Prioritisation Under Extreme Uncertainty

The era of exponential opportunities is here. There's too much to do. People who master their own attention and other people's attention will win. Here's how to master your own attention.

Startups have very few resources. Like OpenAI in the early days had less than google Deepmind, given their goal of AGI. They could run fewer experiments. So, how does one decide what is worth doing in such a resource constrained + uncertain environment ? How to pick a task.

Most prioritisation methods work well when you have a lot of data and resources and are built for large organisations. For very small teams dealing with extreme uncertainty, you don’t have data early on to make clear decisions for prioritisation. So how do you move forward ?

Based on all my reading on prioritisation and my own experience in last one year as a Startup Founder, here’s what I’ve landed on.

At a high level, there are only 2 things to worry about:

• Survive (Stay in the game)

• Achieve goal

Depending on your situation, you may have to shift your main goal between these 2 modes. Once you know which one is your goal, we can move forward.

Step 1: Goal & KPI Alignment

• Action: Does this task serve the ultimate goal ? If not, cut it.

• Filter: Yes/No. No alignment, no go.

Step 2: Minimum Cut

• Action: If you could only do one thing, would this make the cut?

• Question: Is this the most essential move toward the goal?

• Filter: Score necessity (1-5). Only 3+ survive.

Step 3: Leverage Potential

• Action: Does it unlock big wins—second/third-order impacts, multipliers, or compounding effects?

• Question: Will this create exponential progress down the road?

• Filter: Score leverage (1-5). Aim for 4+ to prioritize game-changers.

Step 4: Resource Fit + 80/20

• Action: Can we do it with what we have (time, money, compute, people)? Can we 80/20 it—get 80% of the value with 20% of the effort?

• Question: What’s the bottleneck? How do we hack it leaner and faster?

• Filter: Score feasibility (1-5). Redesign heavy tasks for less. 3+ to proceed.

Step 5: Learning Speed

• Action: How fast will we get feedback to learn or pivot?

• Question: When do we know if it works or fails?

• Filter: Score speed (1-5). Favor 4+ for quick iteration.

Step 6: Team Fit

• Action: Who’s the best mind to nail it efficiently?

• Question: Does this play to our strengths?

• Filter: Score fit (1-5). 3+ ensures execution edge.

Step 7: Worth-It Score

• Action: Add scores from Steps 2-6 (out of 25).

• Formula: Minimum (5) + Leverage (5) + Resource (5) + Speed (5) + Fit (5) = Total.

• Decision:

  • 20-25: Do now—high impact, low risk.

  • 15-19: Plan soon if resources align.

  • 10-14: Backburner—revisit later.

  • Below 10: Cut unless mandatory.

Step 8: Gut Check

• Action: Ask: “If I skip this, will I regret it in 6 months?”

• Filter: If gut screams yes but score’s low, dig deeper. Otherwise, trust the score. (Tiebreaker: Pick fastest or highest leverage.)

Once you have a list of tasks that pass the above regimen, they are worth doing. The final step is to maximise throughput. Here’s how to do it.

Step 9: Maximise Throughput

• Action: Define the smallest, smartest next step. From the shortlist, can tasks run in parallel? How do we sequence them for least time?

• Question: What’s the immediate move? How do we max throughput?

• Filter: Map dependencies, group parallel tasks, order by bottlenecks or speed.

Test it out, remove steps that are not important for your case. Or if you have data, then use RICE framework.