SadaIQDefine a metric once. Trust it everywhere.
The semantic layer holds the business meaning of your data: what a metric is called, how it is calculated, who owns it and at what grain it can be sliced.
Meaning that travels with the number
A metric carries its owner, grain, time logic and lineage into every chart, answer and export. When the definition changes, the version changes with it, so a figure from last quarter can still be explained next year.
Illustrative synthetic data
Flow links from source to audience, in synthetic units
| From | To | Value |
|---|---|---|
| Orders database | Sales dataset | 620 |
| CRM | Sales dataset | 140 |
| CRM | Customer dataset | 180 |
| Files | Customer dataset | 90 |
| Sales dataset | Net revenue | 760 |
| Customer dataset | Repeat rate | 270 |
| Net revenue | Executives | 460 |
| Net revenue | Operations | 300 |
| Repeat rate | Executives | 120 |
| Repeat rate | Operations | 150 |
- An orders database, a CRM and files are the three sources.
- They feed the sales and customer datasets.
- Those datasets resolve into the net revenue and repeat rate metrics; net revenue receives 760 units from sales, 620 of them from the orders database.
- Net revenue then feeds executives with 460 units and operations with 300.