Hi there -

Here is this week’s “1 principle, 2 strategies, and 3 actionable tactics” for running lean…

**1 Universal Principle**

**“Prefer Fermi estimates when uncertainty is high.” **

*(*Share this on Twitter

*)*

How do you prioritize work that delivers the biggest impact for the least amount of work?

One popular technique is the RICE framework, where you score and rank each idea’s Reach, Impact, Confidence, and Effort to prioritize what to work on first.

The real challenge arises when assembling the rubric.

- How do you objectively and fairly score reach, impact, confidence, or effort?
- What are the minimum acceptance criteria?
- How do you prevent worse ideas from creeping over better ones?

The answer lies in NOT aiming for 3-digit precision but in order of magnitude estimation — aka Fermi estimation.

*Enrico Fermi was an Italian physicist famous for making rapid order-of-magnitude estimations with seemingly little available data.*

Fermi estimations work by making justified guesses about a problem’s input assumptions, which are accurate within an order of magnitude (the nearest power of ten). If you’ve ever tried to estimate how many pieces of candy there are in a jar, you’ve been exposed to a Fermi problem.

This is often the best we can do with little data, but it’s surprising how useful this kind of ballpark estimate can be in making a decision.

**2 Underlying Strategies at Play**

**I. Aiming for precision is a false proxy when uncertainty is high.**

When uncertainty is high, which is always the case in the early stages of a new product or feature, aiming for precision hides the truth in layers of compounding innocent lies.

For instance, to estimate effort precisely, you might try and

- Scope out the project into tasks and sub-tasks
- Allocate hours to each task
- Then, compute a total

First, doing this across every project requires additional planning overhead. More importantly, every project has unknowable unknowns which are only discovered over time. Both of these challenges are exacerbated in resource-constrained startups.

Furthermore, aiming for precision is a false proxy because it’s easy to fudge these numbers by downplaying effort. When team members do this to make their favorite idea win, it lowers confidence and sparks needless debating.

**II. Fermi estimates eliminate subjectivity.**

The alternative is to forgo precision for confidence in the estimates. Estimation is indisputably more straightforward by only allowing orders of magnitude choices, with little room for argument.

When estimating effort for a project, the only options I can pick from are:

- 2 hours (probably not worth planning as a team)
- 2 days
- 2 weeks
- 2 months

It’s hard to mistake a 2-week project for a 2-month project.

The key to applying Fermi estimates to the scoping problem is picking big enough intervals between choices.

Powers of 10 work in most cases, but when time, conversion rates, or confidence are involved, it helps to use smaller non-linear jumps like 3x or 5x between choices.

In the next section, I’ll share the graduations I use.

**3 Actionable Tactics**

**I. Scoring Reach and Impact**

10x works perfectly well here.

When estimating reach, I ask how many users or customers would be potentially impacted by the project:

- 1
- 10
- 100
- 1,000
- 10,000
- 100,000

For impact, potential revenue (ARR) is what I use most often:

- $100
- $1k
- $10k
- $100k

It’s important to remember that these are absolute numbers but ranges rounded up to the nearest order of magnitude. So, a potential revenue bump of $5k-30k would still be closer to $10k than $1k or $100k.

**II. Scoring Confidence and Effort.**

I already covered effort above. For confidence, I’ll present two options.

**a. Use a graduated scale with the following options: **

- 10% (low confidence it will work)
- 50% (could go either way)
- 80% (high confidence it will work)

I eliminate 0% of 100% because the first is a bad idea, and the second is impossible.

**b. Compute confidence as evidence strength**

If you subscribe to validating demand before you build a product or feature using a demo-sell-build approach, you can make confidence a lot more objective by estimating the number of people that have positively bought into the project already:

- 0
- 1
- 10
- 100
- 1,000

**III. Selecting high ROI ideas.**

If you’ve gotten this far, compute the potential ROI of each project as:

This is not only a faster way of prioritizing the highest ROI projects, but when coupled with an evidence-based confidence score, it also increases your odds of realizing this ROI.

That's all for today. See you next week.

Ash

Author of Running Lean and creator of Lean Canvas

.

.

P.S.

I revived my old YouTube channel. The first video dropped yesterday. I'd love to hear your thoughts.