The Analytics Translator: the missing link in data projects

April 2026

Does this sound familiar? The management meeting starts. The financial controller opens a PowerPoint with last month's numbers. The operations manager has their own spreadsheet. The sales lead quotes the CRM. Three sources, three stories, one question nobody asks: "What decision are we making with this?"

That is the moment an Analytics Translator makes the difference. Not by adding a fourth report, but by asking the only question that matters.

Why do data projects fail?

Data projects fail in most cases not because of technology but because of a communication breakdown between business and technology. McKinsey introduced the term Analytics Translator back in 2018 and estimated that demand for this role in the US alone would reach two to four million by 2026. The problem McKinsey described has only grown since then.

From the same research: only 18% of business leaders have confidence in their ability to effectively use data. Not because of a lack of tools — Power BI is affordable, AI is available. But because of a lack of translation.

In an enterprise, you have separate data engineers, data scientists and business analysts. The translation problem sits between departments. In an SME, the problem goes deeper: there is no data team. There is a director with a question, a bookkeeper with accounting software, and a line in the budget for "something with data."

What happens next is what I call the contractor's dilemma: you hire a carpenter (Power BI specialist) who can build a perfect cabinet. But nobody asked whether a cabinet was actually needed — maybe you needed a bookshelf. The technology is flawless. The solution answers the wrong question.

The symptoms are recognisable:

The "expensive wallpaper" dashboard. Beautiful to look at, nobody uses it. It doesn't answer a management question — it displays data. Read more about effective dashboard design →

The definition debate. Sales reports €1.1 million revenue, Finance says €980,000. The meeting doesn't discuss strategy — it debates whose numbers are right. Read how data governance solves this →

The Excel fallback. The Power BI dashboard is live, but the operations manager still opens their own spreadsheet. Because "that's where my numbers are."

What does an Analytics Translator actually do?

An Analytics Translator translates business questions into data solutions, maintains KPI definitions in the data model, and translates dashboard results back into concrete actions for management. Three tasks a standard Power BI specialist does not perform.

1. Translate business questions into data questions

Management says: "We want more control over our margins." A technician hears: "Build a margin dashboard." An Analytics Translator asks: "Which margins? Per customer, per product, per channel? Over what period? And what will you do differently once you have those numbers?"

That difference in questioning determines whether the dashboard becomes management information or expensive wallpaper.

2. Maintain the data model and definitions

What is "revenue"? Including or excluding VAT? On invoice date or delivery date? An Analytics Translator locks these definitions into the data model — not in a document nobody reads, but in the technology itself. One source of truth, no more debate. Read more about the Single Source of Truth →

3. Translate data back into action

A dashboard showing inventory is €200,000 higher than last year is descriptive. An Analytics Translator makes it diagnostic: which product categories are stuck? Since when? What is the impact on working capital? And what is the recommendation?

The difference between "here are your numbers" and "here is what you should do with them."

Why is this role critical for SMEs?

In large organisations, the Analytics Translator is often a dedicated position — sometimes even a full team. For SMEs, that is neither realistic nor necessary. What SMEs do need is someone who fills this role. Externally or internally.

Someone who understands what a P&L looks like, how a supply chain works, what a healthy inventory turnover rate is — and who can translate that into a Power BI data model. That combination is rare.

CBS figures from 2025 show that 29.8% of Dutch SMEs with 10-249 employees used AI technology, compared to only 13.8% of micro-businesses. The problem is not that the technology isn't available. The problem is that nobody asks the right question.

Meanwhile, APMG International has offered a formal Analytics Translation certification since 2024 — the concept has moved from niche to mainstream. But virtually all content, training and vacancies target enterprise organisations with existing data teams. SMEs fall outside the scope.

Three signals that your organisation needs an Analytics Translator:

You have dashboards, but nobody looks at them. The problem isn't the technology — it's that the dashboards answer the wrong questions.

You want to "do something with data" but don't know where to start. You don't lack data — you lack direction.

You're considering AI or Copilot but your data isn't in order. AI amplifies whatever you feed it, including the mistakes. Discover which phase your organisation is in →

Analytics Translation in the AI era

AI tools like Microsoft Copilot in Power BI amplify the quality of your data model — in both directions. A clean model with clear column names, documented definitions and correct relationships delivers useful answers. A messy model delivers convincing-sounding wrong answers — at speed.

Microsoft promises you'll soon be able to "just talk to your data" through Copilot. Sounds great. Until you realise that Copilot uses the same messy data as your current dashboards — only faster.

The Analytics Translator is the role that ensures the foundation is solid before you unleash AI on it. Not by programming Copilot, but by translating business logic into a data model that AI can understand. Read more about scalable data architecture →

Governance, definitions, clean data model — these are not technical side issues. They are the conditions under which AI adds value rather than causes damage.

How does Analytics Translation work in practice?

Analytics Translation follows four steps that together ensure data investments actually deliver returns. Not as a one-off project, but as a way of working that fits the rhythm of the organisation. All steps are delivered remotely via video conference and secure file sharing.