Data maturity assessment: which of the 5 stages is your organization in?

Twentynext sees the same pattern time and again in mid-sized and large organizations: a data maturity assessment only becomes truly valuable when it goes beyond technology and shows how trustworthy your data really is for decision-making today. A proper assessment maps out, in a structured way, how well an organization collects, manages, and uses data—usually across five stages: ad hoc, reactive, standardized, predictive, and transformative. Knowing where you are is the starting point for any realistic improvement plan:

• Stage 1 (ad hoc): data is scattered, definitions vary, and reporting is manual.
• Stage 2 (reactive): reports exist, but producing them takes time and the numbers are not always reliable.
• Stage 3 (standardized): shared data sources and consistent definitions are in place.
• Stage 4 (predictive): Data Science models support proactive decision-making.
• Stage 5 (transformative): data and AI create measurable new business value.

Why data maturity is more urgent than ever (Service management)

Twentynext sees a recurring issue in mid-sized and large organizations: the ambition to use AI is growing fast, but data quality and structure are lagging behind. That is no coincidence. In 2024, 22.7 percent of Dutch companies with 10 or more employees used one or more AI technologies, an increase of nearly 9 percentage points compared to 2023. But there is another side to that story: among companies that had considered AI but were not using it, “lack of experience” was by far the main reason (74.6 percent).

In practice, that “lack of experience” is rarely a technology problem. It is almost always a data maturity problem. Organizations launch an AI initiative before they truly understand the quality of the data they are working with. An assessment fixes that by giving them an honest starting point.

In 2024, five EU-27 countries had a higher share of AI-using companies than the Netherlands, while the EU average stood at 13.5 percent. So the Netherlands is one of Europe’s frontrunners, but there is still room to improve the quality of AI adoption—not just the volume. Data maturity is the lever that makes that possible.

On top of that, the EU Data Act, which becomes applicable on 12 September 2025, introduces new requirements around data portability and data management. Organizations that already have strong data governance in place will be in a far better position to meet those obligations.

What an assessment gives you

A data maturity assessment is more than a score. It delivers three practical outcomes: an honest picture of your current state, a prioritized investment list, and a realistic path to the next stage. Organizations that skip this step often invest in tools they are not ready to use, or in AI applications that fail because the data engineering foundation is missing. You can read more about that foundation in the article on why strong Data Engineering makes your AI project scalable.

When is the right time?

An assessment is useful at any point, but it becomes critical in three situations: just before choosing a BI platform, at the start of an AI initiative, and after a merger or acquisition where multiple data systems need to be combined. In all three cases, an unrealistic view of your own data maturity leads to premature purchases or AI initiatives that stall because of poor data quality.

Get started yourself:

• Check whether your organization has more than three active data pipelines that are not centrally managed.
• Ask five colleagues from different departments to look up the same KPI. Do you get five matching answers? If so, you are likely already in stage 3 or higher.
• Check whether your organization has agreed definitions for key business metrics such as revenue, margin, and customer base. If not, you are likely in stage 1 or 2.

The five stages explained: what defines each one?

"Better to spend one extra hour—or have one difficult conversation—than deliver a half-finished solution."

— Martijn

Stage 1 – Ad hoc is the starting point. Data exists, but there is no structure around it. Reports are created on request by individual employees, usually in Excel. KPI definitions are inconsistent, and decisions are made based on instinct or experience rather than data. Leaders rarely recognize this pattern in their own organization, but Twentynext encounters it regularly in companies that are convinced they have “been doing data for years.”

Stage 2 – Reactive is where basic reporting exists, but creating it is still time-consuming. In a typical stage 2 scenario, an operations manager at a mid-sized manufacturing company working across five different systems—ERP, production MES, Excel summaries, a CRM, and a procurement database—may spend several hours each week manually combining data for the weekly report. Reports are always backward-looking and regularly contain conflicting figures because the underlying data sources are not synchronized. After standardizing into stage 3, where a central data environment integrates those sources, reporting time often drops from hours to minutes, while trust in the numbers rises significantly.

Stage 3 – Standardized is the turning point. Shared data sources, consistent definitions, and automated reporting are in place. Business Intelligence is set up as an ongoing service rather than a one-off project. This is the stage where dashboards stop being “just visuals” and start genuinely supporting decisions. See also why your dashboard doesn’t make decisions—and how to change that.

Stage 4 – Predictive is where Data Science starts contributing in a meaningful way. Models forecast demand, detect anomalies, or segment customers. This is where the CRISP-DM methodology used by Twentynext really shows its value: structured iteration between the business question, data understanding, and modeling produces repeatable insights. The CRISP-DM approach ensures every project follows a traceable path, which is essential when you want to move models into production.

Stage 5 – Transformative is where data and AI create demonstrable new business value. The result may be new products, new markets, or fundamentally redesigned business processes. According to industry benchmarks, fewer than five percent of organizations reach this level consistently. Getting there requires more than technology; it also demands mature governance, where the AI governance framework becomes a prerequisite.

Scorecard by dimension

Each stage looks different across four dimensions: data quality, technology, processes, and culture. See the table below for an overview.

Get started yourself:

• Assess your organization separately on each of the four dimensions. A strong score in one area means little on its own; your lowest score determines your actual stage.
• Ask your IT department: what percentage of your reporting is fully automated? Below 50 percent: stage 1 or 2. Between 50 and 80 percent: stage 3. Above 80 percent, with validated models: stage 4 or 5.
• Identify your biggest data source bottleneck: is it technology, process, or culture? The most valuable next action depends on that answer.

Comparing the five stages across four dimensions

How to run a data maturity assessment yourself

An assessment should not start with a questionnaire. It should start with a business question. Which decisions does your organization need to improve? The answer determines which dimensions matter most in your specific situation.

Step 1: define the scope

Do not try to assess the entire organization at once. Start with one department or one critical business process. A logistics team trying to understand delivery delays has different priorities from a marketing team trying to improve customer segmentation. Focusing on one domain leads to faster and more reliable outcomes.

Step 2: interviews and data audit

Speak with stakeholders from different parts of the organization, including IT, business teams, and data governance staff. Analyze existing data to build quantitative insight into usage patterns and trends. In practical terms, ask each department how many unique data sources they use, how long reporting takes, and how many versions of “the same” report are currently circulating.

Step 3: score the four dimensions

Use the scorecard in the table above as your reference point. Be honest about the lowest-scoring dimension: that determines your real stage. An organization with excellent technology but weak data quality is not in stage 4, no matter how advanced the platform it bought may be.

Step 4: create a realistic growth path

Build a structured roadmap to move from your current maturity level to your target stage. Prioritize actions based on business impact and available resources. In practice, jumping two stages at once is rarely realistic without significantly increasing the risk of failure. Moving from stage 2 to stage 3 is usually achievable within six to twelve months, provided the priority and budget are there.

Get started yourself:

• Set a current and target score for each dimension on a scale from 1 to 5.
• Calculate the average per dimension. Your lowest average shows your bottleneck.
• Define no more than three concrete improvement actions for that bottleneck dimension. More than three splits focus and slows progress.
• Repeat the assessment after six months to measure progress.

What CRISP-DM and AI readiness add in stages 4 and 5

Once an organization reaches stage 4, the nature of the work changes fundamentally. Reporting is no longer the end goal; models become the new product. That is where a structured methodology such as CRISP-DM becomes essential.

CRISP-DM as a quality anchor in stage 4

Twentynext applies CRISP-DM consistently in Data Science projects. For organizations aiming to reach stage 4, that brings three clear benefits: transparency around project status, repeatability of successful approaches, and the ability to loop back when new data requires it. For clients, that means every modeling project follows a traceable structure from business question to deployment, with documented evaluation steps along the way.

Generative AI use cases come with additional requirements. The deployment phase now also includes governance, hallucination monitoring, and prompt version control. Organizations in stage 4 that want to use generative AI need a solid data platform underneath it. As Martijn van Grieken, Director AI Development at Twentynext, puts it: “You do not choose a data platform for today—you choose it for the architectural decisions you do not want to undo three years from now.”

AI readiness as the checkpoint between stage 3 and 4

Not every organization that reaches stage 3 is automatically ready for AI. AI readiness specifically requires labeled or structured training data, governance for model validation, capacity for continuous monitoring, and in-house knowledge to interpret model outcomes. Organizations that lack these elements risk putting an advanced model into production that no one understands or trusts. Twentynext describes that pattern in more detail in the article on when your organization is truly ready for an AI use case.

The regulatory link is direct as well: the European Data Governance Act has applied since September 2023, with Dutch implementing legislation adopted in 2024. Organizations in stages 4 and 5 should also test their data governance against these frameworks.

Get started yourself:

• Check whether your organization has labeled historical data for its primary business question. If not, fix that before starting model development.
• Ask your IT department whether monitoring exists for current dashboards or models. No monitoring means no stage 4.
• Review whether your data governance policy aligns with the requirements of the Data Governance Act and the Data Act. This is a prerequisite for responsible AI use.

Real-world scenario: moving from stage 2 to stage 3 in manufacturing

Imagine an operations manager at a mid-sized manufacturing company with roughly 150 to 200 employees. The business runs on five separate systems: an ERP, an MES, two Excel reports, and a manually maintained quality register. Every month-end close costs the team several days of manual data consolidation, and by the time the report is ready, most key decisions have already been made.

After a data maturity assessment, the team discovers that the bottleneck is not a lack of data, but the absence of standardized definitions and a central data pipeline. In this scenario, moving into stage 3 requires three specific steps: a shared data model with unambiguous KPI definitions, an automated pipeline that integrates all five sources, and a central data and reporting environment that generates the monthly report automatically.

The outcome is predictably positive: the reporting cycle drops from several days to a few hours, reliability improves because manual steps are removed, and the team can shift its focus from collecting data to interpreting it. This is exactly the kind of growth path Twentynext supports for organizations like these, from Eindhoven and at client sites across the Netherlands.

Frequently asked questions

What exactly is a data maturity assessment?

A data maturity assessment is a structured evaluation of how an organization collects, manages, shares, and uses data for decision-making. It measures current capabilities, highlights strengths and gaps, and produces a prioritized roadmap to improve data quality, access, and outcomes. The result is not just a final score, but a concrete growth path for each dimension.

How long does a data maturity assessment take?

The timeline depends on the scope and complexity of the organization. A focused assessment covering one department or business process can usually be completed within two to four weeks, including interviews, data audit, and reporting. An organization-wide assessment for a company with multiple divisions or data systems typically takes six to twelve weeks. The time investment pays off: without an honest baseline, follow-up initiatives are often underestimated or poorly prioritized.

How does Twentynext support a data maturity assessment?

Twentynext carries out data maturity assessments as a first step in Data Engineering, Business Intelligence, and AI implementation projects. The approach combines stakeholder interviews, a technical data audit, and a CRISP-DM-based analysis to evaluate both the technological and organizational dimensions. The output is a concrete improvement plan with priorities, timelines, and deliverables—not a generic report. You can read more about the broader approach on the Twentynext solutions page.

Does every organization need to aim for stage 5?

No—not every organization should aim for stage 5. Not every business needs to reach level 5 across every dimension. A stage 4 organization may be perfectly positioned to achieve its goals. The more relevant question is: which stage best supports your strategy? A manufacturing company that needs reliable monthly reporting may be fully served by stage 3. A healthcare provider looking to improve early diagnosis may need stage 4. Stage 5 is most relevant for organizations where data sits at the core of the business model.

What are the biggest pitfalls in a data maturity assessment?

The biggest pitfall is overestimating your maturity because you focus on your strongest dimension instead of your weakest. A second pitfall is confusing technology purchases with maturity: buying a modern data platform does not automatically make you stage 4. A data governance maturity model can help organizations assess current governance practices and build a clear roadmap for improvement, from ad hoc processes to fully integrated governance frameworks. The third pitfall is treating the assessment as a one-off exercise. Data maturity needs repeated measurement, at least once a year.

Conclusion

Data maturity is not a destination; it is a growth journey. The five stages—from ad hoc to transformative—give organizations an honest frame of reference to understand where they stand and what the most valuable next step is. The most important lesson from Twentynext’s day-to-day practice is simple: your stage is defined by your lowest-scoring dimension, not your highest. Forget that, and you invest in the wrong place.

The fact that 74.6 percent of companies that considered AI but did not adopt it cited “lack of experience” makes one thing clear: for most organizations, the leap to AI is not primarily a technology issue—it is a data maturity issue. An honest assessment is where that journey starts.

For organizations in Eindhoven and across the Netherlands that want to understand which stage they are in, Twentynext offers a structured approach that starts with the business challenge, not the technology. Get in touch via twentynext.nl to discuss what a data maturity assessment could deliver for your organization.