In 1989 Nintendo shipped a machine that was already old on the day it launched. The Game Boy carried a monochrome screen with four shades of grey, no backlight, and an 8-bit processor related to those of an earlier generation. A year later Sega answered with the Game Gear: colour screen, backlit, more power, higher resolution. On any spec comparison of the era, the Game Gear won every row.
The result went the other way, and not by a little. The Game Boy, together with its successor the Game Boy Color, sold 118.69 million units. The Game Gear stopped at 10.62 million by March 1996. Eleven of the worse machines for every one of the better ones. The man who chose that grey screen was Gunpei Yokoi, at Nintendo from 1965 to 1996, and he gave his method a name that sounds like a translation error: lateral thinking with withered technology.
And it isn't a story sealed in the nineties. On 6 August 2026 Nintendo reported 3.82 million Switch 2 units in a single quarter, reaching 23.68 million in its first thirteen months: the fastest-selling hardware in the company's history, ahead of the PlayStation 5 in the quarter ending 31 March 2026, with a machine any engineer would place well behind its rivals in raw power.
The comfortable reading is that this is about marketing and games: 'the Game Boy won because it had Tetris and Mario, the technology was irrelevant'. It's half true, and that's why it misleads: it turns the engineering decision into a footnote and the victory into a matter of catalogue or luck. Nothing usable comes from there, because nobody can manufacture a Tetris on demand.
The other reading, more modern, is the praise of the humble: 'you don't need the latest thing, keep it simple'. It sounds good and says nothing operational, because it never explains when simple wins and when it is just worse. There are graveyards full of products built on old technology that lost for exactly that reason.
The error both share is treating the old part as a sacrifice that had to be compensated elsewhere. It wasn't. The grey screen was not the price Nintendo paid to win somewhere else: it was the mechanism by which it won.
This isn't about nostalgia or austerity. It's about what a mature component actually buys, and about the toll the cutting-edge one charges that almost nobody counts.
1. A mature component doesn't cost less: it buys variables. The monochrome screen wasn't merely cheap on the invoice. It drew almost nothing, and that is why the Game Boy ran up to 35 hours on four AA batteries while the Game Gear burned six batteries in three to five hours. It was built on amortised processes, and that is why the launch price was $89.99 against $149.99. It was light, it survived drops, and it was readable in sunlight — which is where a kid actually plays. Battery life, price, robustness and supply are not concessions to engineering: they are the four variables that decide a handheld purchase. The old part didn't sacrifice them, it paid for them.
2. Mature is cheap because its limits are already known. A technology with years of use comes with something the cutting edge does not have: manuals, catalogued failures, multiple suppliers, and people who already know how to work with it. That means you can design right up against its limit without discovering that limit halfway through, and whoever builds on top of you doesn't lose time fighting the part either. The advantage isn't only economic, it's certainty: you know where the wall is before you accelerate.
3. Risk is spent once, and you choose where. Every product has a finite budget of technical risk, money and attention. If you spend it making a new component work at all, you are not spending it making the product good. Nintendo put its entire budget into the use: the form factor, the controls, the price, the games. Sega spent its budget holding up a colour screen that in 1990 still ate the battery. So this mechanism is not 'save money': it is choosing where you want your hard problems to be.
4. The twist: being better on the headline number is what sank Sega. This is the genuinely counter-intuitive part. Colour was not an advantage that wasn't enough; it was the direct cause of the three defeats: it forced six batteries and three hours of runtime, a bigger and heavier device, and sixty dollars more. Superiority in the row that heads the comparison was paid for in the rows that decided the purchase. A spec sheet ranks products by what can be measured in a shop; use ranks them by what you notice on the third hour of a car journey. When those two orderings contradict each other, the second one wins, because it is the one that repeats every day.
Think about the last technical decision you made while building something: the tool, the model, the platform, the language. Be honest about the real reason for the choice. Did you pick it because it solved the problem better for whoever was going to use it, or because it was the option nobody would question, the one that reads well when told?
And the uncomfortable one: if tomorrow someone built your same product with the most boring technology available and sold it at half the price, running three times longer and never failing, what would you have left? Because that person wouldn't have made an inferior version of yours. They would have understood before you which dimension the match was actually being played on.
If the mature component buys variables instead of sacrificing them, then the question of which technology to use is not a matter of technical level but of where you want to compete. The layers:
Not the ones that shine in a comparison: the ones that make someone use it a second time. Price, waiting time, reliability, how long it takes to learn. If your technical choice improves none of those four, you are not making it for the product.
For every mature component the useful question isn't 'what can't it do?' but 'what does it hand me for being amortised?': cost, battery, alternative suppliers, people who already know it, failures already documented. That list is almost always longer than the list of its shortcomings, and it is the one nobody writes.
Decide explicitly which single element of your product is going to be hard, risky or new — and make it the one the user notices. Everything else, mature and without surprises. Two technical bets at once don't double the ambition: they multiply the ways of not arriving.
Facing any decision where the pressure is to use the latest thing, ask which row of the comparison you are winning and which rows you are paying it in, and which of the two orders the real purchase. The other face: mature wins when the customer's problem is no longer one of capability. If what stops people using your product is that it still can't be done, the old part won't save you and the risk has to be spent right there. The method doesn't say 'use old things': it says know which dimension decides, and don't pay for power in the currency that actually matters.
Gunpei Yokoi died in 1997 without being famous outside Japan, and with an idea that keeps sounding like an excuse until you look at the numbers: 118.69 million against 10.62. His method was not settling for less. It was realising that withered technology — old, cheap, its limits already known — is not what you are left with when you can't afford the good stuff, but what lets you afford everything else: the price, the hours of battery, the resistance to being dropped, the calm of whoever builds on top knowing where the wall is. Sega won the comparison in the shop and lost the afternoons. The next time you feel the pressure to use the latest thing so your product gets taken seriously, don't ask which is the best technology. Ask which dimension actually decides whether this gets used tomorrow — and which boring part can buy you that dimension whole.