Compare a KPI across sites, calculated the same way everywhere

Why do cross-site comparisons take so long?

Because the logic is tied to one site's tag names, units and meters, so it gets rebuilt for the next site.

Site A calls the steam meter one thing, site B another, and site C reports in a different unit. Each comparison becomes a spreadsheet exercise, and nobody is sure the numbers mean the same.

How does one KPI definition work across sites?

Write the KPI once, then map each site's tags to it; any exception is approved and recorded.

• The definition has an owner and a version, so a changed formula is visible and can be compared with the old one.
• Every step is checked with unit algebra: mixing units that don't fit is refused with a suggested fix.
• Averages are time-weighted, so deadbanded and report-by-exception tags give correct numbers.
• Two sites with the same tag names are kept apart, and each site can have its own time zone.

Dashboards for every site

A board built for one site can be copied to another by rules, without AI, and gets a site selector.

• A tile with no single equivalent at the other site is left out, with the reason shown.
• Sister assets (rigs, units, lines) can be ranked on a bar chart with statistical outliers marked.
• Every number on a board answers "How was this made?" with its source tags and definition version.

How do we know a number is right?

Saved pipelines run without AI, so the same input always gives the same output.

You can check any dashboard figure against your own SQL reference. Each KPI result also states which exclusion rules it applied.