Building Two Ventures in Parallel — and Why They Needed Each Other
World Accelerator and Global Development 50 were designed as one system with two halves. The reasoning behind the split, and what it cost to run both at once.
Why not one organisation
The obvious question is why this is two ventures rather than one with two divisions. The answer is that the two halves need different operating rhythms, different talent and different counterparties, and organisations that try to hold both usually end up doing one properly.
Technology delivery runs on iteration, tolerates failure as information, and hires people who expect to ship. Governance work runs on consensus, treats a public failure as a durable cost, and hires people who expect to persuade. Housed together, the faster culture tends to dominate the slower one and the governance work becomes a compliance function attached to a product roadmap. That is precisely the failure mode I was trying to avoid.
Why not one at a time
The sequential alternative was more tempting and I considered it seriously. The reason I rejected it is that each half is substantially less useful without the other, and building them in sequence means several years during which the first half accumulates exactly the weakness the second was meant to address.
Technology deployed into weak institutions produces capability that is captured or abandoned. Institutional reform without technical capability produces improved processes that still cannot deliver a service. I had seen both outcomes closely enough to want neither.
That reasoning still holds. The cost of acting on it was higher than I estimated.
What it cost
Running two organisations in parallel imposes a tax that is easy to underestimate.
Attention is the obvious one. Founder attention is the scarcest input in an early organisation, and halving it is not a proportional cost — it is worse, because the residual half is fragmented rather than merely smaller. Context switching between a product decision and a policy negotiation is genuinely expensive, and I lost more to it than I expected.
The subtler cost was in hiring. Senior people reasonably ask how much of the founder's time they will have, and the honest answer was less than they wanted. I lost candidates over this, and the ones I lost were often the ones who would have reduced the problem.
What I would keep and what I would change
The structural decision was right and I would make it again. Two organisations with distinct cultures, linked by a shared strategy rather than a shared management layer, is the correct shape for this work.
What I would change is the sequencing of leadership. I under-invested in bringing in senior operating leadership early, on the assumption that founder involvement was necessary during formation. That assumption was partly vanity and partly a failure to distinguish between decisions that genuinely required me and decisions that merely felt safer with me involved.
Building the leadership layer earlier would have cost more in year one and bought considerably more in year three. That is the trade I would make differently.