Case Study

Getting More Out of Marketing Investments with KLIS

Improving the predictability of sales and supporting efficient investment choices to maximize ROI by identifying and proactively monitoring Key Leading Indicators of Sales (KLIS)

Forward looking performance management

Actionable Marketing and Sales discussion

12% Sales growth due to efficient investment choices

The Business Challenge

Retailers often rely on lagging indicators like market share to determine if their goals have been met.  Additionally, the severe competition and the increasing costs, lead to a lack of results from ad investments and negative ROAS. There is a need to gain a granular understanding of the details of what drives sales performance beyond the amount spent on media.

The Key Leading Indicators of Sales (KLIS) supports performance-focused marketing by using Machine Learning to look at individual relationships between marketing and sales activities as well as systematic interconnectedness. 

With KLIS, the dependencies across online and offline activity, for example: TV spend, print ads, search, pricing, promotion, and messaging are identified and actively managed for maximum impact.

Getting More Out of Marketing Investments with KLIS 1
Getting More Out of Marketing Investments with KLIS 2

Technical Solution:Interlinked with all dependencies (product series)​ elastic net model with drivers of offline and online sales

The Solution

Our client, a leading consumer robotics company was using an MMM study, which concluded that TV advertising was delivering high ROI for the manufacturer and recommended they significantly increase their investment in this particular channel. However, when the key leading indicators of sales analysis was conducted across the whole business, it uncovered a strong interdependence between TV investment, Google paid search and Amazon-sponsored product campaigns.

If the company had followed the MMM recommendation without considering the additional insights provided by the more holistic KLIS model, it was likely the competition would benefit from the increased consumer interest created by further TV advertising , as they had a high share of voice in both paid Google search and Amazon campaigns.

Our Approach

The methods used

KLIS investigates individual relationships between marketing activities as well as systematic interconnectedness. To do this, KLIS uses advanced statistical analytics like Bayesian belief networks and Elastic Net to answer: What drives sales? What drives intermediate factors driving sales?


  1. Hypotheses testing to answer business questions
  2. Advanced modeling to identify drivers of sales
  3. Live tracking for proactive monitoring

ML Model deployment

The final outcome of the study is a model that shows the interactions between predictive and outcome variables. In this particular case, there are 2 intermediate levels (search and traffic) and one final level (sales).

The model shows how each marketing activity affects these outcomes and how they affect each other. Not all marketing activity shows up as significant. This means some marketing activities are not impactful for some outcome variables.

Knowing such information allows the clients to optimize their marketing budget.

Business impact

​​As the elastic net regression model has revealed the extent of this interconnectedness it has enabled the multidisciplinary team to design a successful way forward: deciding to increase investment in all three areas recommended by the KLIS analysis to ensure that awareness at the top of the funnel gets converted into sales rather than feeding competitive growth. The company achieved 12% growth, with Amazon channel sales doubling at a time of increased competitive pressure.