Most manufacturers who tell me they want to break into defence are aiming at the wrong target. They watch national budgets, read about NATO spending commitments and wait for a tender to appear on a portal. Then they wonder why three years pass and nothing lands.
If you make a component, a subsystem, a pump, a sensor, a connector, an enclosure, you are almost never the entity a ministry contracts with. The ministry contracts with a prime. The prime buys from a short list of system integrators. Those integrators buy from you. Your customer is an engineering manager and a supply chain manager two or three levels below the headline announcement, and they settled their supply base long before the programme went public.
I have sold into Saab, Nammo, QinetiQ and BAE Systems North America. Not one of those started with a ministry. Every one of them started with an engineer.
In most industries you send a datasheet and a price. In defence you enter a qualification process that behaves like a project in its own right, with a timeline, a cost and an owner inside the customer.
Expect design reviews, first article inspection, environmental and shock testing, materials traceability, counterfeit parts controls, security questions if your product carries firmware, and an audit of your quality system. Expect obsolescence management early, because the platform will still be flying, sailing or driving in twenty years and somebody has to guarantee your part still exists.
None of that is unique to defence, it is just tighter. When I was selling engine cooling pumps at Gilkes and pushed to get into Caterpillar's engine build programme, they asked for an open book review of the whole supply chain. I pushed back, then I yielded, then I presented it in a format that suited us rather than one that exposed us. Every serious customer I have dealt with, marine, rail, automotive or defence, runs gateways like that. The manufacturers who lose treat the gateway as an insult instead of the price of entry.
Nobody in defence buys a component. They buy a qualified component that somebody will still be able to supply in fifteen years.
The single biggest lever you have is being designed in while the platform is still being drawn. That is not a pass or fail gate. Plenty of good business is won from the approved supplier list on availability, lead time, obsolescence cover, offset credit or a sharper commercial case, and I have won that way many times. But the design tilts the odds harder than anything else you can do.
The trick is knowing when the design opens. Platforms are not redesigned because a salesman called. They are redesigned when something forces the change.
At Gilkes I spotted a change in emissions regulation and worked out that the engine builders would be compelled to update their engines. New engine builds meant new pump designs. So I went round almost every engine manufacturer in the world and opened the conversation on the regulation, not on the product. Marine first, then rail, then the larger truck engines. That business went from £2.5M to £15.2M, up 508 percent, and most of it came from that one insight. The cycle came down from eighteen to twenty four months to roughly twelve, because deal control, stakeholder mapping and direct technical access beat waiting for procurement to call.
GE came out of an exhibition conversation about a pump sitting on a Caterpillar engine. Low volume, big engines, so we charged for the design work. Around £150,000, and the first time in Gilkes' very long history the company had been paid to design a pump. Defence is full of low volume programmes. If the volume does not justify free engineering, charge for the engineering.
In defence the forcing events are the same in shape. A capability upgrade, a mid life update, an obsolete part with no supplier left, a new regulation, a new threat. Find the forcing event and you have found the way in.
Two things govern where your product can actually go, and both need answering before you spend money chasing a programme.
The first is export control. Depending on content and country of origin you may sit under ITAR, EAR, a UK licence regime or a European equivalent. That governs which end users you can serve, which partners you can talk to and what you can put in an email. Get advice early. It is cheaper than an investigation.
The second is offset and localisation. In the UAE, the Tawazun Economic Program requires an offset obligation worth sixty percent of contract value above a threshold around ten million dollars, backed by a bank guarantee. In Saudi Arabia, GAMI is targeting more than fifty percent localisation of military spending by 2030 and reported 24.89 percent at the end of 2024. Read that as a buying preference with teeth. A prime carrying an offset obligation is hunting for credible local content, and a component maker willing to build, assemble or transfer capability in country solves a problem the prime already has.
That is a partner selection decision, not a paperwork exercise, and it is the one most manufacturers rush. Qualify the partner against your standard before you appoint. Software like Parteloa exists to score partners properly instead of trusting the pitch.
Component makers price the initial award and forget what follows. Operations and support commonly account for the large majority of a platform's whole life cost, with published estimates running from around seventy five percent upwards. Spares, repairs, overhauls, obsolescence redesigns and training all sit downstream of one design win, and they run for decades at margins the first order rarely matches.
So do not judge a defence opportunity on the first purchase order. Judge it on the installed base you would own for twenty years.
If you want to know whether your product has a real route into a defence supply chain, or whether you would spend two years qualifying for a programme that never repays it, the Market Diagnostic gives you a straight go or no go in five working days. Or book twenty minutes and talk it through.
The Market Diagnostic gives you a go or no-go in 5 working days, built on real intelligence.