Your data platform cost a lot to build.
It still doesn't work.

We audit what you have, tell you what's actually wrong, and build the improvements. Two phases: audit first, then a fixed-price build. Senior engineer on every project.

Let's talk

What brings most clients here.

The problem:

  • BigQuery costs tripled in six months. No one knows why.
  • Dashboards take 45 seconds to load. People stopped opening them.
  • Pipelines break weekly. You hear about it from tickets.
  • You inherited a platform. Nobody knows what half of it does.
  • Finance says revenue is X. Product says Y. Leadership believes neither.
  • Nobody uses the BI tool. The data isn't trusted.
  • You want AI features but your data isn't structured for them.

What you get:

  • A written audit you can act on, not slideware.
  • A prioritized improvement list, ranked by cost vs impact.
  • Reliable pipelines with alerts before users notice.
  • Queries and dashboards that load in seconds, not minutes.
  • One definition per metric your business runs on.
  • A BI tool people use, because the data is trustworthy.
  • AI features that work, querying governed metrics, not raw tables.

The target. What your platform should look like.

You probably have some of these layers. Some are missing. Some exist but need replacing. The audit tells us which.

Three layers. Same three we improve.

Data Pipelines

Replace legacy Python scripts with Airbyte or Fivetran. Add monitoring and alerting that actually fire. Fix the three pipelines nobody touches because they scare everyone.

Data Architecture & Modeling

Refactor the models nobody understands. Cut BigQuery costs with partitioning, clustering, and query rewrites. Add tests so data quality issues stop reaching your dashboards.

Dashboards, Reports & AI-ready

Migrate off BI tools your team has given up on. Rebuild trust in the numbers by rebuilding the numbers. Configure the AI features on top (conversational analytics, AI agents) so they query consistent definitions, not raw tables.

The tools we work with: yours and ours.

We work with your existing stack first. If something needs replacing, we tell you why and what we'd use instead.

BigQuery
BigQuery Serverless data warehouse with built-in Gemini AI and EU data residency.
Dataform
Dataform SQL transformations built into BigQuery. Zero extra licensing. Our default.
dbt
dbt Industry-standard SQL transformations. Full support for your existing dbt projects.
Airbyte
Airbyte Managed ELT with 600+ connectors and a builder for custom APIs.
Fivetran
Fivetran Fully managed ELT with 600+ connectors and automated schema handling.
Looker
Looker Enterprise BI with LookML governance and Gemini conversational analytics.
Lightdash
Lightdash BI with metrics in code, dbt or native YAML. Flat-rate pricing, unlimited users.
Metabase
Metabase Approachable BI with dashboards, governed metrics, and Metabot AI.
Steep
Steep AI analytics built on governed metrics. Non-technical teams ask, get trusted answers.
Dagster
Dagster Asset-based orchestration. Declare pipelines and dependencies as code.

See all technologies and when we pick each →

How an optimization runs.

Audit (1–2 weeks, €3,000 – €5,000)

Architecture, costs, pipelines, and data quality reviewed. Team interviewed. You get a written report and a prioritized improvement list. Available standalone.

Your decision (no pressure, no fees for deciding)

Proceed with the build, do it yourselves, or stop here. Plenty of clients take the report and execute internally.

Build (8–12 weeks, bi-weekly demos)

We build the improvements from the audit, in the priority order you approved. Fixed price, agreed before we start.

Handover (1–2 weeks)

Training, documentation, recorded walkthroughs. Your team owns the platform when we leave.

Post-launch support (30 days, included)

30 days of reachability for questions, tweaks, edge cases.

Pricing.

Phase 1: Audit
€3,000 – €5,000

1–2 weeks, fixed price, paid 100% upfront. Written audit report and prioritized improvement list. Same audit available standalone.

Phase 2: Build (optional)
€17,000 – €45,000

9–14 weeks (build + handover), priced after the audit based on scope. Fixed price, agreed before we start.

Total range when clients do both phases: €20,000 – €50,000.

What affects the price

Factor Lower price Higher price
Current state Mostly working Fundamentally broken
Technical debt Isolated issues Systemic across layers
Migration scope Keep most tools Replace legacy stack
BI tool migration Stay on current tool Migrate to new tool

Example scenarios

Example: query tuning + cost optimization
€20,000 – €25,000
10–12 weeks

Performance and cost optimization: schema redesign, query tuning.

Example: replace legacy Python with Airbyte
€25,000 – €35,000
10–14 weeks

Tool modernization: move legacy Python to Airbyte or Fivetran, add Dataform.

Example: multi-layer migration
€40,000 – €50,000
12–16 weeks

Major overhaul: multiple migrations, complex data quality issues.

Why is this more expensive than building from scratch?

Two reasons.

Live systems

We improve while your platform is in use. Migrations run in parallel, dashboards don't go dark, pipelines keep flowing while we replace them underneath. Building from scratch doesn't have that constraint.

Interconnection

Most platforms have connected problems. Fixing the slow dashboard means fixing the model, which means fixing the pipeline. The scope is wider than it looks.

Payment

Audit 100% at audit kickoff. Build 50% at start · 30% at the mid-project milestone · 20% on final handover. Fixed price. No change orders without a signed amendment.

Not sure this is the right fit?

No data platform yet? Start fresh.

Build end-to-end from scratch: ingestion, transformations, dashboards. Audit first, then a fixed-price build.

See Implementation

Scope is narrower than a full optimization?

Focused help on one or two layers: Dataform, Looker, BigQuery costs. Fixed-price, 2–6 weeks.

See Expert Services

Already have a working platform?

Monthly retainer for monitoring, incident response, improvements, and a senior engineer on call.

See Ongoing Support

From minutes to seconds.

Plus a 77% drop in infrastructure costs and 119 Looker derived tables collapsed into a single Dataform layer.

Photo of Phoebe Pham
“Dashboards that used to take minutes now load in seconds, and the whole team has real trust in the data again.”

Phoebe Pham BI Lead, Aeyde

Read the Aeyde case study