Before purchasing a platform, sensor, or fleet of devices, an organisation must state the question the object is meant to answer. In technology and environment, that question includes raw materials, electricity, maintenance, data, labour, and end of life. Responsible innovation begins by defining a testable need and by identifying limits that cannot be shifted onto other people or the local environment.
The question sets the boundary#
A digital tool is not neutral merely because it is digital. A form can speed up an unnecessary purchase; a dashboard can hide the infrastructure it consumes; an app can turn a right into an inaccessible procedure. The problem must therefore be described before vendors are compared. It should say who experiences it, how often, what alternative exists, and what harm would be avoided. Without that account, features are mistaken for benefits.
A digital tool is not neutral merely because it is digital. A form can speed up an unnecessary purchase; a dashboard can hide the infrastructure it consumes; an app can turn a right into an inaccessible procedure.
The first question also determines scale. If a fault affects twenty devices, an inventory, available parts, and basic training may be better than a full migration. If a building wastes energy through badly set schedules, the diagnosis should inspect controls and maintenance before installing hundreds of sensors. A proportionate response reduces materials, expenditure, and future dependence.
Material, energy, and time#
The cost of technology does not end with a purchase invoice. It includes extraction, manufacture, transport, servers, updates, batteries, packaging, and collection. It also includes staff time spent configuring accounts, resolving incidents, and explaining decisions. Counting that journey does not require a perfect figure; it requires refusing to treat as zero what has been moved outside the project budget.
Energy efficiency deserves identical care. A management system may reduce heating, yet its outcome depends on building quality, selected thresholds, and the people interpreting alerts. Measuring kilowatts without listening to occupants can save in one column while increasing discomfort, travel, or private workarounds in another. Useful measurement comes with context and a route for challenging anomalous results.
Case 1: replace or repair#
Case 1. A library is offered three hundred new computers. Its inventory shows that many faults are replaceable drives and batteries, while screens and cases remain suitable. The question is not which model is fashionable, but which service is failing: access, speed, security, or maintenance. Certified repair, refurbished devices, and a partial purchase are compared with expected warranties and energy use.
The case establishes a practical rule: service life is measured by ability to provide the service, not by launch date. Parts, compatibility, security updates, repair time, and local availability should be documented. Procurement can require manuals, parts for a defined period, and secure erasure at the end. Those terms make circularity a verifiable obligation rather than marketing language.
Case 2: measurement in a public building#
Case 2. A municipality installs occupancy and temperature sensors to adjust ventilation in a sports centre. The pilot detects cold rooms and proposes switching heating off earlier. Cleaning staff point out that readings occur before the busy afternoon period. The schedule is corrected, the adjustment rule is published, and manual control remains. The project improves because situated knowledge changes the design.
In such work, privacy and material conditions connect. A sensor needs power, connectivity, replacement, and sometimes new network equipment. Only necessary signals should be collected, retention must be fixed, and authority to alter rules should be clear. Planning removal prevents abandoned devices and open accounts after a pilot. Reversibility is both an environmental and an administrative requirement.
Case 3: the phone as a requirement#
Case 3. A company assigns shifts through an app requiring a recent phone and continuous mobile data. It reduces office calls, but shifts cost to temporary staff and excludes people who share a device. Assessment must include a lightweight web route, alternative notices, human support, and compensation when work requires connectivity. A free download does not eliminate a technical requirement’s economic and labour effects.
Accessibility is not an appendix. Screen size, language, coverage, battery, plain language, and support hours determine who can use a service unaided. Testing ordinary and edge cases reveals barriers before rollout. Maintenance, public-facing staff, procurement, and affected people should participate; each sees failures that a sales demonstration will not reveal.
Governance, sources, and decision#
Governance makes these questions routine. A short record can state the need, rejected alternatives, life-cycle budget, exclusion risks, owner, and review date. It should include modest indicators: repaired equipment, energy per service, resolved incidents, time without access, and complaints. Indicators do not replace judgment, but they show whether an efficiency promise shifted labour or waste elsewhere.
Institutional sources test assumptions. UNEP’s Global E-waste Monitor documents electronic waste; the ITU publishes connectivity and digital-divide data; the International Energy Agency examines digitalisation and energy; and the European Commission brings together repair and circular-economy policy. These reports do not make a local decision, but they prevent assessment as though the screen were the whole system.
The final decision must be understandable. If a purchase proceeds, it should state the need, retained equipment, responsibility for parts, consumption measurement, and complaint route. If it is delayed, missing evidence should be named. If a tool is withdrawn, data migration, erasure, reuse, and support for dependent people must be planned. Responsibility is shown when stopping an initiative as well as when approving one.
Technology adds value when it expands capability without making its material conditions invisible. Asking the first question prevents impulse acquisition and leaves room to repair, adapt, and learn. Looking beyond the screen does not reject innovation; it requires innovation to be proportionate, legible, and corrigible. That standard lets a technical solution serve people and environment rather than asking them to absorb silent costs.
Public and private procurement can alter the outcome before the first device arrives. Tenders and contracts should value durability, repairability, interoperability, and operational consumption as well as initial price. Requiring comparable technical information prevents vague promises. Ownership of hardware should also be distinguished from licences that block reuse. When one supplier controls repair, parts, and access to data, the organisation needs to know what happens if that supplier changes direction or disappears.
Security alone does not justify accelerated replacement. A device without updates can be a risk, but that conclusion requires checking its actual use, network segmentation, support options, and the availability of maintained software. The answer may be to remove it from a critical task, not automatically send it to waste. Asset and erasure records protect both information and the possibility of a second life.
Place matters in technical decisions. A rural centre with uneven connectivity, a neighbourhood of small homes, and an office with varied shifts do not bear a cloud-dependent solution in the same way. Designing for average conditions creates exclusion where the service may matter most. Assessment should state minimum requirements and provide degraded operation, local copies, or in-person support. Such measures reduce frustration and parallel purchases made to bypass foreseeable failure.
Evaluation continues after launch. At three and twelve months, consumption, repairs, abandonment, and effects on tasks should be reviewed. Short interviews with people who use and sustain the service reveal signals telemetry misses: waits, duplicated steps, anxiety about monitoring, or learning effort. If results do not confirm the original need, correction must be possible without blaming the person who identified the problem.
A just transition distributes benefits and burdens explicitly. When automation saves time, it should say who keeps that time and who takes on new checking work. When a service is centralised, additional journeys and loss of local support should be assessed. Environment is not an abstract compensation offered at the end; it is part of access, health, and work. Equity is tested by examining the exception and offering a genuine route out.
Finally, documentation must be readable outside the technical team. One page in plain language, with dates, owners, and review routes, lets a community discuss priorities. Specialist knowledge remains necessary, but it must not become a barrier to questioning a decision. Explaining uncertainty builds more trust than promising total control. A mature organisation records what it does not know and sets a date to check again.
Training should accompany delivery, yet it cannot transfer responsibility for weak design to each individual. Useful guidance explains what to do after a fault, how to request an adjustment, and when to stop using equipment. Time for learning and maintenance also belongs in the budget. When that work falls informally to one person, a fragile dependency is created and essential labour is hidden. Sharing knowledge and maintaining instructions is a measure of continuity and care.
Inventory data can guide environmental decisions without becoming surveillance. Age, functional condition, repairs, and aggregate consumption are enough for planning; it is not reasonable to turn every individual use into a performance indicator. Proportionality protects trust. The declared purpose should limit collection and later matching, while independent review can check that information is not reused to punish or discriminate.
A retirement schedule deserves its own budget. Keeping devices until nobody knows who owns them encourages information loss and poorly controlled exports. Chain of custody, verified erasure, and agreements with authorised reuse organisations or managers close the process. Donating equipment without checking condition or support can shift a problem to organisations with fewer resources. The best exit preserves value and respects the person who will receive the material.
Reporting should include what did not work. Publishing a corrected schedule, an unexpected maintenance cost, or an access barrier lets other teams avoid repeating it. This transparency is not an exercise in displaying failure; it turns failure into shared knowledge. For it to be credible, the accountable person needs authority to pause purchasing, change configuration, and respond within a known period.
Fuentes / Fonts / Sources: UNEP Global E-waste Monitor: https://www.unep.org/resources/report/global-e-waste-monitor-2024 ; ITU Facts and Figures: https://www.itu.int/itu-d/reports/statistics/facts-figures-2024/ ; International Energy Agency, Digitalisation: https://www.iea.org/topics/digitalisation ; European Commission, Repair: https://environment.ec.europa.eu/topics/circular-economy/repair_en
