Apache Superset vs Whatagraph

Whatagraph builds client reports for marketing agencies by pulling from dozens of ad, analytics, and call-tracking sources. The recent releases split into two groups. One keeps widening the source catalog — Snowflake as a warehouse source, WhatConverts lead data, and Ahrefs Rank Tracker as a keyword-level integration separate from the existing domain-level Ahrefs. The other attacks reliability: a Data Storage destination that pre-syncs data so widgets stop waiting on live APIs, the Basis integration rebuilt on that model, and automatic failover when the team member whose account owns a source loses access.

The reporting layer is being made boring in the good sense. Live-API reports break in ways an agency cannot explain to a client — a token expires, someone leaves, a widget goes blank mid-presentation — and the last few months answer each of those failure modes directly: store the data, align spend figures with the vendor's own billing totals, and silently switch to a colleague's credentials when one account stops working. The presentation work is following the same instinct, replacing per-metric threshold rules with automatic shading and consolidating scattered style settings into themes. Less configuration to maintain, fewer ways for a report to be wrong in front of a client.

Credential failover currently rescues a source when someone else on the team happens to have access; the gap it exposes is agency-owned connections that do not depend on an individual at all, which is where this thread points next.