Cloud data platforms

Choosing the cloud data platform that fits the work you actually run, and setting it up so your own people can keep it going.

Write a brief Book a meeting

What the work covers

A platform decision reaches into how data arrives and what the monthly bill turns out to be. These are the parts the work tends to touch.

The workload first: What you actually run, from dashboards to machine learning, and which of those should drive the choice.
Platform comparison: The main options read against your own queries and data volumes, instead of a feature grid.
A real cost picture: What the platform costs under your usage pattern, including the charges that do not show up in a list price.
Who will run it: Matching the platform to the skills your team already has, so it does not need specialists you cannot hire.
Setup and loading: Standing up the platform and getting data into it, with the pipelines that keep it current.
Access and separation: Who can reach which data, and how development is kept apart from what runs in production.
Cost controls: Warehouse sizing and query limits set so a single job cannot run the bill up overnight.

Choosing on the workload

Most platform choices are made from a feature comparison and a vendor demo, where the main options all look capable. Each of the leading platforms can run a reporting warehouse or a machine learning workload, so a demo settles very little. The choice that holds up starts from the work you already have: the queries you run most, and how much data they touch.

Cost needs the most care. Each platform prices differently, so the same workload can be cheaper on one and dearer on another depending on how it is used, and teams who read only the per-credit or per-terabyte rate are often surprised by the bill. Most of the money goes on warehouses left running and on queries that scan more data than they need. The useful comparison is your own workload priced on each platform, with the charges that do not appear in the headline rate written out.

The other question is who will run the platform once it is chosen. Databricks suits a team comfortable with notebooks and Spark, and BigQuery suits people who think in SQL and want little to administer; Snowflake is closer to BigQuery in that respect. A platform that needs skills your team does not have, and cannot easily hire, becomes a cost of its own. The aim is a setup your own people can operate, with cost controls in place from the first day.

How to start

The brief

Say what you run on data today and what you want to run next, and which platforms you are already considering.

A small first piece

An engagement usually opens by pricing one of your real workloads on the platforms in contention, so you can see the numbers before committing to a build.

Who does the work

Amos Weiskopf has spent more than twenty years building software and data pipelines, including pipelines that feed machine learning models, so he has worked on the kind of jobs a platform has to carry.

Questions buyers ask

What is a cloud data platform?

A cloud data platform stores your data and runs queries and analytics on it without servers of your own. Snowflake and Databricks are the best known, along with Google's BigQuery, and each suits a different kind of work.

Which cloud data platform should we choose?

There is no single answer. It depends on what you run, how much data you hold and what your team already knows. The right choice is worked out from your own workload and priced on each option.

How much does cloud data platform consulting cost?

It turns on how many workloads are in play and how much setup and migration they need. The price is settled after your written brief has come in.

Will a cloud data platform cost more than we expect?

It can, when a warehouse is left running or queries scan more than they need. Pricing your real workload on each platform before you choose, and setting cost controls at setup, keeps the bill close to the plan.

Can our own team run the platform afterwards?

That is one of the main things the choice should turn on. The platform is matched to the skills your team already has, so it can be operated without hiring specialists you cannot find.

Does our data stay under our control?

Yes. The platform runs in your own cloud account, and the data stays with you throughout. If you want to sign a non-disclosure agreement before sending anything sensitive, that can be done first.

Engagements by written brief.

Work begins from a documented problem statement, not a sales call.

Write a brief Or book a meeting