dbt Consulting
We build and clean up dbt projects on Snowflake: models your reports can trust, tests that catch a bad load before anyone sees it, and documentation your team can maintain without us.
- 1
Raw source tables
Loaded as-is from each system, never edited by hand
- 2
Staging models
One per source table: renamed, typed and deduplicated
- 3
Intermediate models
Joins and business rules, written once
- 4
Reporting models
The tables dashboards read, with each metric defined
- 5
Tests and docs
Checks on every run, and a description for every column people use
Sound familiar?
If one of these is true, this work usually pays for itself.
- The same metric is calculated in several dashboards, and they disagree.
- Business logic lives in BI tool formulas or one-off SQL that only one person understands.
- A source change breaks reports and nobody finds out until a manager asks.
- You have a dbt project, but it grew without a structure and runs slowly.
What we build in dbt
- 01
New dbt projects
Staging, intermediate and reporting layers set up from the start, with naming, folders and run order your team can follow.
- 02
Tests that catch bad data
Uniqueness, null and relationship tests, plus record-count checks against each source, so a broken load stops before it reaches a report.
- 03
Metric definitions
Each business metric written once in a reporting model and documented, so Sigma, Power BI or Tableau all show the same number.
- 04
Cleanup of an existing project
Remove duplicate logic, move formulas out of the BI tool, speed up slow models, and document what each model is for.
Start with the decision, then trace the data
- 1
Start from a number people argue about
We pick a metric that two reports disagree on and trace it through the models to the source.
- 2
Model it once, test it, document it
The fix becomes a reporting model with tests and a written definition, and the reports switch to it.
- 3
Hand over a project your team can run
Code in your repository, runs on your schedule, and notes on how to add the next model.
What the work changed
120 dbt models
One governed model for a multi-brand ecommerce portfolio
Seven operating domains modeled in dbt on Snowflake. One Shopify migration reconciled 10,105 records after 53 tests passed.
Read the case study →
74 dbt models
Snowflake platform run in-house by a national legal organization
Eight source systems modeled in dbt and handed to the internal team, who have run it since.
Read the case study →Free tools and guides
Free tools
Reporting Time CalculatorThe hours and cost of building and reconciling reports by hand, and how long automating them takes to pay back.Calculate- Data Practice Launch PlannerEight quick questions turn into a plan: what is done, what is next, a BI pick and three first actions.
- Build Your DTC Data StackOne pick per layer, from ingestion to BI, with the reasons and the phase to start with.
- BI Tool QuizEight questions about your warehouse, report builders and budget, and a pick among Sigma, Power BI, Tableau, Looker, Metabase, Hex and Omni.
Guides and comparisons
- What Is a Semantic Layer?Why each metric should be defined once, and where that definition should live.
- Klaviyo Dbt Models in SnowflakeThe staging and reporting models we build on Klaviyo events, profiles and campaigns.
- DTC Metric DictionaryEvery metric from gross sales to LTV, with grain, date, sources and owner.
More in the Snowflake guides and free tools.
Start with the problem
You don't need a finished requirements document. Start with a number the team doesn't trust, a report that takes too long, or a warehouse bill that keeps climbing.
Discuss your data problem