![\"2x2](\"/article-assets/digital-transformation-pmi/digital-transformation-pmi-matrice-assi-frequenza.jpg\")
A company's digital transformation is the process that redefines how the business works, not just the tools it uses. It involves processes, data, people and the service model. It is not the same as digitizing an activity (which replaces a manual operation with a digital one), nor as introducing a single management system (which is a project, not a program). A sum of isolated adoptions does not produce a transformation: it produces fragmented data and tools that don't talk to each other.
\nSurveys by ISTAT, Italy's national statistics institute, point to a gap in technology adoption between smaller and large Italian companies that remains structural [1]. The lag is not only about tools: it is about method in managing the transition. The sections below cover scope, roadmap, process choices, indicators and common mistakes.
\nThe term \"digital transformation\" covers very different realities: from moving to the cloud to a complete overhaul of the service model. ISTAT surveys show that the gap in technology adoption between smaller Italian companies and large ones remains structural and is not closed by individual one-off investments [1]. The difference between digitizing an activity and transforming the business is the first distinction to establish.
\nDoes a company's digital transformation start from processes or from tools? Buying a new management system doesn't transform the business; it just moves the data problem somewhere else.
\nDigital transformation works on four distinct dimensions. The first is processes: workflows are rethought before being automated or digitized. Digitizing a flawed process makes it faster at producing flawed results. The second dimension is data: the transformation generates and accumulates data that must be governed (who owns it, in what format, with what access). The third dimension is people: new tools require new skills and new behaviors, not just new logins. The fourth dimension is the service model: in some cases the transformation also changes how the business interacts with customers, suppliers and its own team members.
\nThe gap by company size is measured by ISTAT and is widest precisely where skills and organizational coordination are needed: in Italy in 2025, integrated management software is used by 48.8% of companies with 10-249 employees versus 85.9% of those with at least 250, and data analytics tools by 41.9% versus 83.6% [2]. These are measures of tool adoption, not of the quality of the transformation: they tell you how many companies bought, not how many changed the way they work.
\nThe operational distinction is between digitization and transformation. Digitization is replacing a manual activity with a digital equivalent (e.g., e-invoices instead of paper ones): it is necessary but not sufficient. Transformation is redefining the process upstream, using technology to change how the work is done, not just what it is done with. Digitization is a technical component; transformation is an organizational choice.
\nDigital transformation starts from a map of current processes: without that map, technology investment replicates the inefficiencies of the old way of working. Priority processes are usually selected on three criteria: frequency, value generated for the customer, and organizational pain. Starting with low-frequency processes is a typical mistake, because the return on the investment shows up late.
\nWhich processes should you digitize first when resources are limited? Digitizing a process that generates little value doesn't improve it; it just makes it faster at not generating value.
\nThe frequency × value × pain matrix is the operational tool for choosing where to invest first. Frequency measures how often the process is carried out (daily, weekly, monthly): high-frequency processes deliver faster returns on investment. Value generated for the customer measures the process's impact on the result the customer perceives: high-value processes deserve priority attention. Organizational pain measures how much inefficiency, frustration or error the process currently produces: high pain points have the greatest potential for improvement.
\nProcesses that are typically a priority, because they are high-frequency and high-pain, include order management (from acceptance to fulfillment), invoicing and payment reminders, and document management (contracts, certifications, technical documentation). For the process analysis method that guides this selection, it is worth reading how to map your processes.
\nA common mistake is to start by choosing the tool before mapping: software is evaluated on its sales features without having defined which process it must fix and by what success criterion. The result is a tool that meets generic needs, not the specifics of the company's process.
\n\n\n
A digital transformation roadmap works when it combines an 18-24 month horizon with quarterly operating cycles. The OECD survey on the digitalization of small businesses shows how often the sequence stays implicit: a third of the companies surveyed (34%) decide on digitalization from the top down, 13% collect proposals from employees, and about one in five (19%) say they have no structured process at all [3]. The roadmap exists precisely to make the sequence visible before execution.
\nHow many simultaneous initiatives can a company in transformation realistically pursue? A roadmap with every initiative running in parallel is not a roadmap; it is a wish list.
\nThe operational scheme breaks the roadmap into three successive waves. The first wave (0-6 months) is the quick wins: low-complexity, high-visibility initiatives that deliver fast results and build internal trust in the program. Typical examples: automating outgoing invoices, adopting a centralized document management system, standardizing data exchange formats with your main suppliers. The second wave (6-12 months) is the foundations: initiatives that enable the following ones, typically infrastructural. Examples: migrating to the cloud, integrating existing management systems, building a shared data system. The third wave (12-24 months) is evolution: initiatives that use the foundations you have built to transform higher value-added processes or to change the service model for customers.
\nFor each initiative on the roadmap, it helps to define: the process it changes, the expected measurable result, its dependencies on other initiatives, and the internal owner. For integration with the strategic planning that governs resource allocation, see the guide to strategic planning. For the specific process automation tools that often belong to the second wave, see also business process automation.
\nThe real outcome of a digital transformation is the quality of the data left behind. The Bank of Italy finds that the use of advanced digital technologies (cloud, big data, artificial intelligence) is positively associated with productivity in Italian companies with at least twenty employees, but warns that the analysis is purely descriptive and that adoption is uneven across the country [4]. The difference often lies not in the technology but in the governance of the data it produces.
\nWho is responsible for data during and after the transformation? Without an explicit data owner, every new tool creates a new information silo.
\nData governance doesn't require a complex structure: it requires three explicit decisions. The first is ownership: for each type of data (customer, order, invoice, product, team member), who is responsible for its accuracy? The second is quality: by what criterion do you decide whether data is reliable (completeness, consistency, timeliness)? The third is access: who can read, edit and export which data?
\nThe three data quality criteria are operational and measurable. Completeness measures how many records have all mandatory fields filled in (e.g., how many customer records have a tax ID, email address and phone number). Consistency measures whether the same data has the same form across different systems (e.g., the customer code is the same in the management system and in the invoicing system). Timeliness measures how long the data has gone without updates relative to how often it typically changes (e.g., a price list last updated 18 months ago).
\nA common mistake is to treat data governance as a future problem, to be tackled once the company is \"big enough.\" In reality, the best time to set ownership and quality rules is before starting the transformation, when systems are still few and habits are not yet entrenched. For measuring data-based indicators, see how to choose business KPIs.
\nPeople's adoption is the most underestimated component of a digital transformation. In the OECD survey on the digitalization of small businesses, among those who say they have not digitalized, the main obstacle is not price: 43% don't know where to start and 40% say they don't know enough about digital tools, while cost is cited by one company in four. Among those dissatisfied with their level of digitalization, the most cited cause is lack of time for training (43%), ahead of hardware and maintenance costs (37%) and a shortage of skills (27%) [3]. A transformation that doesn't include structured training and transition time produces tools that are paid for and not used.
\nHow much time is reasonable to give a company to absorb a significant digital change? The most efficient tool left unused is worth less than the least efficient tool used well.
\nThe three levers of operational support work on different levels. The first is structured training: not a two-hour course at go-live, but a path that distinguishes initial training (how the tool works), operational training (how it applies to the specific process) and advanced training (how to extract value from the data it produces). The second lever is communicating the why: people who understand the reason for a change adopt it more easily than those who are simply told about the new procedure. The third lever is protected transition time: a defined period (e.g., 4-8 weeks) in which the old and new systems coexist, and usage errors are treated as training signals rather than inefficiencies.
\nThe clearest sign that support is insufficient is the spread of \"workarounds\": people use the new tool for the mandatory functions and go back to the old method for everything else. This pattern can only be reversed with another round of targeted training, not with hierarchical pressure.
\nThe return on a digital transformation is not the \"ROI of the software\": it is the measurable difference in time, errors, data quality and service capacity. You need indicators that the head of a function can read and use in operational decisions, not just dashboards for top management. The distinction between adoption indicators (actual use) and outcome indicators (effect on the process) is central.
\nWhich indicators should you focus on to understand whether the transformation is working? A transformation that isn't measured is not a transformation; it is a purchase.
\nThe essential set of indicators for monitoring a digital transformation falls into two categories. Adoption indicators measure whether people actually use the new tools: active usage rate of the tool (active users / enabled users, monthly), percentage of transactions handled through the new system vs. the old one. Outcome indicators measure whether the process has actually improved: average process times (from input to output), error rate per output (e.g., invoicing errors per month), data quality (percentage of complete and consistent records).
\nFor each indicator, it helps to define the baseline (the value before the transformation), the 6-month target and the 12-month target. Without a baseline, it is impossible to measure improvement objectively. For integration with your operational performance measurement system, see business performance measurement and how to choose business KPIs.
\nThe most common mistakes in digital transformation are not about the technical choice of software: they are about sequence, data governance and supporting people. They take different forms depending on the sector, but some patterns recur. Recognizing them before you start reduces the most expensive waste.
\nWhich of these mistakes does more damage: choosing the wrong tool or not training people? Both lead to the same outcome: technology investments that don't change the way people work.
\nThe six most recurrent mistakes, each with a small operational correction:
\nDigitizing without mapping processes first. The new tool replicates the inefficiencies of the old process. Correction: before choosing the software, document the current process and identify the pain points to fix. The choice of tool follows the process analysis; it doesn't precede it.
\nToo many projects in parallel. Launching three or four initiatives at the same time scatters management's resources and attention. Correction: define a roadmap in waves, with no more than 1-2 active initiatives per quarter, and stick to the sequence even under pressure.
\nUngoverned data. Every new tool introduces a new information silo with no ownership or access rules. Correction: before the go-live of each new system, define who is responsible for the data, at what quality level and with what access.
\nUntrained people. Training is reduced to an initial webinar and then dropped. Correction: structure a three-phase training path (initial, operational, advanced) and allocate the necessary time in the project plan.
\nTop-level indicators only. Only aggregate costs and revenues are measured, with no adoption or process indicators at the operational level. Correction: for each initiative, define at least one adoption indicator and one outcome indicator, with a baseline and a 6-month target.
\nDependence on a single vendor. Choosing proprietary solutions that don't interoperate with the rest of the ecosystem creates lock-in that limits future choices. Correction: assess in advance the options for exporting data, the available APIs and the support for open formats.
\nThe tools described in this article (process mapping, a three-wave roadmap, data governance, supporting people, adoption and outcome indicators) produce different results depending on the context.
\nFor very small companies (under 10 employees), an 18-24 month roadmap may be out of proportion to their capacity to absorb change. In these cases, it is better to start with two well-defined initiatives and measure their outcome before planning the next ones.
\nThe ISTAT data cited [1][2] cover Italian companies with at least ten employees and measure tool adoption, not the quality of the transformation: micro-businesses are outside the scope of observation, and adoption profiles vary widely by sector. The OECD survey [3] is an international survey of companies already active on digital platforms in seven countries, with just 33 Italian respondents: it is an indication of trends, not a measure of any one country's business landscape. How far the conclusions transfer should be assessed against your own specific context.
\nThe association found by the Bank of Italy between advanced digital technologies and productivity [4] is described by the authors themselves as purely descriptive and does not imply a direct causal relationship. Other factors (quality of management, sector, competitive structure) contribute significantly. Digital transformation is a necessary but not sufficient condition for improving performance.
\nWhat is the difference between digital transformation and digitization?\nDigitization replaces a manual activity with a digital equivalent (e.g., paper form → PDF, paper invoice → e-invoice). Digital transformation redefines the process upstream, using technology to change how the work is done, not just what tool it is done with. Digitization is a technical component; transformation is an organizational choice.
\nWhere should you start when resources are limited?\nThe most efficient starting point is mapping the processes with high frequency and high organizational pain. Once 2-3 priority processes have been identified, you select the tool that specifically addresses those processes, not the most complete or most advertised tool.
\nHow long does a company's digital transformation take?\nA transformation that touches the core processes of a company with 15-50 employees typically takes 18-36 months of structured work, in successive waves. The first year is usually devoted to quick wins and foundations; the second and third years to evolving the higher value-added processes.
\nHow do you measure the success of a digital transformation?\nWith adoption indicators (are people actually using the tools?) and outcome indicators (has the process improved compared with the baseline?). Without a baseline defined before the start, it is impossible to measure improvement objectively.
\nWhen is the right time to tackle data governance?\nBefore the start of each new digital initiative, not after. Data governance is easiest to define when systems are still few and habits are not yet entrenched.
\nA useful digital transformation is not a sum of purchased software: it is a coherent redefinition of how the business works, held together by a roadmap that states the sequence before execution. Four elements distinguish a living transformation from a series of purchases: mapping priority processes before choosing tools; a three-wave roadmap (quick wins, foundations, evolution) over an 18-24 month horizon; data governance (ownership, quality, access) defined before go-live; structured support for people, with three-phase training and protected transition time.
\nMeasurement is not an accessory: it is the condition that lets you understand whether the transformation is producing results or only costs. For each initiative, defining a baseline and 6- and 12-month targets before the start makes results readable and grounds decisions to continue or stop in data, not impressions.
\n[1] ISTAT, \"Imprese e ICT — Anno 2024\", Istituto Nazionale di Statistica, 2024. Available at: https://www.istat.it/comunicato-stampa/imprese-e-ict-anno-2024/
\n[2] ISTAT, \"Imprese e ICT — Anno 2025\", Istituto Nazionale di Statistica, 2025. Available at: https://www.istat.it/comunicato-stampa/imprese-e-ict-anno-2025/
\n[3] OECD, \"SME Digitalisation to Manage Shocks and Transitions: 2024 OECD D4SME Survey\", OECD SME and Entrepreneurship Papers no. 62, OECD Publishing, Paris, 2024 (survey of companies active on digital platforms in France, Germany, Italy, Japan, Korea, Spain and the United States). Available at: https://www.oecd.org/content/dam/oecd/en/publications/reports/2024/09/sme-digitalisation-to-manage-shocks-and-transitions_735fc44d/eb4ec9ac-en.pdf
\n[4] Baltrunaite A., Formai S., Linarello A., Mocetti S., \"Ownership, governance, management and firm performance: evidence from Italian firms\", Banca d'Italia, Questioni di Economia e Finanza no. 678, March 2022 (2019 wave of the Invind survey, about 3,200 firms with at least 20 employees). Available at: https://www.bancaditalia.it/pubblicazioni/qef/2022-0678/QEF_678_22.pdf
\nA useful digital transformation is not a sum of purchased software: it is a coherent redefinition of processes, data, people and the service model, held together by a roadmap that states the sequence before execution. A map of priority processes, waves of initiatives over an 18-24 month horizon, data governance and support for people are the four elements that distinguish a transformation from a series of purchases.
\nThe thread that ties these elements together is the fit between the tool and the way of working: software amplifies the process it finds; it doesn't rewrite it. When the process hasn't been mapped and rethought, the tool speeds up inefficiency. For the strategic level upstream, it is also worth reading the guide to business strategy; for the operational mapping of the processes the transformation will need to touch, how to map your processes.
\nA company that governs its own digital transformation stops piling up unused licenses. It has visibility over the sequence of initiatives, knows which data is reliable, and its people actually use the tools it paid for. It is a change of posture within reach of organizations of any size, provided the transformation remains a process choice, not a technology exercise.
","headings":[{"level":2,"text":"Define what really changes when a company starts its digital transformation","id":"define-what-really-changes-when-a-company-starts-its-digital-transformation"},{"level":2,"text":"Map the processes the digital transformation will need to touch first","id":"map-the-processes-the-digital-transformation-will-need-to-touch-first"},{"level":2,"text":"Build an 18-24 month roadmap that survives changing priorities","id":"build-an-18-24-month-roadmap-that-survives-changing-priorities"},{"level":2,"text":"Choose and govern data as a company asset","id":"choose-and-govern-data-as-a-company-asset"},{"level":2,"text":"Support people through the transition without forcing it","id":"support-people-through-the-transition-without-forcing-it"},{"level":2,"text":"Measure the return on the transformation with indicators that are useful for the work","id":"measure-the-return-on-the-transformation-with-indicators-that-are-useful-for-the-work"},{"level":2,"text":"Common mistakes in digital transformation","id":"common-mistakes-in-digital-transformation"},{"level":2,"text":"Limits and conditions of applicability","id":"limits-and-conditions-of-applicability"},{"level":2,"text":"FAQ — Frequently asked questions","id":"faq--frequently-asked-questions"},{"level":2,"text":"Operational summary","id":"operational-summary"},{"level":2,"text":"Sources and references","id":"sources-and-references"}],"tldr":"Is it better to invest in a structured digital transformation program, or to proceed with one-off projects, choosing one tool at a time based on the urgency of the moment? The answer depends on how codified your current processes are, how dependent you are on a single legacy software system, and how many people need to be trained at the same time.","tldrItems":null}