“You Say Tomato…” Avoiding the Single Source of Truth trap.
The most common challenge companies face when combining data from different sources—whether for simple reporting or advanced automation and AI-based personalization—is determining which data source to use. This challenge is exacerbated when two data sources appear to represent the same thing but yield different results. A classic example is "leads." To marketing, a lead might be someone engaging with a website; to sales, it might be someone who can "open the door" to a new account; and to finance, it might be an account that has started the sales negotiation process. These differences can make organizational alignment difficult and can lead to misunderstandings, mistrust, conflict, and failure.
A typical reaction to this challenge is to call for a "single source of truth"—choosing or creating one data set that everyone understands represents a specific thing. The idea is that if everyone agrees on the definition of a lead, they can choose the appropriate data source and start working.
While well-intentioned, this approach is misguided. Although it may clear up misunderstandings about the term "lead," it doesn't address the root of the issue—there is no single truth. All three definitions of "lead" are valid. Each type of lea" is important to its respective department, but a single source of truth only serves one group.
Determining which definition to prioritize is no easy task. Assuming the groups are even willing to address the issue, they must first understand each other's use cases, agree on which to prioritize, and hopefully agree on alternate names for the other types of leads. This effort can be so time-consuming that it may leave little bandwidth to meet the needs of all groups.
A better approach is to address the issue at its root. Each team has a different definition of "lead." Instead of focusing on the term, focus on the definitions.
Ask each team to describe in detail what they mean by the term "lead."
Describe the term using consistent, structured language that all groups can understand, with a level of detail that leaves no ambiguity.
Capture the definition and make it easily accessible to any person or system that needs to understand it.
Refer to the definitions whenever two teams need to cooperate.
Once these definitions are established, teams might start modifying the common term to "Marketing Leads," "Sales Leads," or "Finance Leads." While this is helpful, it is not critical. The most important aspect—what they mean—has already been captured.
This approach isn’t without its challenges. Creating detailed definitions using commonly understood terms requires effort, prioritization, and resources. The amount of effort and resources depends on how companies approach it. Too little effort or governance will lead to incomplete and unstructured definitions that exacerbate the problem. Too much governance and control will result in rigid solutions that are difficult to use and hinder adoption.
Companies are better served by taking a hybrid approach:
Adopt a single controlled framework that breaks definitions down into a finite set of small, basic, and common descriptive terms.
Integrate this framework into existing workflows and applications to automate or ease the definition capture process.
Develop a rules-based governing model to automate oversight, empowering teams to use and repurpose the data without needing to ask for permission.
I touch on many of these solutions in my earlier posts. I’d love to hear your thoughts. Let me know what approaches your company has taken in the comments below.
