You are working hard and it is not compounding. Output is up, hours are up, and the curve is still flat. The usual diagnosis is that you need more discipline or more resources. Usually the real answer is arithmetic, and it is not the variable you have been pushing on.
- Drive, Fidelity and Velocity multiply, so the weakest one sets the ceiling.
- In the model, 0.5 on all three gives 0.125, just under the 0.15 threshold.
- Find the lowest number and change one structural thing about it this week.
Three things determine whether a person, a team or a system compounds: Drive, Fidelity and Velocity. The important part is not the three names. It is the operator between them.
They multiply. They do not add. Which changes everything about where you should spend the next month, and you can find out where that is by scoring yourself in about ten minutes, using the steps below.
Multiply
Additive thinking hides the bottleneck. Multiplicative thinking exposes it.
If the three variables added together, being exceptional at one would compensate for being weak at another. Push Drive to the ceiling and you could carry mediocre Fidelity and Velocity along with it. That is how most people implicitly model their own performance, and it is why "just work harder" feels like a reasonable plan.
Under multiplication, that compensation does not exist. Here is the case that should stop you:
A 0.9 in Drive is a person running at the edge of what they can sustain. Paired with 0.2 in Fidelity and 0.2 in Velocity, the product is 0.036. Effectively nothing. Every additional unit of effort is being multiplied by a number close to zero, which is the mathematical description of burnout with no output.
Your weakest variable is not a weakness. It is a ceiling on the other two.
The threshold
There is a line, and being balanced is not enough to cross it.
You might reasonably conclude that the fix is balance. Even that out, be decent at all three, and the product recovers. It does, but not as far as you would expect.
A solid, respectable 0.5 across all three variables produces 0.125, and the threshold is 0.15. Balanced competence sits just under the line. This is the uncomfortable finding, and it is the reason the framework is worth ten minutes of your attention: most functioning teams are in the bounded regime, working steadily, never quite compounding, and cannot see why.
Below the line, growth is bounded, you get out roughly what you put in. Above it, the system starts generating some of its own momentum. The distance between those two regimes is not a matter of effort. It is a matter of which variable you fix.
The 0.15 figure is a result of the model in the full paper, derived with ideas from percolation theory, which studies how connected structures change state past a critical point.12
The three
What each variable actually measures.
Intent, energy, commitment. The variable everyone already optimises, and usually the one least in need of attention.
Whether your model of the problem is true. High Drive on a wrong model scales the error. This is the variable that quietly sits at 0.2 in most organisations.
Not speed of activity, speed of learning. How long from doing something to knowing whether it worked. Long feedback loops cap everything.
The scorecard rates each one from 0 to 1 across three statements, then has you multiply. The number at the end is less useful than the thing you notice while getting there, which is that one of your three scores is visibly worse than the others and you already knew which.
A worked example
Same effort, different fix.
Imagine a team that scores itself 0.8 on Drive, 0.3 on Fidelity and 0.5 on Velocity. The product is 0.8 × 0.3 × 0.5 = 0.12, below the 0.15 threshold. Now compare two ways to spend the next month.
| Move | Drive | Fidelity | Velocity | Product |
|---|---|---|---|---|
| Where they started | 0.8 | 0.3 | 0.5 | 0.12 |
| Push harder on Drive | 0.9 | 0.3 | 0.5 | 0.135 |
| Fix the weakest variable | 0.8 | 0.5 | 0.5 | 0.20 |
Working harder raises the product by about 12 percent and still leaves the team under the line. Raising Fidelity from 0.3 to 0.5 raises it by about 67 percent and crosses the line. The effort is the same size. What differs is where it lands.
What a structural change looks like.
- Fidelity: put someone in front of the real data every week, and test the one assumption the plan depends on.
- Velocity: shorten the loop between doing something and knowing whether it worked.
- Drive: remove a source of friction instead of adding pressure.
Four scoring mistakes.
Scoring on a good week. Giving all three variables the same number. Rating Drive on effort when it measures intent. Scoring alone, which is why the page ends by asking you to compare with your team.
Do it this week
Find the lowest number. Change one structural thing.
Rate yourself on all three. Multiply. Compare to 0.15. Then name your weakest variable and make one structural change to it this week, not a resolution, a change to how the work is set up. Then hand the scorecard to your team and compare, because the interesting conversation is where your scores disagree.
Effort is the variable you can feel, which is why it is the one you keep pushing. The ceiling is somewhere else. The mathematics does not negotiate.
Sources
- Pustovetova, T. The Scaling Triad: Drive, Fidelity, and Velocity as Determinants of Exponential Value Creation in Autonomous Multi-Agent and Human-Agentic Systems.
- Stauffer, D., and Aharony, A. (1994). Introduction to Percolation Theory (2nd ed.). Taylor and Francis.
Writes The Visionary, field notes from the AI-native frontier. Building in public: practical builds, named methodology, receipts, no fluff.