Working with Attributions

Creating a New Attribution

To create a new attribution in Alvíss AI, follow these steps:

1. Select a Model:

   • Choose models that correspond to the countries, regions, and groupings of interest. These models will be used to calculate the impact of each variable on the Key Performance Indicators (KPIs) for the selected areas.
     2. Choose a Date Period:

   • Specify the date period you want to include in the attribution analysis.
     3. Default Settings vs. Advanced Settings:

   • You can create an attribution using Alvíss AI’s default settings, which are pre-configured for general use.

   • For more customized attributions, navigate to the Advanced section. Here, you can:
     • Select the number of samples drawn from a posterior distribution.
     • Set baseline values for each variable.
     • Note that increasing the number of samples will extend the calculation time.

Why Change Baseline Values

Alvíss AI’s attribution calculations use a default set of baseline values designed to work well for most business cases. However, as an advanced user, you might want to customize these baseline values to better align with your specific business context and goals.

Default Baseline Values:

• The default values provide a general reference point suitable for a wide range of scenarios.
• They are pre-configured to ensure reliable and consistent attribution calculations across different industries and applications.

Customizing Baseline Values:

• Tailor to Your Business: By adjusting the baseline values, you can make the attribution analysis more relevant to your unique business conditions and challenges.
• Leverage Your Expertise: Use your industry knowledge and insights to set baseline values that reflect realistic scenarios and benchmarks for your business.
• Enhance Precision: Custom baseline values allow for more precise and meaningful impact assessments, improving the accuracy of your decision-making process.

For instance, in the context of media investment:

• Default Setting: The baseline value for media investment is set to 0.
• Custom Setting: You can change this baseline to a specific amount that better represents your typical media spend, providing a more accurate measure of the impact on KPIs like sales.

By modifying baseline values, you gain greater flexibility and control over your attribution analyses, enabling you to derive more actionable and business-specific insights from Alvíss AI’s powerful data-driven platform.

Analyzing a Single Attribution

After calculating an attribution, you can delve into the details page for in-depth analysis:

1. Dashboards:

   • Alviss AI provides various dashboards for each variable group, displaying the impact on KPIs.
     2. Explore Tab:

   • Use the Explore tab to visualize the effect of a single variable or compare multiple variables’ effects on a single graph. This helps in understanding the contribution of each variable to the overall outcome.

Working with Attributionsets

To analyze multiple attributions simultaneously, you can group them into an Attributionset:

1. Creating an Attributionset:

   • Group several attributions based on your analysis needs.

   • Once the Attributionset is activated, it becomes visible in the Effect dashboard, allowing you to view aggregated impacts.
     2. Handling Overlapping Dates:

   • If chosen attributions with the same Country, Region, and Grouping have overlapping dates, the newer attribution will overwrite the older ones.
     3. Extending an Attributionset:

   • If you want to add new data while maintaining old results for consistency, you can extend an existing Attributionset. This feature is useful for continuous analysis and historical comparison.

Active Attributionset

Your active Attributionset is the one visualized in the Effect dashboard. To activate the Attributionset press a lighting bolt next to it.

Ensuring Sensible Results

To ensure that your attribution results are meaningful and reliable, follow these steps:

1. Verify Direction of Impact:

   • Check if the effects of your variables on each KPI are in the expected direction (positive or negative).

   • For example, if you increase your marketing spend, you would typically expect a positive impact on sales.
     2. Assess Magnitude of Impact:

   • Evaluate whether the size of the effects (magnitude) aligns with your expectations and business understanding.

   • For instance, if a minor adjustment in pricing results in an unusually large change in sales, this might indicate a need to review your model or data accuracy.
     By following these steps, you can effectively utilize Alviss AI to create, analyze, and manage attributions, gaining valuable insights into the factors driving your business outcomes.