Small or 1-person data teams
Set the foundations for a trusted data warehouse without over-engineering
Work directly with Michael Kahan to re-design, model and automate the pipeline for one high-value deliverable over 8 weeks so you have an end-to-end example to follow.
Trusted by 30+ data teams
Sound familiar?
You're the primary data person at the company
You inherited the setup, or you built it yourself while learning. It works. But every decision so far has been yours, with no one to tell you if you're doing it the right way, and the business is about to put the next few years of reporting on top of it.
- Views on views on viewsBusiness logic is buried in nested views, stored procedures or reports.
- Inconsistent metricsTwo numbers show up for the same KPI, and nobody is sure which is right.
- Every report is one-offEven small requests start from scratch and turn into painful, time-consuming projects.
- Building straight in productionThere's nowhere else to work, so changes get tested on the live reports.
- Tribal knowledgeNothing is documented and conventions only exist in the minds of one or two people.
- Leadership is asking about AIEveryone wants it, but the data underneath can't support it yet.
None of this gets better on its own. Every month you keep building on the setup you have now is more that has to be cleaned up later.
My approach
The power of one deliverable
When it comes to "modernizing" a data warehouse, most teams try to move everything at once and agree to overly ambitious timelines to get it all done. Months go by, a lot of activity happens, but nothing is finished enough to satisfy the business.
You know you've been working hard. But from the perspective of your stakeholders who rely on the data, nothing has changed. They still don't trust the numbers and continue to go around the new process you're trying to introduce.
That's why I take a different approach.
Together, we identify one deliverable the business actually cares about and re-build just that pipeline all the way through. From source to the data marts, with the workflow around it. It gets finished, it gets used, and the decisions we make along the way become the pattern for the next one.
- You see the whole path, so the gaps show up in a real build instead of a diagram
- Your team learns it by doing it once, and the second one goes faster
- You have something working to point at while the rest is still in progress
How it works
A simple 3 step process
You don't need to build everything at once to deliver a big impact. We pick one deliverable, build it properly, and set the standards your team follows for everything after.
-
Audit the Current Design
First I review your current setup based on 36 proven conventions across ingestion, modeling, and workflow. You get written findings, a recorded walkthrough, and a list of what to fix first.
Deliverable: Architecture Audit & Findings
-
Build a Foundational Pipeline
Then, we'll implement a proper data pipeline end-to-end guided by one clear deliverable. We'll fix issues in real-time during weekly working sessions and ensure you know exactly what to do at each step.
Deliverable: Structured data model & end-to-end workflow
-
Document & Repeat
Lastly, I document how everything works, along with the standards your team now follows. Your team leaves with the confidence to maintain and repeat the new process on their own.
Deliverable: Data Architecture Handbook & Training
The finish line
What you'll have by the end
In just 8 weeks, you'll have a working example of a properly structured data warehouse pipeline in place and know exactly how to repeat it going forward.
- Numbers you can trustbuilt on an organized data model, not scattered queries
- Faster turnaround on requestsusing repeatable conventions with clear documentation
- A strategy not dependent on one personwith standardized workflows, environments and more
The alternatives
What it costs to keep going the way you are
Salary, benefits and more for a role most small-midsize companies cannot keep busy full time.
They build it their way and bill until you stop them. When they leave, you own a system nobody on your team understands.
Your team keeps adding to the current design which makes resolving it even more difficult in the future.
About
Hi, I'm Michael Kahan
Since 2015, I've supported data platforms for startups and enterprises alike but over time recognized a common pattern: small teams wanted to modernize their data infrastructure but were often overwhelmed by the complexity and noise in the space.
That's why I launched Kahan Data Solutions in 2021. To help growing teams build well-structured analytics architectures with modern tools, while avoiding costly mistakes.
Whether you're looking for direct support or just trying to get clarity on your architecture, I'm happy to help.
FAQ
Common questions
Who is this for?
Small data teams (1-4 people) looking to build their data warehouse with best practices and ready to put in the work.
What counts as a deliverable?
An output the business already uses today or has asked for. Common examples are a report/dashboard, data feed, spreadsheet, etc.
Do you do the development, or advise?
Both options are available depending on your needs, which we can discuss further on the intro call.
How is this different from any other agency or consultant?
Most others focus on billable hours and keeping you dependent on their services. For this reason, they also often overlook small (or 1-person) data teams.
Meanwhile, I believe in ownership, education, and personal growth which is why my approach is all about "teaching you to fish". My offer is fixed fee with a clear scope so you know exactly what to expect, rather than an open-ended contract.
What stack do you work with?
The only requirement is that you use dbt for data transformation. Otherwise, I support most modern databases, such as Snowflake, BigQuery, Postgres, etc.
What pipeline components do you help with?
The 3 core areas of focus are Data Ingestion, Modeling and Workflow. This includes items such as setting up your "landing zone", creating a star schema (facts & dimensions), establishing environments, automating testing and database security.
What do the working sessions look like?
Every week we will meet 1:1 for 60 minutes via video call to dive deep into your project, brainstorm solutions and review your work to make sure you stay on the right track with your deliverable.
Between sessions, you'll also have your own private Slack channel to communicate directly with me anytime to ask questions, submit work for review and get the support you need. I'll record custom screenshares, share custom resources and more to keep you moving without adding more meetings to your calendar.
I'm interested, what's the next step?
Click the "Book an intro call" button below to schedule a 30 minute call with me. We'll discuss your current state, your future goals and if there's a good fit to work together.
Lay the data foundations your team can build on
Book a 30-minute intro call. We'll look at where you are today, what's in the way, and whether this is a fit.
30 minutes, no pitch
Book an intro call
A short call to see where your data stands today and what's in the way.
- Where your data and setup are today
- What's blocking the projects you want to deliver
- Whether the engagement is a fit for where you are
You'll leave with a clear next step either way.
Loading the calendar…
