Analytics Connector

01

A contract, not an export

Every feed has a stable name, stable column names, declared types and one sentence saying what a single row is. A column is never renamed and never changes type. That is the difference between a data source and a spreadsheet somebody emailed you: the report you build this month still works after the next release.

02

Five minutes, not two days

Download one file, open it, put your key in once. Power BI starts with the table list already loaded and every feed you are allowed to read in it. Excel takes one pasted URL. Looker Studio gets a ready-made script with your addresses already filled in.

03

The line-level grain, not a summary

Invoice lines, journal lines, purchase order lines with what was actually received, stock movements one at a time. Whoever builds your reports gets the same grain the accounting engine works at, so their numbers tie to yours instead of nearly tying to yours.

04

One key, one list of tables

The key you give the sales dashboard reads the sales tables and nothing else. It cannot reach the ledger, and the ledger does not even appear in the table list it is offered. Revoke it and it stops reading the same second.

05

Who took what, and when

Every read is recorded: which key, which table, how many rows, how long it took, and from where — including the attempts that were refused. The record is append-only, so when an auditor asks who has a copy of your books, the answer is a screen rather than a guess.

06

A switch you can reach in a hurry

One table off, one transport off, or the whole connector off — one click each, from inside the workspace, with a reason recorded so whoever finds the broken report knows why. Nothing about this needs a support ticket.