A data fabric is an approach to data architecture, not a single product. It connects data held across many systems, companies and locations under one shared layer. That layer holds the definitions, the access rules and a record of where each figure comes from. People can then use the data under the same rules wherever it is held, instead of each team exporting its own copy. 110 Analytics, a data analytics consultancy in Malta, builds the first form of it as a warehouse in the client's own account.
Does this sound like your group?
If several of these are said in your business, the rest of this page is written for you.
- Each company closes its month in its own system, and the group figures are pasted together afterwards.
- Revenue means something slightly different in every company we own.
- The same customer buys from two of our companies, and nobody can see the total.
- We run more than one accounting system, and nobody wants to migrate them all.
- Every team keeps its own copy of the data, and each copy is a little different.
- Nobody can tell me where the number on the board slide came from.
- We want to try AI on our data, but we do not know which version it should read.
- Showing a director their own company's numbers means sending them the whole group file.
Why does group data stay in pieces?
Group data stays in pieces because each company was set up to run itself, not to report alongside the others. Each has its own systems, its own codes and its own way of naming things. A company that joined the group later may have brought its systems with it.
So group reporting is assembled by hand. Someone exports from each company, maps the codes, adjusts for the differences and pastes the result into one file. Every team that needs the data makes its own copy, and the copies can drift apart.
Hours are only part of the cost. When a figure is challenged, nobody can quickly show where it came from or which definition it used. The meeting then turns into a discussion about the number rather than the business.
What is a data fabric, in plain terms?
A data fabric is a way of organising data, not one piece of software you buy. It puts one shared layer over the data in all your systems, companies and locations. That layer holds the agreed definitions, the access rules and a record of where each figure comes from.
The aim is that data can be found and used under the same rules wherever it is held. A team that needs a figure reads it from the governed layer, instead of exporting its own copy and defining it again.
The word is a metaphor: separate threads woven into one piece of cloth. Each system keeps doing its own job. The fabric is what ties their data together, connecting the systems rather than replacing them.
Does a smaller company need a data fabric?
A single company needs the idea behind a fabric more than the full architecture. For established companies turning over roughly €1M to €10M a year, the practical first form is one warehouse in your own account. Every system loads into it overnight, and the definitions are agreed once with your team.
That warehouse is not a rival to a fabric; it is the foundation one grows from. Its definitions, access rules and loads are the same building blocks a fuller fabric uses. The work is designed to carry forward if the business adds a company, a system or a country.
That band describes who we typically work with, not a limit.
One word, one group definition
What sits in the shared layer?
The shared layer is where the rules live, so every report, team and tool reads the data under the same ones.
Agreed definitions
Revenue, margin, customer and the other terms the group counts. Each is agreed once with the people who use it, written into the model and inherited by every report.
Matching rules
How a customer, product or supplier in one company's system is recognised in another's. The rules are written down, not kept in somebody's head.
Access by role
Who sees what, set by role. A company's directors see their own company; the group board sees all of them.
Where each figure comes from
A record of which system and which load a number was drawn from, so a challenged figure can be traced back to its source.
Schedules and flags
When each source loads, and a flag raised by rule when a load fails or a key figure moves off target.
When does a fuller fabric earn its place?
A fuller fabric earns its place when data is spread across several companies, several ERPs or accounting systems, or several countries. At that point, one overnight load into one store may not be the whole answer.
Some companies in the group may need to keep their data in their own account. Some data may be better left where it is, with the shared layer describing it and governing who can reach it. The layer is what keeps the definitions and access rules the same across all of them.
Larger and more complex projects like these draw on the partnerships we have aligned. The front door stays the same: a data strategy call with Glen, then the data audit. The audit is where the right shape for your group is worked out.
Copies everywhere, or one shared layer?
Who owns the data fabric?
You own it outright, from start to finish. The warehouse and the shared layer are built in your environment, under your account, with your credentials.
If we parted company, it would keep running and still be yours: the loads, the definitions, the access rules and the reports. There is no 110 platform underneath it and no licence to renew with us.
Our job is to make the data usable, not to hold it. We ask for read-only access wherever the system allows, and only the access you agree to. In a group, you decide whether ownership sits with the holding company or with each company.
Where does AI fit in a data fabric?
AI sits at the top of a fabric, reading from the governed layer rather than from loose files. The AI layer answers plain-language questions from the same definitions and data as every report, not from a guess.
That matters more in a group than in a single company. Ask an AI tool about group revenue while each company defines revenue differently, and it has no agreed version to read.
So the AI layer is the last thing switched on, once the definitions and access rules can support it.
Spreadsheets, one warehouse or a fuller fabric?
A company or group can bring its data together in three ways, and each suits a different stage.
| Exports and spreadsheets | One warehouse | A fuller data fabric | |
|---|---|---|---|
| Suits | A quick answer, once | An established company | A group with several companies or systems |
| Where the data sits | In whoever's file | One store, in your account | Where it belongs, under one layer |
| Definitions | The builder's | Agreed once | Agreed once, across the group |
| Access | Whoever has the file | Set by role | Set by role and by company |
| Where figures come from | Hard to tell | Recorded in the load | Recorded across every source |
| Updated | When someone gets to it | Overnight | On a schedule agreed for each source |
| Who owns it | Whoever built it | You | You |
What a shared layer changes, in one group
A group in Malta owns a distribution company, a retail company with several shops and a small services company. The distribution company runs an ERP, and the shops run a point of sale and their own accounting system. The services company keeps its books in a separate package. Customer lists and budgets live in spreadsheets.
- Revenue is counted at a different moment in each company, so the companies cannot be compared fairly.
- Trade customers who buy from both distribution and retail appear under different names, so nobody sees their total.
- Each company's directors receive the whole group file, because there is no other way to share their part.
- When the board questions a figure, tracing it back takes a round of emails.
- Each company's systems load overnight into a warehouse in the group's own account.
- Revenue has one group definition, agreed with each company's finance lead, and every report inherits it.
- Trade customers are matched across the companies, so the group can see what each one buys in total.
- Each company's directors open their own numbers; the board sees all of them.
- A challenged figure can be traced to the system and the load it came from.
None of the companies changed its systems. Their data now meets under one set of rules, and the same layer can take in the next company the group adds.
What is in place by the end of the first month
The first month is implementation. Three to four weeks from an accepted quote, a dashboard is live, and this is the foundation underneath it.
The foundation, in your account
On a platform we recommend before the build, based on what the group already runs, and set up under your own credentials.
The first sources, loading overnight
The systems behind the first dashboard, from one or more companies in the group, read-only wherever the system allows.
The first shared definitions
The terms the first dashboard depends on, agreed across the companies involved and written into the model once.
Access by role
Set from the start, so each company and each role sees the numbers it needs.
A live dashboard
The first view on top of the shared layer, current to yesterday.
How a data fabric starts
- A data strategy callThirty minutes with Glen, free, with no pitch deck. You leave knowing where your data is fragmented across the business, and with a straight answer on whether 110 Analytics is the right fit.
- The data auditAbout three hours with Glen and your team, mapping each company's systems and where each figure comes from. Six dimensions are scored with you, five levels each; level 3 is where reports stop depending on a person. The audit closes with a quote.
- The buildThe foundation and the first dashboard, three to four weeks from an accepted quote. Your team agrees the definitions with us; we do the building.
- The fractional teamAn ongoing team keeps the loads running and the definitions current, and connects the next company or system as the group changes. Larger projects draw on the partnerships we have aligned.
What we do not do
- We do not sell a fabric product, a platform or a licence.
- We do not hold your data for you. It stays in your own account.
- We do not replace the systems each company runs. We are not IT.
- We do not start with the full architecture when one warehouse is the right foundation.
- We do not switch on the AI layer before the definitions can support it.
Glen Sultana founded 110 Analytics after a career built where the numbers move daily. Our people have run analytics for tier-one companies, S&P 500 firms and high-growth technology businesses. Glen is the person on the first call, and he runs every data audit himself.
- Built in your environment, under your account
- No lock-in: if we part company, it keeps running and stays yours
- Based in Malta
Straight answers
Does a smaller company need a data fabric?
It needs the idea, not the full architecture. For an established company turning over roughly €1M to €10M a year, the practical first form is one warehouse in its own account. It loads overnight, with definitions agreed once, and it is the foundation a fuller fabric grows from, if the business grows.
Is a data fabric the same as Microsoft Fabric?
No. Microsoft Fabric is the name of a product sold by Microsoft. This page describes the data fabric approach, which can be built on more than one platform. We recommend the platform before the build, based on what you already run.
How is a data fabric different from a data warehouse?
A warehouse is one store that data is loaded into. A fabric is the wider approach: one shared layer of definitions and access rules that can span several stores, companies and systems. In the way we build, the warehouse comes first and the fabric grows from it.
Who owns the data fabric?
You do, outright. It is built in your environment, under your account and with your credentials, so if we parted company it would keep running and stay yours. There is no 110 platform and no licence to renew.
Do the companies in our group have to change their systems?
No. Each company keeps its own systems, and their data is connected rather than replaced. We read from them on a schedule, read-only wherever the system allows.
Can each company see only its own numbers?
Access is set by role, so a company's directors can see their own company and the group board can see all of them. Who sees what is agreed with you during the build.
What does a data fabric cost?
We quote the build once the data audit is done. The figure depends on how many companies and systems are involved and what the first dashboard needs to show. The platform sits in your own account, so the cloud provider bills its running costs to you directly.
How does a data fabric project start?
With a data strategy call: thirty minutes with Glen, free. If there is a fit, the next step is the data audit. It takes about three hours with Glen and your team, maps each system and closes with a quote.
Fabric terms, in plain English
- Data fabric
- An approach that puts one shared layer of definitions, access rules and source records over data held in many places.
- Shared layer
- The part of a fabric that holds the rules: what each term means, who can see what, and where each figure comes from.
- Lineage
- The record of where a figure came from: which system, which load and which rules were applied on the way.
- Ontology
- Your business's terms, each defined once and linked to the others. In a group, it is the set of definitions every company reports against.
- Data warehouse
- One store of copies of your business data, organised for questions. In our approach, the first form a fabric takes.