Your move this week: the one-bottleneck rhythm

Theory you cannot act on is just trivia, so here is the rhythm to run, starting this week. It is deliberately small, because the whole point is to do one thing properly rather than ten things partway.

Build the view and name the constraint out loud

First, build the stage-by-stage funnel view with benchmarks beside each stage, all the way through retention and expansion. If you cannot see your stages yet, wiring up that measurement is the work for this week and nothing else, because you cannot break a constraint you cannot see. Once you can read the funnel, do the thing most founders never do: name your single constraint out loud. Write it in one sentence. 'My constraint this cycle is the MQL-to-SQL handoff at 11%.' Naming it commits you, and the commitment is half the discipline.

Pick one exploit lever and ship it before you spend

Second, choose exactly one exploit lever for that constraint and ship it before adding any spend. Faster response, tighter qualification, a churn-save flow, a sharper follow-up. One lever, shipped, on the constraint, with the resources you already have. Resist the urge to also tidy up three other stages while you are in there; that urge is the trap this whole playbook exists to break.

Subordinate your agents to the one stage

Third, point your AI agents and channels at that single stage for the cycle and actively hold them off everything else. This is the subordination step, and for a solo operator running on agents it is the highest-leverage habit you can build. Every agent you own should be serving the constraint this cycle, not running its own independent optimisation. If an agent is producing more input than the constraint can process, throttle it. Feeding the bottleneck at the right rate beats flooding it every time.

Re-diagnose on a fixed cadence

Fourth, set a fixed review rhythm, monthly or per cycle, and re-run the diagnosis. When the constraint has moved, and it will move the moment you break the current one, follow it. Do not keep polishing the stage you already fixed out of habit. Growth leadership is constraint-hunting on repeat, and the founders who compound are the ones who run this loop with discipline while everyone else is busy optimising everything. Find the one bottleneck. Break it. Find the next one. That is the machine.