Description
What this service covers
Most disputes over reports are really disputes over definitions, because each team works out its customer count from a separate export. We set up a managed warehouse database, load your data sources on a schedule and model them using a transformation framework. That way a term like active customer is defined once, kept in version control and protected by a test that fails as soon as the underlying data changes. The same foundation supports reporting now and applied AI projects later.
What you receive
- Managed warehouse database set up with defined access roles
- Scheduled loading of up to five sources, such as your app database, payment gateway and ad accounts
- Three modelling layers (staging, intermediate and mart) built with a transformation framework
- Tests for uniqueness, referential integrity and freshness on key tables
- Auto-generated documentation site showing how models depend on each other
- Daily scheduled runs with failure alerts sent to email or team chat
- Handover explaining how to add another model
The work in 4 stages
-
01
Stage 1: Question gathering
We collect the business questions your current reports cannot answer.
-
02
Stage 2: Data design
We trace each source and shape the reporting layer to answer those questions.
-
03
Stage 3: Pipeline build
Loading, models and tests are built, then historical data is backfilled.
-
04
Stage 4: Handover
We walk you through it, hand over documents and answer your questions for two weeks.
Who it suits, and what we need from you
Ideal for
- Businesses where every report comes from a different spreadsheet
- Analysts who spend most of their week combining exports
- Companies heading into due diligence that need figures to agree
What we need from you
- Read access to each source system, database and platform
- An hour each with the owners of any definition that is currently contested
- A decision on where the warehouse will be hosted and whose budget pays for it
You share these through the brief on your order page after checkout.
What changes for your team
- One agreed number for each metric, defined in code and protected by tests
- Problems upstream appear as failing tests, not customer complaints
- Analysts spend their time analysing, not assembling data
Questions about this service
Do we need a premium cloud warehouse, or will an open-source database do?
Our source data is untidy. Can you still work with it?
Are dashboards part of this?
This is a remote service: nothing is shipped physically, and the work is delivered into systems you control. See delivery and handover and refunds and cancellation for the full terms.
Reviews
There are no reviews yet.