Every system lives on its own
Ads in one console, web analytics in another, CRM and call tracking in a third. The summary is stitched together by hand at month end — after the decisions were made.
29 data sources, each with an honest status
Data Hub shows where your data comes from, what it is assembled into and where it goes: 29 sources, 14 destinations and five ready-made datasets. Flows ship switched off, and writing to an external system waits for the owner's approval.
Why the numbers don't agree even when everything is 'connected'
A zero means two different things
A zero on the dashboard is either a real zero spend or a sync that failed. Until those are told apart, the report and the AI summary are wrong with equal confidence.
Connecting stalls at platform review
Your own developer app, a scope list, screenshots and a review on the platform's side. The integration sits for weeks and nobody remembers which step it stopped at.
How data moves through Data Hub
Pick the sources
The catalog holds 29 connectors. Each one shows its readiness — live, partial, planned or waiting on platform review — and the current state of the connection.
Assemble the dataset
Five ready templates: PPC, SEO, ecommerce, lead quality and an executive summary. Currencies are normalised, rows deduplicated, suspicious zeros flagged.
Send it to destinations
Dashboard, report builder, client portal, spreadsheets, BI tools and exports. Writing turns on only with the owner's approval, and every run is logged.
What you get in Data Hub
A catalog of 29 sources
Ad platforms, web analytics, ecommerce, CRM, finance, call tracking and aggregators — each with its auth type and the list of streams it provides.
Readiness stated honestly
Every source says whether it is live, partial or still planned, and what exactly is blocking the connection right now.
14 destinations
From the dashboard and report builder to Google Sheets, Excel, Looker Studio, Power BI, BigQuery, Snowflake, webhooks and CSV/JSON exports.
Suspicious-zero protection
The dataset separates a zero the platform returned from data that never arrived, and shows source coverage next to the metric.
AI grounded in the dataset
An AI answer has to cite the dataset, the source, the date range, freshness and row count. On stale or thin data the recommendation is blocked.
App setup assistant
Each provider gets a packet: required scopes, the developer portal link, the evidence list and the current status. Review and approvals stay with a human.
In numbers
- 29
- data sources in the connector catalog
- 14
- destinations: sheets, BI, warehouses, webhooks
- 0
- writes to an external system without owner approval
Questions about Data Hub
Which data sources can I connect right now?
The catalog holds 29 sources: ad platforms, web analytics and Search Console, ecommerce, CRM, finance, call tracking and aggregators. Each carries a readiness status — some are live, some partial, some waiting on the platform's own review. The catalog states this openly instead of hiding it behind a wall of logos.
Can Data Hub write into my Google Sheet or warehouse?
Writing is supported: every destination lists its write modes — replace, append, merge, snapshot or dated history. It stays off by default, because it needs the owner's approval, an explained write mode and a run log before it is enabled. Several warehouse destinations are still marked as planned.
How do I know whether a zero in the report is real?
Data Hub separates a zero the platform returned from a sync that never landed. Next to the metric you see source coverage, the time of the last successful run and the row count, so a zero day does not automatically become the conclusion that the ads stopped working.
Do I need my own developer app for every platform?
Often yes: platforms require their own app, a scope list and a review. The setup assistant builds a packet per provider — scopes, the developer portal link, the evidence list and the current status. Passing review, entering verification codes and accepting legal terms stay manual on purpose.
Can I let AI work on this data?
Yes, with a limit: the answer must cite the dataset, the source, the date range, the freshness and the number of rows behind it. If the data is stale or the sample is below the threshold, the recommendation is blocked rather than delivered with false confidence.
Get your data into one place
The source catalog is open to browse and the flows ship switched off. Nothing is written outward until you allow it.