The Second Human: Who Design Forgot

Every design school teaches the same diagram. A circle labelled "user" sits at the centre; concentric rings radiate outward — needs, context, constraints. ISO 9241-210, the governing standard for human-centred design (HCD), frames its scope around "interactive systems" and the people who operate them.¹ The installer who wires the system into a wall, the technician who replaces its battery four years later, and the recycler who pries it apart at end-of-life are not in any of those rings. They are, at best, edge cases. The data suggests they should be the main case.

437 US Workers Died in Installation and Repair Occupations in 2024

The Bureau of Labor Statistics recorded 437 fatal workplace injuries among installation, maintenance and repair workers in the United States in 2024, down marginally from 443 the year before.² Vehicle and mobile-equipment mechanics alone accounted for 157 of those deaths. The physical profile of the occupation explains part of the toll: 96.2% of workers must perform tasks in low postures, 80.8% spend their shifts standing, and only 7.6% can choose to alternate between sitting and standing.³ Noise exposure affects 91.7%, yet only 38.8% use hearing protection.³

These are not construction labourers caught in structural collapses. These are people servicing products that someone else designed, manufactured and sold. The product defined the posture, the clearance, the fastener. In renewable energy — the sector HCD-heavy firms most like to celebrate — falls account for 80% of fatalities, and 650 photovoltaic installers were injured in the US between 2011 and 2019.⁴ ⁵ Roofing deaths rose 15% between 2020 and 2021, tracking the growth in rooftop solar installations.⁵ Wind turbine maintenance has produced 147 documented worker deaths and 5,778 total accidents worldwide through December 2024.⁶

None of these numbers appear in a typical design brief. The person the product was designed for is not the person it kills.

Glue Is a Design Decision That Costs Someone Else's Hands

The adhesive that bonds a smartphone battery to its chassis is a user-experience decision disguised as a structural one. It makes the device thinner and lighter — for the buyer. For the technician who replaces the battery, or the informal-sector worker who pries the device apart with a screwdriver, it is a source of lacerations, punctured cells, and chemical exposure.

iFixit's repairability scores make the gap measurable. The Fairphone 6 scores 10 out of 10, with screen and battery repairs that are adhesive-free and take minutes.⁷ The iPhone 16 scores 7 out of 10, partly because Apple introduced an electrically released adhesive for its battery in 2024 — a 9-volt current debonds the cell in roughly one minute, residue-free.⁸ That single material substitution eliminated a step that previously required manual prying, heat guns, and risk of puncture. Samsung's Galaxy S25 Ultra scores 5 out of 10; its battery remains, in iFixit's phrasing, "tightly glued."⁷ Google's Pixel 9 sits at the same mark, its battery "stubbornly glued down."⁷

A market is forming around the gap. The debonding-on-demand adhesive tape segment — products from 3M, Nitto Denko and tesa that release under heat, voltage or UV — was valued at $0.62 billion in 2025 and is projected to reach $1.19 billion by 2030, a compound annual growth rate of 14.1%.⁹ The demand exists because the original adhesive was chosen without asking who would undo it.

62 Million Tonnes of E-Waste Meet 322 Authorised Recyclers

The World Health Organisation estimated global e-waste generation at 62 million tonnes in 2022, of which only 22.3% was documented as formally collected and recycled.¹⁰ The remaining 48.2 million tonnes entered landfill, informal processing, or unknown flows. Each discarded device carries the design decisions of the firm that made it — the type of adhesive, the number of fastener varieties, the presence or absence of material markings on plastics. Improper processing releases up to 1,000 chemical substances, including lead, mercury and dioxins.¹⁰

India generates 1.4 million metric tonnes of e-waste annually.¹¹ Just 322 authorised recyclers operate in the country; the informal sector handles an estimated 95% of discarded electronics.¹¹ In Mustafabad, New Delhi's largest informal e-waste hub, workers earn roughly $1 per dismantled mobile phone and $8 for a 12-hour shift, without protective equipment.¹¹ They report frequent hand cuts, infections and chemical burns — injuries whose likelihood is set, in part, by the product architecture they never influenced.

Design for Disassembly (DfD) is not a new idea. It has been discussed in industrial-design literature for three decades. What is new is that regulation is beginning to enforce it. The EU's Ecodesign for Sustainable Products Regulation (ESPR) now mandates that smartphones carry spare parts for a minimum of seven years and large household appliances for up to ten.¹² Physical repairability labels graded A to E became mandatory for smartphones and tablets in the EU from June 2025, and member states must transpose the Right to Repair Directive into national law by July 2026.¹² Housing components must separate without damage using standard tools; technical repair manuals must be freely accessible.¹²

The regulation addresses end-of-sale obligations. It does not prescribe the ergonomics of disassembly, the maximum number of fastener types in a device, or the posture a recycler must adopt to reach a glued cell. The second human is acknowledged but not yet centred.

France's Repairability Index Shows What Self-Reporting Misses

France launched its indice de réparabilité in January 2021 — the first national repairability score, applied initially to smartphones, laptops, televisions, washing machines and electric lawnmowers.¹³ The structure is instructive for what it reveals about institutional blind spots. Manufacturers compute and declare their own scores using a spreadsheet from the Ministry of Environment; no independent verification is required.¹³ The consumer group HOP (Halte à l'Obsolescence Programmée) has questioned the impartiality of the process.¹⁴

The index evaluates spare-part availability, documentation and disassembly — all relevant to the repair technician. But it does not measure the time a trained technician takes to perform a common repair, the force required to release adhesive bonds, or the risk of component damage during disassembly. A device can score well on paper while remaining hostile to the person doing the work.

Samsung improved its documentation and Apple adjusted certain iPhone 13 repair procedures in the index's wake.¹⁴ Those are useful second-order effects. They are not the same as designing the product around the technician from the start.

The Standard Itself Defines the Blind Spot

ISO 9241-210:2019 structures human-centred design around "interactive systems" and the people who use them.¹ Its four principles — understanding context, specifying requirements, producing solutions, evaluating against requirements — are scoped to the user-facing interaction. Maintenance, installation and decommissioning fall outside that scope unless the design team elects to include them.

This is not an error in the standard; it is a boundary condition that most practitioners treat as permission. When a design team conducts user research, they interview buyers, operators, perhaps administrators. The HVAC installer who must reach a condenser coil through a 400mm gap, the wind technician who climbs 60 metres to a nacelle, the recycler in Seelampur who cuts a lithium cell free of its adhesive bed — these are not in the research panel.

IEC 60300-3-14 covers dependability management and maintenance support.¹⁵ MIL-STD-470 and its successors address maintainability in defence procurement. These standards exist. They are not taught in design schools, and they are not referenced in the HCD canon. The field has two bodies of knowledge — one for the person who buys a product, one for the person who keeps it alive — and they do not speak to each other.

What Would Falsify This Argument

If a longitudinal study showed that products designed under strict HCD processes produced lower technician injury rates and faster mean time to repair than products designed without HCD, the argument would collapse. That study does not exist. The absence of the data is itself evidence of the blind spot: no one is measuring because no one is asking.

Alternatively, if the ESPR's new disassembly and spare-part mandates produce a measurable decline in recycler injuries within three years, the regulatory path may prove sufficient without reforming HCD itself. That outcome is plausible but unproven.

One Sentence About What Changes

The next time a design team in Bengaluru or Berlin opens a user-research plan, the question is whether "user" includes the person in Mustafabad who will take the product apart with a screwdriver and bare hands for a dollar.