Platform · What we sell

Semantic model

A database workbench — connect their warehouse, write SQL, name datasets. Analytics and Halo read only what you promote to live.

This is

The connection and SQL layer between their database and SKUsight reports / Halo.

This is not

Not a preset SKU table catalogue and not a second Power BI.

  • Connect Postgres, MySQL, SQL Server, SQLite, or another engine they already run
  • Browse their schemas once the connection inspects
  • Write any SQL and pin it as a named dataset
  • Promote staging → prod → live so Halo and reports share one bind

How it works

Their warehouse. Your SQL. One live bind.

  1. 1

    Connect their database

    Postgres, MySQL, SQL Server, SQLite — host, user, the engine they already run.

  2. 2

    See their tables

    Schemas and tables from that connection. Not a SKUsight catalogue.

  3. 3

    SQL and named datasets

    Write any SQL. Pin it as a dataset for reports or Halo.

  4. 4

    Analytics and Halo read live

    Promote staging → prod → live. Only live binds Halo and published reports.

Reports provided

What teams actually open

  • Connection health
  • Named datasets for analytics
  • Halo allow-list on live

Glossary on this page

Words this module uses

Connection
Host, credentials, and engine — like a DataGrip data source.
Dataset
A physical table or a saved SQL script reports and Halo may read.
Live
Approved prod, deployed to main — the only bind Halo and published reports use.

Full glossary

There is no SKUsight-owned table list here. Their database is the source. We connect, write SQL, and name datasets.

Staff work in staging, open a promotion into prod, and an Admin deploys approved prod to live. Halo never reads staging.

One client’s connections never appear on another client. Isolation is the app boundary.

How this differs

Not the same module twice

  • vs Halo

    The semantic model is the workbench and the bind. Halo is the analyst allowed to query live datasets.

    Open Halo
  • vs Retail Enterprise

    Enterprise is the report menu. The semantic model is how those reports reach their warehouse.

    Open Retail Enterprise

30-day data assessment

Connect a subset. See the operating picture. Decide with evidence.

A paid assessment, scoped in writing. If you move to a full contract, the fee is credited back — confirmed commercially, not as a website coupon.

  1. Week 1

    Connect

    We take a subset of your retailers, marketplaces, distributors, and ERP extracts and land them in DataHub.

  2. Week 2

    Health & mapping

    SKU matching, pack standardisation, and a Data Health Score so you can see what is trusted — and what is not.

  3. Week 3

    Opportunities

    Availability, leakage, mix, and coverage exceptions on your own SKUs — the operating picture your teams will actually use.

  4. Week 4

    Insights & executive report

    Diagnostic and predictive framing, plus a leadership pack. If you convert to a full contract, the assessment fee is credited back — in writing.

ConnectHealthMappingOpportunitiesAI insightsExecutive report