Understanding Unique Metrics and Dataset Discrepancies In The NinjaCat Template Builder
When working with data in NinjaCat, especially when using Datasets or comparing Quick Connector results to template widgets, it’s important to understand how unique metrics behave. Differences in how data is ingested, stored, and queried can affect how results are calculated and displayed, especially when using metrics like Unique Users or Unique Clicks. This guide outlines key considerations to help ensure your reports and dashboards reflect the most accurate values possible, and to clarify why you might see differences between Dataset-based widgets and those pulling directly from a provider.
🧩 Unique Metrics: Direct Provider vs. Dataset Behavior
Unique metrics (like Unique Users, Unique Sessions, etc.) behave differently depending on how your widget is connected to data:
- Quick Connector (Direct Source): When a widget pulls data directly from a provider like Google Ads or Facebook, NinjaCat sends your exact request—including dimensions, filters, and date ranges—each time the widget loads. The provider then calculates unique metrics in real-time. This typically results in accurate values, even for aggregated views (e.g., weekly or monthly reports).
NinjaNote: If your widget filters by a field (e.g., Campaign) that’s not shown as a dimension, the provider still calculates uniqueness at the filtered level. This can cause totals to appear lower than expected
- Datasets: Datasets store data that’s ingested once, based on a defined set of dimensions and metrics. If unique metrics are included, they’re calculated at the ingestion level—often daily. When you later aggregate this data (e.g., into weekly or monthly views), you’re summing those pre-computed values, not recalculating uniqueness. 🚫 Unique metrics are not additive. Summing daily unique values over a month can result in inflated or misleading totals.
Best Practices:
If accurate unique metrics are required over longer date ranges, we recommend building widgets that pull directly from the provider. For daily reporting, dataset-based widgets may still be accurate.
🔍 Comparing Quick Connector vs. Dataset Data
It’s natural to compare widget results between Quick Connector and Dataset sources in the Template Builder—but be cautious.
Even if your widget appears to reference just a few fields, all dimensions and metrics used in the Dataset’s ingestion stream are active behind the scenes.
To accurately compare values between Quick Connector and Dataset widgets: • Make sure the Quick Connector query uses all the same dimensions and metrics used in the Dataset. • Be sure date ranges and filters match exactly. • Understand that calculated fields or naming conventions in Datasets may further impact results.
Only when both are truly “apples to apples” will you see a 1:1 match.
Updated about 7 hours ago