Every decentralised industry goes through a phase that, in hindsight, becomes infamous. Telecoms had it in the 1980s. Broadband had it in the early 2000s. Battery storage is in the thick of it now and solar had its own equivalent during its first decade of rapid expansion. It created jobs, momentum and belief at a time when all three were needed.
It becomes easy to oversimplify it as the “cowboy phase,” but that label is usually applied with a level of blame that misses the underlying truth: when a market grows faster than its ability to train, standardise or coordinate, variability is not a moral failing – it is a structural inevitability. The early solar market did not attract bad actors so much as it created conditions no young industry could reasonably withstand.
Low barriers to entry played a part. In those early years you could become a “solar installer” with a van, a ladder, a logo and a certificate from a short course designed more to meet demand than to build depth. Training frameworks were still forming, standards were still settling, supply chains were still working out which components were even appropriate for the British climate. The market was expanding faster than its own competence pipeline and the distance between opportunity and preparation widened by the month.
Method, too, was inconsistent. Not because people were cutting corners deliberately, but because there was no shared definition of what “good” looked like. One company torque-tested every connector; another relied on hand-tightening. One used matching DC connectors; another mixed whatever the merchant had on the shelf. Documentation might be immaculate one week and improvised the next, depending on who surveyed the job or who was available to commission it. Decentralisation invites variation and without a method to channel that variation, the work naturally diverges.
The customer saw none of this. To most building owners, a solar system was binary: panels on the roof, inverter turns on, job done. The subtleties that define long-term quality – torque values, polarity checks, cable routing, IR readings, string configuration, oversizing logic, component compatibility – were invisible. It is difficult to demand excellence in things you cannot see, cannot interpret and often do not know exist. Installers operated in a world where speed was measurable, but quality was not. That is not negligence, that is an information asymmetry baked into the model.
The consequence is the landscape contractors inherit today. Systems without drawings. Arrays with mismatched connectors. DC isolators whose fuses have run hot for years. String configurations that made sense at the time but now limit performance. Earthing arrangements that were adequate under one set of rules but insufficient under another and documentation gaps that turn even simple faults into investigative work.
These problems are not modern, they are historical. They belong not to today’s contractors, but to yesterday’s incentives and yet it is today’s contractors who carry the responsibility for resolving them.
Across Britain, the maintenance teams who never installed these systems are the ones diagnosing intermittent faults on cold rooftops, opening isolators that have seen a decade of thermal cycling, tracing cable routes through buildings that have changed hands three times and explaining to asset owners why certain failures have roots older than the company maintaining them. These are not the consequences of poor workmanship. They are the consequences of a market that matured before its support structures did.
Seen this way, the so-called “cowboy era” was not an era of bad intent. It was an era of speed, novelty and incomplete architecture. Solar was becoming mainstream before its systems for training, documentation, handover and maintenance had the chance to form. Early chaos is a predictable stage in any decentralised transition. What matters is not how it began, but what we build next.
The task now is not to point fingers at a past no one can change. It is to create a future where long-term reliability is not a matter of luck. That requires method. It requires shared definitions of competence. It requires documentation that survives company turnover and changes of tenancy. It requires standards that do not constrain innovation but channel it and it requires tools that make good practice easier than poor practice.
The early phase expanded the rooftop landscape.
The next phase must strengthen it.
Contractors did not create the inconsistencies of the past decade – but they will be the ones who define the consistency of the next one and if the first era of decentralised energy taught us anything, it is that growth without structure creates a backlog. The era we are building now has the chance to do more: to create a hive that learns, aligns and improves, no matter who installed the system or when the work was done.
That is how decentralisation matures. Not by avoiding its early turbulence, but by designing the architecture that prevents it from returning.
