Repair the onboarding screen first. Three prospects lost in one week are evidence of a live constraint, while a stronger award answer offers only a possible future benefit.
In April 1970, Apollo 13’s crew had a carbon dioxide problem. The lunar landing had already been abandoned after an oxygen tank failed, and the astronauts were using the lunar module as a lifeboat. Its scrubbers were built for two people, not three. The command module had spare cartridges, but they were square and the lunar module’s openings were round.
NASA engineers in Houston had to choose what mattered now. They designed an adapter using materials already aboard the spacecraft, then guided the crew through assembling it. NASA’s account of the Apollo 13 mission documents the failure, the improvised carbon dioxide removal system and the crew’s safe return.
The mechanism matters more than the drama. A large mission had become dependent on one small interface that could no longer be ignored.
The active constraint deserves the night
At 11:47 PM, the founder has two drafts open.
One is an ECOWAS award answer. The argument is sound, but the opening feels weak. Another hour could make the story sharper, improve the description of the market and give the judges a cleaner reason to remember the company.
The second draft is code for an onboarding screen. Three prospects reached that screen during the week and failed to complete the process. Each had already shown enough interest to try the product. The failure happened after attention had become intent.
Both tasks carry deadlines. Only one sits directly inside the path from interested prospect to active user.
That distinction decides the order.
An award application may create credibility, introductions or funding. Those outcomes depend on selection, timing and what happens after selection. Broken onboarding is already producing an observable result. People arrive. They attempt the next step. They leave.
When runway is limited, current evidence should usually outrank imagined upside. Fix the square-cartridge problem first.
Compare consequences, not prestige
Founders often compare tasks by visibility. Awards, investor updates and public launches feel important because other people will see them. A repair inside onboarding can look smaller, especially when the defect is a confusing field, an unclear button or a missing confirmation state.
The better comparison is consequence.
If the award answer remains imperfect, what changes tomorrow? The application may be weaker, although it can still be submitted if the core answer is accurate.
If the onboarding screen remains broken, what repeats tomorrow? The next prospect can hit the same point, with no reason to assume they will explain why they left.
This is where a simple decision rule helps: prioritise the task whose failure is already recurring and closest to customer value.
The rule does not make every product task urgent. A speculative redesign should not automatically displace a fixed external deadline. The evidence here is narrower. Three prospects encountered the same commercial path in one week, and the founder can identify a repair within that path.
That is a constraint, not a preference.
I use the same reasoning when a contract threatens to pull a small team away from its product. The cheque is visible; the roadmap damage arrives later. Kwame’s decision to protect the core product while extending runway came down to tracing each option through to its operational consequence.
Set a stopping rule before touching either draft
The dangerous version of this night ends with two unfinished tasks. The founder opens the onboarding code, discovers adjacent problems, starts cleaning them up and misses the award deadline.
So the repair needs a boundary.
Fix the specific point that blocked completion. Test the path available to you. Record the broader issues for daylight. Then return to the award response and submit the strongest honest version already on the page.
The stopping rule might be based on scope rather than time: one broken step, one verified path, no redesign. That prevents a customer-facing repair from turning into an unplanned rebuild.
The award draft needs its own boundary. Remove any claim that lacks evidence. Make the decision and result legible. Stop polishing sentences that do not change the judge’s understanding.
This is close to the choice examined in Should I Fix Broken Onboarding or Finish the Award Application?, but the three lost prospects make the priority clearer. The product has supplied evidence that the founder does not need to manufacture.
Leave the system safer than you found it
Apollo 13 did not need the best possible filtration system. The crew needed a workable adapter assembled from what was available, under constraints that had already changed the mission.
A founder’s stakes are smaller, but the decision shape is familiar. Protect the path carrying real demand before improving the story about what the company might become.
Repair the onboarding step. Verify the prospect can get through it. Submit the award answer without pretending another rewrite will remove uncertainty from the judging process.
Before closing the laptop, add one note for the morning: contact the three prospects, tell them the blocked step has been repaired and offer a direct route back in. The repair stops the next loss. The follow-up gives the previous three a reason to try again.
Comments
No comments yet.