Essay · AI & Capital Discipline · Part 06 of 12
I've been early twice in ways that cost real money, and both times I told myself I was being visionary.
Early looks identical to right at the moment you commit. The difference only appears later, and by then you've built the org chart around it.
What early actually costs
The obvious cost is the capital. That's the smallest part.
The bigger costs are the ones that don't show up as a line. You standardize on a tool that gets superseded in eighteen months and now migration is a project. You hire for a capability the market hasn't priced yet and you overpay, then the market catches up and your people get recruited away at a premium you set. You educate customers who then buy from the vendor who arrives second with a cleaner product.
In mining we used to say the pioneer takes the arrows and the settler takes the land. That's not a joke about being cautious. It's an observation about who pays the discovery cost and who captures the return.
How to tell early from wrong
The distinction isn't about the technology. It's about which constraint is binding.
If the technology works and your organization can't absorb it, you're early. Absorption is a solvable problem with a known timeline. You can decide to invest in it or decide to wait, and both are legitimate.
If the technology doesn't work yet for your use case, you're not early. You're wrong about the technology, and waiting doesn't fix that, because you'll be equally wrong when it does work if you never understood why it didn't.
People conflate these constantly. "We were too early" is the most comfortable post-mortem available, because it makes the failure a timing error rather than a judgment error. Sometimes it's true. Often it's a way of not looking at the actual mistake.
The thing that makes early defensible
There's one condition under which being early is clearly worth it: when the learning is proprietary and it compounds.
If moving first gets you data nobody else can capture, a workflow position nobody else occupies, or a relationship that gets harder to displace over time, then you're not paying for the technology. You're paying for a position, and the technology is the entry fee.
If moving first just gets you an earlier version of what everyone will have in two years, you paid a premium for nothing. Worse, you paid it in your own credibility, which you'll want later when the timing actually is right.
The version of this I got right
At Microsoft in the nineties I was pushing enterprise deals in markets the company hadn't prioritized. Africa, India, the Middle East. That was early, and it worked, but not because I saw the future more clearly than anyone else.
It worked because being early there bought relationships that stayed bought. The partners we signed when nobody was competing for them were still partners a decade later. The learning compounded. If it had only bought us a head start on a product cycle, it wouldn't have been worth the travel.
On Monday
For any commitment you're about to make ahead of the market, write down what being early buys that being on time wouldn't.
If the answer is a position that gets harder to take later, go. If the answer is a head start on something everyone gets eventually, wait and spend the money on absorption instead.
The operating principle. Being early is only worth paying for when the learning compounds. Otherwise you funded everyone else's tuition.
Juan Vegarra is the author of An Outsider’s Playbook. The views here are his own. More essays · Advisory · Write me