those dollars into direct mail,” etc. This modeling gives marketers the tools to prove ROI when planning budgets. This Media Mix Model achieved a 91% accuracy rate when predicting future results.


By analyzing your spend across all of marketing channels and looking at their performance, we’ll help determine an optimal strategy for allocating your resources.