Guide library
Getting started (3)
What is AI readiness?What does an AI readiness assessment involve?AI readiness glossaryBuilding the foundations (5)
Process mapping before automationWhen has an SME outgrown spreadsheets?How to choose business software that fitsHow to prepare your data for AIDo SMEs need a data warehouse?By market (5)
AI readiness for Singapore SMEsAI readiness for Malaysian SMEsAI readiness for Indonesian SMEse-Invoicing in Malaysia: a guide for SMEsAplikasi stok barang untuk UMKM (Bahasa Indonesia)The framework
SEL Foundation IndexAll guides →Resources · Data guide
Do SMEs need a data warehouse? The single source of truth explained
Vendors and consultants often tell growing businesses that they need a data warehouse and a “single source of truth”. Both terms describe something useful. Neither is always the right first step for an SME. This guide explains what the terms mean, when a data warehouse is worth having, and the simpler options to try first.
The short answer
What a single source of truth is.
A single source of truth is one agreed place where each of the business’s key figures comes from, so that every team works from the same numbers. A data warehouse is one way to create it, but most SMEs can start with simpler steps.
The problem a single source of truth solves is common. Sales reports one revenue figure, finance reports another, and meetings are spent arguing about which is correct. The cause is usually that each team takes its numbers from a different system or spreadsheet, and calculates them in a different way.
In our view, the most important part of the solution is agreement, not technology. A business needs agreed definitions for its key figures and one official system for each type of data. Once those are in place, a data warehouse can make reporting faster. Without them, a data warehouse simply stores the disagreement in a more expensive place.
The terms
What the terms mean.
These terms are often used as if they mean the same thing. They do not, and the difference affects what you should buy.
| Term | What it means | Why it matters for an SME |
|---|---|---|
| Single source of truth | One agreed place where each key figure comes from, so that every team uses the same numbers. | It is a goal, not a product. It can be reached in several ways. |
| System of record | The one system treated as the official source for a type of data, such as the accounting system for invoices. | Most SMEs need this before anything else. |
| Data warehouse | A central database that collects data from several systems and organises it for reporting and analysis. | One way to reach a single source of truth, usually once data must be combined from several systems. |
| Data pipeline | The automated process that copies data from source systems into a central store, often cleaning it on the way. It is sometimes called ETL, for extract, transform and load. | The part that needs ongoing maintenance when source systems change. |
| Business intelligence (BI) tool | Software that turns data into reports and dashboards. | Can read directly from systems, or from a data warehouse. |
| Data lake | A store for large amounts of raw data in many formats, kept before it is organised. | Rarely what an SME needs first. |
More definitions are in our AI readiness glossary.
Four options
Four ways to reach a single source of truth.
The options below run from simplest to most involved. Most SMEs should work through them in order, and many never need the fourth. Each option depends on the ones before it, which follows the sequence in our SEL Foundation Index: processes come before systems, and systems come before data.
- 01
Agree definitions and a system of record.
Write down one definition for each key figure, such as revenue, gross margin and active customer. Then name the official system for each type of data. This costs staff time rather than money, and it solves many reporting disputes on its own.
- 02
Connect the systems you already have.
Set up integrations so that data moves between systems automatically, for example from the POS to the accounting system. This removes re-keying, which is a common source of mismatched figures.
- 03
Add a reporting tool on top.
Use a business intelligence tool that reads from your connected systems and applies the agreed definitions. For many SMEs, this gives reliable dashboards without a separate data store.
- 04
Build a data warehouse when the need is clear.
When reports regularly combine data from many systems, need long history, or must feed AI tools that use data from several sources, a data warehouse becomes worth its running costs. By then, the first three options have prepared the data it will hold.
How to tell
When a data warehouse is worth considering.
Compare your business with the two columns below. If most of your answers sit in the right-hand column, a data warehouse is likely to be worth the investment.
| Question | Probably not yet | Worth considering |
|---|---|---|
| Where your reports come from | Most reports come from one or two systems. | Reports regularly combine data from four or more systems. |
| How long reports take | Monthly reports take hours, not days. | Staff spend days each month joining exports in spreadsheets. |
| Whether the numbers agree | Teams usually agree on the key figures. | Finance, sales and operations each report different numbers for the same thing. |
| History | Your systems keep enough history for the questions you ask. | You need to compare years of data that sit in old or replaced systems. |
| Your AI plans | Your first AI use needs data from one system. | Your AI use needs data from several systems combined, such as sales, stock and finance. |
These are signs rather than rules. A short assessment can confirm which option fits your business.
Common questions
Questions about data warehouses.
Is a data warehouse the same as a database?
A data warehouse is a type of database. Most business systems have their own database, designed for recording transactions such as sales and payments. A data warehouse is a separate database, designed for reporting and analysis across several systems.
What is the difference between a data warehouse and a data lake?
A data warehouse stores data that has been cleaned and organised for reporting. A data lake stores large amounts of raw data, in many formats, before it is organised. Most SMEs that need either will find a data warehouse more useful, because their questions are about known business figures.
Do we need a data warehouse before we can use AI?
Not usually at the start. Many first AI uses work directly with data from one or two well-maintained systems. A central data store becomes useful when AI needs to combine data from several systems, such as sales, inventory and finance.
What does a data warehouse cost to run?
The cost depends on how many systems feed it, how much data it holds and who maintains it. Besides the setup, budget for the ongoing work of keeping the connections running and the definitions up to date. These running costs are often larger than the setup cost over a few years.
Where should we start?
Start by agreeing definitions for your ten most important figures, and naming one system of record for each type of data. Our free Readiness Snapshot can also show whether data is the foundation most likely to hold your business back.
Find out where you stand
See how ready your foundations are.
The free Readiness Snapshot takes about five minutes. It emails you a PDF report showing where your processes, systems and data stand on the path to AI, and which gap to close first.