Services
Data architecture, data engineering, business intelligence.
Three capabilities, one team of two. Most real projects need more than one of them — which is precisely why having all three under the same roof matters. The architecture, the pipeline and the report end up agreeing with each other, because the same people built them.
We work with industrial manufacturers across Spain and Europe, and we have delivered in Telco, Defense and Education. Certified on Google Cloud and Microsoft Azure; we build on AWS too.
Data architecture
Before anyone writes a pipeline, somebody has to decide where the data will live, how it will be modelled and who is allowed to see what. That decision outlives every tool you buy on top of it — and it is the one most companies make by accident, one urgent request at a time.
We start from the decisions the platform has to support, not from the product. Half the time that changes the shortlist. The other half it confirms what you were already leaning towards, and now you can defend it with reasoning instead of a vendor deck.
Data architecture in detail →What you get
- A map of the current state: sources, flows, and where things actually break
- A target architecture with the reasoning written down, not just a diagram
- A build sequence with dependencies and rough effort per stage
- The options we rejected, and why
You probably need this if
- You are being quoted for a platform and cannot tell whether it is the right one
- Two reports disagree and nobody can say which one is correct
- Adding a new data source takes a month
Data engineering
The unglamorous part: getting data out of systems that were never designed to share it, cleaning it, and delivering it on a schedule people can rely on. When this works, nobody notices. When it doesn’t, everything downstream is a guess with a chart on top.
Batch or streaming depending on the problem, not on the trend. Real-time costs more to build and considerably more to operate, so we only recommend it when the decision it feeds actually happens in real time. If a nightly run solves your problem, we will tell you that.
Data engineering in detail →What you get
- Pipelines running in production, with tests and alerting
- Every transformation documented — the definition of each field, written down
- Monitoring that tells you something broke before your CFO does
- A handover your own team can actually maintain
You probably need this if
- The monthly close depends on one person running things by hand
- A job fails at six in the morning and somebody restarts it manually
- “The numbers were wrong last week” happens more than once a quarter
Business intelligence
A dashboard is not the deliverable. The deliverable is a number your team stops arguing about. That takes fewer KPIs than most companies build and far clearer definitions than most companies write down.
We build for the person who has to defend the figure in a committee, which means every metric comes with its formula, its owner and its source. It also means we will push back on the request for a fortieth chart nobody opens.
Business intelligence in detail →What you get
- A defined KPI set: formula, owner and source for each one
- Reports designed around the decision, not around the data that happened to exist
- Access and row-level security that matches how your company is organised
- Training, so your team can build their own without breaking the model
You probably need this if
- Three departments report the same metric with three different numbers
- Management still asks for Excel because they don’t trust the dashboard
- You bought Power BI licences and almost nobody logs in
How they combine
One project usually needs two of these. Often all three.
This is what a typical industrial engagement looks like from the inside. The handoffs between the three stages are where most data projects lose weeks — which is exactly the seam that disappears when it is the same two people throughout.
-
01
Decide
Map what exists, agree what the platform has to support, choose the architecture and write down why. Usually two to three weeks.
-
02
Build
Extract, model and deliver. Pipelines in production with tests, alerting and documentation as they are written, not afterwards.
-
03
Surface
Define the KPIs, build the reports, hand over the model and train the people who will own it once we are gone.
You do not have to buy all three. Plenty of clients bring us in for one stage because they already have the rest covered — and that is a perfectly good reason to work together.
How to bring us in
Three ways to work together.
All three capabilities are available under any of these models. Pick the one that fits how your organisation buys, not the one that sounds most impressive.
Fixed-scope project
A defined outcome with a start and an end: an architecture assessment, a pipeline in production, a reporting layer rebuilt. You get a scope, a price and a date before we begin.
Team reinforcement
Your roadmap, your tools, our hands. A senior profile embedded in your team for a defined number of days per month, reporting to whoever already leads the work. No onboarding theatre.
Advisory & enablement
You have the team but not the criteria. We review the architecture, sit in the decisions that matter, and train the people who will maintain the result long after we have left.
These are our standard hourly rates, not a rate card padded for negotiation. Hour estimates on each project page are our best read of the scope, not a fixed quote — we confirm the real range on the first call once we understand your systems. We’ll flag it before we go more than 15% over the original estimate.
Questions
The things people ask before the first call.
What is the difference between a data architect, a data engineer and a BI analyst?
A data architect decides where data lives and how it is modelled — the structural decisions that outlast any single tool. A data engineer builds and operates the processes that move and transform that data. A BI analyst turns the result into defined metrics and reports people use to decide. Architecture is the plan, engineering is the plumbing, BI is the interface.
Do I need all three?
Rarely at the same time. If your data already arrives clean and on schedule, you need BI. If you have a warehouse nobody trusts, you probably need architecture before anything else. Most engagements start with one capability and pull in a second once the first uncovers what is actually broken.
Which cloud do you work with — Azure, AWS or Google Cloud?
All three. We are certified on Google Cloud and Microsoft Azure, and we build on AWS. Which one we recommend depends on what you already run, what your team knows and what you are already paying for. A company with its ERP in Microsoft and a team fluent in SQL usually has a different right answer than a company already invested in AWS.
How long does a data engineering project take?
An architecture assessment takes two to three weeks. A first pipeline in production, from access granted to numbers arriving reliably, typically takes six to ten weeks. A full reporting layer on top of clean data takes four to eight. The variable that moves these dates most is not technical — it is how quickly we get access to the source systems.
How much does a data project cost in Spain?
Small specialist agencies in Spain generally start around 4,000 to 6,000 € for a scoped project, and public programmes such as Kit Consulting price basic data advisory from 6,000 €. Our own starting points are listed above. Anyone quoting a number before understanding your source systems is guessing.
Do you only work with companies in Murcia?
No. We are based in Murcia, Spain, and we work remotely with clients across Spain and Europe, on site when a project genuinely needs it. Most of the work — access, modelling, building, reviewing — happens perfectly well remotely. The parts that benefit from being in the room are the first workshop and the handover.
Who actually does the work?
The two founders. There is no pyramid and no handover to a junior after kick-off — whoever scopes your project is who builds it. That limits how many projects we run at once, which is deliberate. If we cannot take your work on in a reasonable timeframe, we will tell you rather than queue you.
Start here
Not sure which of the three you need?
That is the normal starting position, and figuring it out is the first thing we do anyway. Thirty minutes, no slides. Bring the problem as it actually is and we will tell you what we would do, roughly what it costs, and whether you need us at all.
Or write to info@tobolist.com