Most digital maturity assessments have the same problem: they look serious, ask a few broad questions, generate a polite radar chart, and then stop exactly where the useful work should begin. They tell you whether a company feels “advanced” or “behind,” but they do not do much to explain why certain capabilities are holding others back, where the real bottleneck lives, or what kind of improvement would actually change the picture.
L'architettura a Digital Maturity Scorecard is a browser-based assessment built by diShine to turn an early discovery conversation into a more structured diagnostic. It evaluates an organisation across Data, AI, esperienza, Governance, e Performance, using thirty evidence-based prompts and a five-level maturity scale. As answers are entered, the tool calculates dimension scores, builds a live radar profile, detects structural patterns, generates cross-dimensional analysis, and outputs a concise Digital Health Report that can be used in workshops, internal reviews, or client conversations.
This is not a generic quiz dressed up as a consulting framework. The interesting part is what happens dopo the answers are entered. Under the hood, the scorecard runs a fairly rich analytical layer: it computes averages and standard deviations, classifies maturity stages, calculates composite strategic indices, estimates digital risk, detects nineteen score patterns, analyzes dependency relationships across dimensions, surfaces the weakest question-level capabilities, and generates context-sensitive strategic commentary from the resulting profile.
What the tool actually is
At a practical level, the Digital Maturity Scorecard is a single-file interactive assessment designed to be fast to deploy and easy to adapt. It runs as a self-contained HTML, CSS, and vanilla JavaScript application, with no framework dependency and no backend requirement for the core experience. That matters because it makes the tool genuinely portable: you can open it locally, host it as a static file, fork it, brand it for a client context if needed, or extend the scoring logic without inheriting a complex application stack.
The scope is deliberately focused. The assessment is built around five dimensions that tend to determine whether digital work is coherent or fragmented.
| Dimension | What it looks at | Why it matters |
|---|---|---|
| Data | Collection, quality, unification, attribution, predictive use | Weak data foundations distort almost every downstream decision |
| AI | Use-case clarity, operational deployment, governance, enablement | AI maturity without operational readiness is mostly theatre |
| esperienza | Journey design, consistency, personalisation, testing, segmentation | Customer experience quality determines whether strategy is actually felt |
| Governance | Ownership, policy, approvals, tooling discipline, cybersecurity, change management | Governance is what allows good ideas to scale without becoming chaos |
| Performance | KPI discipline, experimentation, ROI measurement, optimisation loops | Performance is where strategy proves whether it is working |
Each dimension contains six questions, bringing the full assessment to thirty prompts. The language of the prompts is intentionally operational. The goal is not to ask abstract questions about digital ambition; it is to ask whether the organisation actually has the habits, systems, and decision structures required to turn ambition into repeatable execution
Why we built it that way
A lot of maturity models flatten organisations into a single score and call it insight. The trouble is that digital capability is rarely flat. One company may have strong performance reporting but weak governance. Another may have energetic AI experimentation sitting on top of poor data discipline. Another may have good customer experience instincts but almost no measurement layer to validate whether those experiences are working.
That is why the scorecard is designed to reveal shape, not just level. The live radar profile is useful because it makes imbalance visible immediately. But the more important layer is analytical: the tool tries to explain the consequences of that shape. If Data is weak and AI scores higher, it flags the fact that AI ambition is likely running ahead of the data foundation. If Experience is strong but Performance is weak, it can point out that the organisation may be delivering work it cannot properly prove. If Governance is the floor, it can identify governance as the constraint preventing stronger areas from scaling cleanly.
In other words, the scorecard is less interested in declaring a company “mature” or “immature” than in identifying the specific structural logic of its current state.
What happens under the hood
This is where the tool becomes more interesting than a normal checklist.
The first layer is straightforward scoring. Every question is rated from 1 to 5, ranging from Fragmented to Optimised. Each dimension score is the arithmetic mean of its six question scores. Importantly, the denominator remains fixed at six even when some answers are missing, which means partial completion does not artificially inflate maturity. Incomplete input produces a more conservative dimension score by design.
From there, the tool computes an overall maturity score as the mean of dimensions that have at least one answered question. That score is then mapped to one of eight maturity stages, from Foundational gap to Transformational leader. The result is a classification that is granular enough to be useful in a discussion, without pretending to be more precise than the data justifies.
The second layer is statistical. For each dimension, the scorecard calculates the population standard deviation of the answered question scores. This matters because a dimension average can hide very different realities. A score of 3.0 might mean six uniformly average capabilities, or it might mean a mix of two strong capabilities and four weak ones. Standard deviation helps distinguish between those cases by measuring internal consistency. In practical terms, it shows whether a capability area is genuinely stable or just unevenly assembled.
The third layer is pattern detection. The analytical engine looks at the distribution of dimension scores and checks for nineteen distinct patterns. Some of these are broad distribution patterns, such as uniform, polarized, all low, or one outlier low. Others are relationship patterns, such as dataAiGap, governanceBottleneck, experienceDisconnect, or measurementLast. These patterns matter because they turn a set of numbers into a diagnostic interpretation. A low score is one thing. A low score that is also the single bottleneck constraining otherwise stronger capabilities is a much more useful finding.
The fourth layer is cross-dimensional intelligence. Here the scorecard applies explicit rules to identify dependency problems between areas that should reinforce one another. If AI is relatively advanced but Governance is weak, the tool flags the scaling and compliance risk of ungoverned AI adoption. If Data is strong but Performance is weak, it flags a “data rich, insight poor” condition: the infrastructure exists, but it is not yet producing decision quality. This is one of the most valuable parts of the tool, because digital maturity problems usually do not sit inside a single box. They emerge at the junctions between capabilities.
The fifth layer is question-level gap analysis. Within each dimension, the engine surfaces the lowest-scoring questions, so the diagnostic does not stop at “Governance is weak” or “Performance is weak.” It can point to the specific capability holes driving that weakness. That makes the output substantially more usable in follow-up discussions, because it bridges the gap between a high-level assessment and an actual action plan.
The math is simple on purpose, but not simplistic
One thing I want to outline, is that the mathematics are transparent. There is no black-box scoring model pretending to be magic. You can read more details about Algorithm & Mathematical References of the solution here, and about the Scoring Model & Analytical Engine right here.
At the core, the scorecard uses arithmetic means, weighted averages, threshold classifications, and standard deviation. That is a strength, not a limitation. A maturity tool like this needs to be explainable to the people using it. If a consultant cannot explain how a score was produced, or if a leadership team cannot understand why a certain risk level appeared, then the tool creates distance instead of clarity.
The project documentation makes the formulas explicit. Dimension scoring is a simple mean over six questions. Overall maturity is the mean of scored dimensions. Composite indices are weighted combinations of dimensions. Risk is computed on a 0–100 scale by combining four factors: the lowest dimension score, score variance, governance risk, and data risk. The result is then classified into levels from Minimal to Critical.
| Mathematical component | What it does | Why it is useful |
|---|---|---|
| Arithmetic mean | Collection, quality, unification, attribution, predictive use | Keeps the core scoring model readable and auditable |
| Population standard deviation | Measures internal inconsistency within a dimension | Distinguishes stable capability from uneven capability |
| Weighted averages | Builds composite strategic indices | Shows broader readiness across related dimensions |
| Threshold logic | Maps scores to stages, tiers, and patterns | Makes interpretation consistent and explainable |
| Risk aggregation | Combines weak-floor, variance, governance, and data risk | Translates maturity into exposure, not just level |
Those composite indices are worth mentioning because they add another interpretive layer. The scorecard computes a Digital Foundation Index, an Innovation Readiness Index, an Operational Excellence Index, and a Customer Value Index by combining selected dimensions with explicit weights. That allows the output to say something more strategic than “your AI score is 2.8.” It can also say whether the organisation has the foundational conditions required for innovation, whether its operational discipline is keeping pace, or whether customer-facing effort is supported by data and measurement.
There is also a compact Maturity DNA fingerprint, which encodes the five dimension scores into a short alphanumeric string. It is a small detail, but a useful one: it gives each assessment profile a concise signature that can be referenced quickly in discussions or reports.
Why the open-source release matters
This project is not just available to use online. It is released publicly on GitHub under the MIT License, which is an important part of what makes it genuinely useful.
That means the tool is not locked into a closed product model where the only valid relationship is “use our version exactly as we provide it.” You can inspect the code, understand the logic, fork the repository, adapt the questions, change the wording, translate it further, modify the report output, or extend the analytical enginefor your own internal workflow or for client work. If you are an agency, consultant, transformation lead, or in-house team, that openness matters because maturity frameworks become much more valuable when they can be adapted to the context they are being used in.
MIT licensing makes this a tool you can actually work with, not just look at. You can use it personally, adapt it for clients, fork it into your own process, or improve it for a more specialized context, provided you respect the terms of the license.
That openness fits the design of the product itself. The scorecard is intentionally lightweight, self-contained, and readable. It does not hide behind infrastructure. It invites inspection.
What makes the output more than a radar chart
The visible interface is clean, but the real value is that the assessment output does not stop at visualisation.
The tool generates a Digital Health Report with key findings, maturity interpretation, strategic implications, and tailored next-step recommendations. It also supports multiple export formats, including PDF, Markdown, and plain text, which makes it much easier to carry the output into real work: proposal writing, internal reviews, strategy workshops, or the first stage of a consulting engagement.
The important point is that the report is not a generic template blindly repeated from one assessment to the next. The narrative engine adapts to the score profile by considering the overall score level, detected patterns, cross-dimensional relationships, company size, and industry context. In other words, the output is designed to be profile-specific, not merely formatted.
This is the difference between a scorecard that looks useful and a scorecard that actually becomes useful in practice.
What this is not
As always, it is worth being clear about what the tool does not try to be.
It is not a replacement for a full transformation audit. Thirty questions can reveal a lot, but they cannot substitute for stakeholder interviews, data architecture reviews, process mapping, martech audits, analytics validation, or governance analysis. What the scorecard does well is provide a structured first diagnostic that makes those later conversations more focused.
It is also not trying to win through technical complexity alone. The application is dependency-light, runs directly in the browser, and keeps the core model interpretable. That restraint is one of its strengths. Many internal tools become less useful as they become more elaborate. This one is more interesting because it stays compact while still doing enough analysis to move beyond surface-level scoring.
Why we think this format works for pre-audits
There is something very practical about a tool that can start as a self-serve assessment, become the basis of a discovery call, and then turn into a downloadable discussion artifact without forcing anyone into a heavyweight platform.
That flexibility is what makes the Digital Maturity Scorecard a good fit for agencies and consultants, but also for internal teams that want a more disciplined way to talk about digital capability. It gives people a shared frame. It highlights imbalance. It makes hidden constraints visible. And because the logic is documented and the repository is open, it does this without asking users to simply trust a mysterious score.
For us, that is the right balance: simple enough to use immediately, structured enough to be credible, and open enough to be adapted.
If you want to explore it, the live version is already available, and the full codebase is public.
Use it. Fork it. Adapt it. Improve it.
🔗 View the repository on GitHub
Documentation:
Readme | Algorithm & Mathematical Reference | Scoring Model & Analytical Engine | Changelog