Attributions

What is Attribution?

In Alvíss AI Attribution refers to the process of determining the impact of different business components on outcomes such as sales, customer satisfaction, and marketing effectiveness. This is achieved through a systematic approach that involves running predictions with both baseline and observed values for each variable. By analyzing the differences between these predictions, Alvíss AI provides actionable insights into which factors contribute most to your business outcomes.

How Attribution Works

Attribution is calculated through a three-step process:

1. Baseline Prediction:

   • For each variable we run a prediction using baseline values. These baseline values represent the default or standard levels for the variables under analysis.

   • This prediction serves as a reference point, illustrating what the expected outcomes would be without any changes or influences from the observed variables.
     2. Observed Prediction:

   • We run a second set of predictions using the observed values for each variable. These observed values are the actual, real-world data points collected from your business operations.

   • This prediction reflects the true impact of the variables based on real data, showing how the outcomes change when influenced by these observed values.
     3. Difference Calculation:

   • The final step involves calculating the difference between the predictions from steps 1 and 2.

   • This difference represents the attribution value, highlighting the impact that each variable has on the overall outcome. A positive difference indicates a positive impact, while a negative difference indicates a negative impact.

Example of Attribution in Action

To illustrate how attribution works, let’s consider an example where we want to determine the impact of marketing investment on sales:

1. Baseline Prediction:

   • Marketing Investment (Baseline): $0

   • Predicted Sales (Baseline): 1,000,000 units
     2. Observed Prediction:

   • Marketing Investment (Observed): $150,000

   • Predicted Sales (Observed): 1,200,000 units
     3. Difference Calculation:

   • Attribution Value for Marketing Investment: 1,200,000 (Observed) - 1,000,000 (Baseline) = 200,000 units

In this example, the attribution value of 200,000 units indicates that your investment of $150,000 in marketing resulted in an increase of 200,000 in sales.

Using Attribution Results

Understanding attribution values allows businesses to:

• Identify Key Drivers: Determine which variables have the most significant impact on desired outcomes, enabling more focused and effective strategic planning.
• Optimize Investments: Allocate resources more efficiently by investing in areas with the highest positive impact.
• Improve Decision-Making: Make data-driven decisions by understanding the true effect of different factors on business performance.

Conclusion

Attribution in Alvíss AI provides a clear and quantifiable method for assessing the impact of various business components. By comparing baseline and observed predictions, businesses can gain valuable insights into which factors drive success and make informed decisions to optimize their strategies.