Skip to content
Beach Street Labs logoBeach Street Labs

§ Productized · Fixed scope · Fixed price

Production data lake.
Two weeks.

ADF + Databricks + dbt + a custom app on top. Most consultancies hand you a dashboard. We ship a working interface your team or customers will actually use — and we do it on a calendar your CFO can plan around.

§ The stack

Boring, proven, fast.

We chose this stack because it ships in days, not months. Every piece is industry-standard and well-staffed. No exotic dependencies. No vendor lock-in beyond what your cloud already has.

  1. 01

    Azure Data Factory

    Ingestion + orchestration

  2. 02

    Databricks

    Lakehouse + compute

  3. 03

    dbt

    Transformations + tests

  4. 04

    Custom app (Next.js)

    What your team and customers actually use

§ Week 1

Pipelines and the bronze layer.

By Friday of week 1, real production data is flowing into Databricks. You can already see it.

  1. Day 1–2

    Discovery + access

    Source-system kickoff, Azure tenant access, credential exchange, scope confirmation. By end of day 2 we have the keys and a target schema sketch.

  2. Day 3–4

    ADF pipelines + landing zone

    Ingestion pipelines for every source, scheduled or event-driven. Bronze landing in ADLS Gen2, partitioned and cataloged.

  3. Day 5

    Databricks workspace + bronze

    Workspace provisioned, cluster sized, bronze layer materialized. Initial sanity dashboards so we can see real data flowing.

§ Week 2

Models, the app, and handoff.

The transforms get clean and the app gets built. Day 14 is the handoff — you walk away with a running system and the keys.

  1. Day 6–8

    Silver + gold dbt models

    dbt project with sources, staging, marts, and tests. Gold tables that match how the business actually thinks. Documentation auto-generated from dbt.

  2. Day 9–10

    Custom app build

    Next.js app that reads from the gold layer — the interface your team or customers actually use. Auth, role-based views, the screens that matter most.

  3. Day 11–12

    Testing + handoff

    End-to-end tests, runbook, Loom walkthrough of the whole stack, repo handoff to your engineering team (or our retainer).

  4. Day 13–14

    Iteration buffer

    Two days to absorb whatever showed up — schema surprises, business-logic edge cases, dashboard requests. Built into the price; not a change order.

§ The differentiator

Why a custom app,
not a dashboard.

BI tools optimize for analysts who already know the question. Custom apps optimize for the people who don’t — your sales reps, ops leads, customers, frontline staff.

A dashboard answers what happened. An app answers what should I do, and how do I do it. One produces insight. The other produces action.

Building the app on top of the data lake is also why we can do this in two weeks instead of two months — when you control the transforms and the interface, you can change a column definition and ship a new screen in the same afternoon.

§ What’s included

  • Azure Data Factory pipelines (up to 8 sources)
  • Databricks workspace + bronze/silver/gold layers
  • dbt project with tests, docs, and CI
  • Custom Next.js app deployed to Vercel
  • Documentation + Loom walkthroughs
  • 30 days of iteration support post-handoff

§ What’s not

  • New source-system contracts (you bring API keys; we wire them up)
  • ML model training (separate engagement)
  • Long-term ops + on-call (transitions to your team or our retainer)
  • Multi-tenant productization or white-label work

§ What we need from you

  • Azure tenant access (or we provision a fresh one)
  • Credentials for every source system in scope
  • One SME for domain questions (~45 min/day)
  • One decision-maker available for sign-off

§ FAQ

The questions you should be asking.

Can you do AWS or Snowflake instead?
Yes. The 2-week scope assumes Azure + Databricks because that stack is fast for us, but we adapt to your existing cloud. AWS + Snowflake + dbt is a clean swap; AWS + Redshift + dbt also works.
Who owns the code?
You do. We hand off the repo, the Azure resources, and a runbook on day 14. No lock-in, no ongoing license fees from us.
What if our sources are messier than expected?
We flag it on day 1 — if the scope shifts to 3 weeks, we tell you upfront, before we start, with a fixed adjustment. We don't quietly burn the iteration buffer and then ask for more.
Why a custom app instead of Power BI or Tableau?
BI tools are for analysts. A custom app is for the people who actually act on the data — sales, ops, customers. A dashboard answers “what happened?” An app answers “what should I do, and how do I do it?” That's the difference between insight and execution.
Can we start before signing the SOW?
Yes. We can run a paid 1-week discovery that credits toward the build. Useful when scope is genuinely uncertain.
What happens after week 2?
Three options: (1) take the keys and run it yourselves, (2) retainer with us starting at $8K/mo for ongoing iteration and on-call, or (3) phase 2 build (more sources, ML, customer-facing scale).

§ Pricing

Agency execution.
Studio pricing.

We deliver what a traditional consulting agency would scope as a multi-month, six-figure engagement — in two weeks, fixed scope, on a studio’s overhead. The structure is different, not the rigor. Send us your sources and constraints; we’ll send a fixed number within two business days of a 30-minute scoping call.

Tell us about your data

Reply within two business days