Single Source of Truth

A single source of truth (SSOT) is a data management principle that consists of designating a reference repository where each piece of business data (a customer, a sale, a price, a stock level) exists only once and authoritatively. The other systems do not keep their own editable copy: they query or sync against that source, so the whole organization reads the same reliable value.

The problem it solves is old and very common: customer data lives differently in the CRM, in the ERP, in the sales rep's spreadsheet and in the marketing tool. When those values do not match, reports stop adding up and nobody knows which one to believe. An SSOT is the foundation for a reliable KPI dashboard, and it is part of what Metrix puts in order when it consolidates a company's scattered information into a single layer before measuring it.

Why it matters

Without a single source of truth, each department makes decisions with its own version of reality. The sales team says it sold one number, finance reports another and leadership gets a third one in the committee meeting. Time goes into arguing about whose data is right instead of acting on it. An SSOT cuts that friction: there is one agreed value, and the conversation shifts from "is this number right?" to "what do we do with this number?".

How it works in practice

Establishing a single source of truth is not about buying software; it is about making a series of governance decisions about data:

• Designate the system that owns each piece of data: define, data point by data point, which application is the authority. For example, the CRM owns the customer record, the ERP owns the account balance and the inventory system owns the stock.
• Sync, do not copy by hand: connect the systems through integration, APIs or middleware so the data flows from its source, instead of retyping it into every spreadsheet.
• Maintain quality at the source: if the data is dirty at the source, it is dirty everywhere. An SSOT goes hand in hand with data quality practices (deduplicating, normalizing, validating).
• Define each metric only once: make "active customer" or "net revenue" mean the same thing in every report, ideally in a shared semantic layer.

A concrete example (Argentina, B2B)

A wholesale distributor in Buenos Aires has the same customer loaded three times: "Supermercado El Sol SRL", "El Sol S.R.L." and "EL SOL", with a slightly mistyped tax ID (CUIT). Marketing sends it promotions, sales cannot see its full history and collections cannot understand why the debt does not reconcile. Only after defining the CRM as the single source of truth for the account record (with the CUIT as the unique identifier) and merging the three records into one does the portfolio report by customer start to make sense, and the volume discount is calculated on the real total purchases, not on three fragments.

Common mistakes

• Believing that having a data warehouse "already is" the SSOT. The warehouse helps consolidate, but if each upstream system keeps editing its own copy without rules, the inconsistency gets in anyway.
• Designating two owners for the same data. If both the CRM and the ERP "can" edit the customer's address, the contradiction comes back.
• Doing it all at once. It is better to start with the two or three most critical data sets (customer, product, price) before trying to unify the whole universe.

How it differs from a data warehouse

They are often confused because both pursue "consistent data", but they operate at different levels:

In short, a single source of truth is an organizational decision about who is in charge of each piece of data; the data warehouse is one of the tools that help make it real. Adopting an SSOT is, in practice, what turns a dashboard from a nice picture into a credible basis for decisions.