Learn the key concepts behind Datama to better understand how the solution works


1. What problems we solve

The Compare extension helps you identify the exact causes behind performance variations.

Whether you’re comparing different time periods or regions, this tool provides clear insights into what’s driving changes.

Results are presented through an intuitive waterfall workspace: the bridge chart, smart comment, and optional table / tree surfaces stay aligned as you explore.


2. Required data types

The analysis requires two main types of business data:

  • Business Metrics (revenue, sales volume, marketing campaign reach, etc.)
  • Associated Dimensions (customer segments, product types, countries, vendors, demographics, time periods, etc.)

Specifically, you’ll need to configure four key elements:

  1. Main KPI: The primary metric you want to analyze (e.g., Revenue, Leads, Margin)
  2. Comparison Dimensions: The elements you want to compare your KPI across (e.g., Month-over-Month, Country-to-Country)
    • By default, Datama selects the first two elements or splits a date range in two
  3. Steps: The funnel stages leading to your main KPI (e.g., eCommerce or Finance funnel) See dataset examples
  4. Explanation Dimensions: Additional dimensions that help explain performance variations, each with an interest score indicating its importance


3. Metric relations

A metric relation describes how different metrics in your data source combine to compute your target KPI.

Simple Example

Consider a retail scenario where we want to understand Revenue. The main KPI is Revenue.

We can break Revenue down as: Revenue = Volume * Revenue/Volume. This creates two meaningful performance indicators (PIs):

  • Volume (number of products sold)
  • Revenue/Volume (unit price)

This breakdown helps separate effects managed by different teams (sales volume vs. pricing strategy).


Advanced applications

Each PI becomes a “Step” in your waterfall chart. While there’s no technical limit to the number of steps, we recommend keeping it under 10 for clarity.

A step represents a ratio between two metrics. For simple metrics like Volume, the denominator is 1.

For example, with 5 metrics (Metric1, Metric2, …, Metric5), your KPI equation might look like:

Metric5 = Metric1 * Metric2/Metric1 * Metric3/Metric2 * Metric4/Metric3 * Metric5/Metric4

Don’t worry about complexity - Datama Compare lets you zoom into specific parts or aggregate components using the “Skip steps” feature.


4. How to interpret waterfall charts in Datama

The waterfall chart is the core visualization in the extension (rebuilt on the Plotter engine with hierarchy layers and smooth open/close updates).

Datama shows how each data variation affects your Key Performance Indicators (KPIs):

Each step shows a performance ratio’s variance and its contribution to overall change.

Example interpretation:

  • The chart explains a -28.1% Revenue drop from last year (45,549€) to this year (32,747€)
  • Key factors include -11.7% in Users and -26.7% in Checkout/Session rates
  • The Checkout/Sessions decrease of -26.7% impacts Revenue by 12,298€


To dive deeper (demonstration with another use case, but the logic is the same):

  • Click any step to see which dimension best explains its variations (in this case, Canada contributes most negatively to transactions this year vs last year.)

By default, the waterfall groups segments with smaller effects into the Remaining category. Click a Remaining bar to expand more elements:

  • Use Split by (right-click on a bar or on value / top-line labels) to view interest scores of other dimensions and change the display. Smart title, subtitle, and comment follow the active drill.
  • Switch exploration to Hierarchy or Top N from the context menu or Tree settings to focus on top drivers
  • Optionally pin the Table view to read the same hierarchy as numbers
  • Insert custom pillars (right-click › Add pillar) to mark net checkpoints with End-style labels
  • Each variation can break down into mix and performance effects (explained in modeling settings)



Please share your thoughts with us. Whether you have questions about the solution, your analysis, or the documentation, we’re always happy to help and hear your perspective.