The Install-Base Subset Trap
Never ship an extension that reaches only a slice of your existing install base.
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 6
- Confidence
- 78%
A console add-on looks cheap: you reuse hardware customers already own and ship a capability upgrade for a fraction of a new platform's cost. The economics run the other way. Console businesses are razor-and-blades, so profit comes from software, and the only number that matters is how many units a developer can amortise a game across. An add-on's addressable market is a strict subset of the base platform's install base, and it shrinks further with every owner who declines to buy it. Developers price that in and refuse to build exclusives; consumers see a thin catalogue and refuse to buy the add-on; the falling attach rate then confirms the developers' judgement. The loop compounds downward. The rule: if an extension can only ever reach fewer users than the thing it extends, ship the next platform instead.
Origin
Extracted from Acquired's post-mortem of Sega's collapse, where two consecutive Genesis add-ons burned the budget and credibility Sega needed for its next generation.
Core principles
- 01In a razor-and-blades platform, profit lives in the software, not the hardware.
- 02The only number a third-party developer cares about is how many units they can amortise a build across.
- 03An add-on's addressable market is a strict subset of the platform it extends, and shrinks from there.
- 04Cheap bill-of-materials is not the same as a cheap strategic bet.
- 05A shrinking catalogue and a shrinking attach rate reinforce each other downward.
How to run it
- 1
Locate the profit pool
Write down which side of the business actually earns margin. In razor-and-blades platforms it is software attach, and hardware is a customer-acquisition cost.
Pro tip If nobody on the team can state the profit pool in one sentence, the add-on decision is being made on gut feel.
- 2
Compute the extension's hard ceiling
Multiply the current install base by a realistic attach rate for a paid accessory. That product, not the total category, is the market a developer sees.
Pro tip Use the attach rate of your last paid accessory, not the marketing forecast.
Watch out Any owner who skips the add-on is permanently outside the market, so the ceiling only ever falls.
- 3
Run the developer test
Ask whether an outside studio would fund an exclusive title against that ceiling. If the answer is no, the catalogue will be thin and the consumer case collapses with it.
Pro tip Ask three real studios rather than modelling it internally; their answer is the market's answer.
Watch out First-party titles can mask the failure for a quarter or two while third parties quietly walk away.
- 4
Price the clean-generation alternative
Put the add-on's ceiling next to the addressable market of a new platform that starts at zero but has no subset constraint. Compare lifetime software revenue, not launch-year unit sales.
Watch out A new platform's slower start is often mistaken for a worse outcome.
- 5
Kill or commit, and say which
If the ceiling is a subset, cancel the extension and move the budget and engineering staff into the next generation. Announce a single roadmap so partners are not asked to bet on two targets.
Pro tip Cancelling loudly buys back more developer trust than shipping quietly.
Watch out Shipping the extension and the next generation together is worse than either alone, because it splits the developer base.
- 6
Set a falsification trigger before launch
If you still ship, agree in advance on the attach rate and date that end the programme, and on who has authority to call it.
Pro tip Write the shutdown memo before launch, while nobody is defending sunk cost.
Watch out Without a pre-agreed trigger, the standard failure mode is a second add-on to rescue the first.
In the wild
In late 1992 Sega shipped a CD-ROM add-on for the Genesis, beating Nintendo and Sony to the new medium. The launch title, Night Trap, sold the idea that full-motion video was the future. But the add-on could only ever be sold to people who already owned a Genesis, and only a fraction of them bought it. Developers, seeing a small and shrinking target, kept their best work on cartridges, which in turn gave consumers no reason to buy the add-on.
→ Roughly 3 million Sega CD units against a Genesis install base of about 30 million, and the add-on was abandoned.
Rather than treat the Sega CD as a lesson about add-on economics, Sega followed it with the 32X in autumn 1994, a 32-bit processor that stacked on top of the Genesis. It carried the same constraint as its predecessor and arrived while Sega was simultaneously rushing a new native 32-bit console to market, so developers were asked to choose between two Sega targets, neither of which had a credible install base.
→ Under one million 32X units sold, and the confusion damaged retailer and developer trust ahead of the Saturn launch.
Common mistakes
Counting the base install as the market
Teams quote the platform's total install base when justifying an add-on. The real market is that base multiplied by accessory attach rate, which is usually under ten percent.
Confusing a cheap BOM with a cheap bet
Reusing existing hardware makes the add-on look low-risk on the cost line. The cost that matters is the engineering, marketing and developer attention diverted from the next platform.
Rescuing a failed add-on with another add-on
The second extension inherits every constraint of the first and adds market confusion. It converts a contained product failure into a credibility failure with retailers and developers.
Is it for you?
Best for
Platform and hardware teams weighing a mid-cycle capability upgrade against committing to a clean next generation.
Not ideal for
Pure services or software businesses where an add-on costs almost nothing to ship and carries no ecosystem promise.
From the transcript
“there is a fundamental problem with what Sega is doing here, and that is that add-on technology for consoles is never really a good idea.”
“So when you're selling an add-on console, you are limiting the target market to the 30 million people that already own a Genesis.”
“Consumers don't want to go buy a platform that's only going to have a very limited number of good games, and it quickly becomes a…”
From the episode
Short: The Death of Sega