Metric relation
Metric relation describes the equation between metrics of your source, in order to compute the KPI you want to explain.
Read more: At Datama, we call it the « Market Equation ». Find more information on how to create a metric relation in our article .
In order to have the most relevant analysis of your data, Datama allows you to manually define your market equation.
To do so, click on the Datama block, define your metrics and dimensions, then define your market equation.
Add the number of steps you want and define the function to join your different steps !
By default, Datama generates a funnel market equation.
Building it can be a little bit tricky, but generally you have it behind your business KPIs. All you need to do is come back to basics and try to write down how your metrics are related.
1. Simplistic example
A simplistic retail example: As a retailer I want to understand my Revenue. So my main KPI is Revenue .
Now, I can easily split Revenue that way: Revenue = Volume * Revenue / Volume . That makes a lot of sense, because then I have two underlying performance indicators (PI) that actually helps me split the different effect behind what I observe on my main KPI Revenue . First underlying PI is Volume (the number of products I sell), and the other is Revenue/ Volume , which is basically my unit price.
Why do I want to split those? Well basically because it’s not the same team in charge of volumes and the one in charge of pricing. So, in order to be able to identify where I need to improve, I need to split those two effects.
2. Generalizing
We call each underlying PI a « Step », because it will be steps of the waterfall you’re building. Number of steps can be theoretically unlimited, but for obvious visual and business reasons, you probably want to limit to no more than 10 steps.
One « Step » is a ratio between two metrics of your source. You have a numerator and a denominator. When no denominator exists (e.g. Volume ) we just say that Denominator is 1.
Steps gets multiplied together in order to get your KPIs. So, if your file has 5 metrics (call it Metric1, Metric2, …, Metric5 ) and your main KPI is Metric5, Metric Relation would typically look like this:
Don’t be afraid of putting too many PIs in that equation: with Datama Compare, you can always zoom in to part of the equation, or aggregate some components using the « Skip steps » functionality.
Datama goes further with Metric Relation by adding multiple properties to refine your analysis for your use case, like Unit, focused dimension and excluded dimensions
3. Get inspired
Datama examples can help you get idea of Metric Relation you may want to build.
4. AI-Generated Market Equation
The AI-Generated Market Equation feature enables you to leverage the power of artificial intelligence to automatically suggest a market equation for your dataset clicking on the ✨ CTA in the Quick setup options. This functionality uses the names of your columns and any contextual information you choose to provide to generate relevant equation suggestions that may accelerate your data modeling and preparation process.
How It Works:
When you click on the ✨ CTA, first it will ask you if you agree with the IA term and condition (for the first time you use it). Note that no actual data values are shared or processed—only column names and context are used to ensure full privacy of your dataset, just the names of columns will be shared. When activated, a AI chat pop-up will appear and the AI will analyze:
- The column names of your dataset.
- Any additional context you enter (e.g., currency which is by default in ‘$’, language, rename a step).
It then suggests a possible market equation aligned with the provided information in a table. You can then iterate with comment to adjust the market equation, names or unit if needed. When you are satisfied with the result you can apply the equation by clicking on ‘Use This Equation’.
If needed, a ‘Reset’ CTA is available at the bottom-left, it will clear all the historic from the chat and restart a compute without any additional information.