Two projects with the same expected return but different variance get treated as equivalent.
You're running a P&L with $200K left in your annual Capital Allocation Budget. Two projects land on your desk the same week. Project A: rebuild the billing pipeline - your team estimates a 90% chance it returns 15% and a 10% chance it returns 5%. Project B: launch a new Upsell feature - 50% chance it returns 30%, 50% chance it returns 0%. You calculate the Expected Value of each return. Both come out to 14%. Your CFO says pick whichever you want, they're equivalent. Are they?
Expected Return applies Expected Value to Returns - it's the probability-weighted average of all possible return outcomes on a Capital Investment. It gives you a single number to compare unlike projects, but it deliberately hides Variance, which means two investments with the same Expected Return can feel very different when the actual outcome lands.
You already know Expected Value turns a spread of uncertain outcomes into one number: multiply each outcome by its probability, sum them up. And you know Returns measure economic value produced relative to what you invested.
Expected Return is what happens when you combine the two. It's the Expected Value of a return distribution.
Formally: E[R] = p₁·R₁ + p₂·R₂ + ... + pₙ·Rₙ
Where each Rᵢ is a possible return outcome and pᵢ is its probability.
This is the single number you use when comparing Capital Investment options. If Project A has an Expected Return of 14% and Project B has an Expected Return of 9%, all else equal, you pick A.
The key word is all else equal. That phrase is doing enormous work, and most of this lesson is about understanding what it hides.
As an Operator with P&L ownership, you face Capital Allocation decisions constantly: should you invest in cost reduction, Expansion Revenue, a new hire, or a tooling rebuild? These investments have different costs, different Time Horizons, and different likelihoods of success.
Expected Return gives you a common scale. Without it, you're comparing raw dollar amounts ("Project A might make $50K") without accounting for how much capital each project ties up or how likely each outcome is. That's how you end up over-investing in projects that sound big but actually produce mediocre returns per dollar deployed.
But here's the P&L trap: Expected Return is a planning number, not an outcome number. When the quarter ends, you don't get the expected return - you get one actual return drawn from the distribution. If your Budget depends on hitting a specific number and you picked the high-Variance option, you might blow your operating plan even though the math said the Expected Return was fine.
This is why understanding Expected Return matters for operators: it's necessary for comparing options, but it's not sufficient for making good decisions. It's step one of Capital Budgeting, not the whole process.
Calculating Expected Return
Start with the possible outcomes and their probabilities. Suppose you're evaluating a $100K investment in a new feature:
| Outcome | Return | Probability |
|---|---|---|
| Feature ships, adoption is strong | 40% | 0.30 |
| Feature ships, adoption is moderate | 15% | 0.50 |
| Feature ships late, minimal adoption | -10% | 0.20 |
E[R] = (0.30 × 40%) + (0.50 × 15%) + (0.20 × -10%)
E[R] = 12% + 7.5% + (-2%) = 17.5%
So you'd expect, on average, a 17.5% return on that $100K - meaning $17,500 in value created.
Comparing two investments
Now suppose another option is a $100K spend on Cost Reduction in your operations:
| Outcome | Return | Probability |
|---|---|---|
| Full savings realized | 20% | 0.70 |
| Partial savings | 12% | 0.25 |
| Project stalls | 0% | 0.05 |
E[R] = (0.70 × 20%) + (0.25 × 12%) + (0.05 × 0%)
E[R] = 14% + 3% + 0% = 17%
The Expected Returns are close: 17.5% vs 17%. But look at the distributions. The feature investment ranges from -10% to +40%. The cost reduction ranges from 0% to 20%. Same Expected Return neighborhood, wildly different spread.
Expected Return alone says "these are roughly the same." Your risk appetite and your Budget's tolerance for a miss should tell you they are not.
Where the probabilities come from
In software, you rarely have a neat probability table. You build one from:
The probabilities don't need to be precise. They need to be honest. A common failure mode is assigning 80% to the outcome you want and 5% to the downside because it feels uncomfortable. That's not Expected Return - that's wishful budgeting.
Use Expected Return when:
Don't stop at Expected Return when:
You run a $2M annual P&L. You have $80K in discretionary Budget left for Q3. Two proposals:
Option A - Pipeline automation tool: Reduces Time-to-Fill on sales Pipeline by 20%. Your team estimates:
Option B - New Upsell campaign: Target existing customers for Expansion Revenue.
Calculate Expected Return for Option A: E[R_A] = (0.60 × 30%) + (0.30 × 15%) + (0.10 × 0%) = 18% + 4.5% + 0% = 22.5%
Calculate Expected Return for Option B: E[R_B] = (0.40 × 50%) + (0.35 × 15%) + (0.25 × 0%) = 20% + 5.25% + 0% = 25.25%
Convert to expected dollar value. Option A: $80K × 22.5% = $18,000. Option B: $80K × 25.25% = $20,200.
Note the spread. Option A's outcomes range from 0% to 30%. Option B ranges from 0% to 50%. Option B has a higher Expected Return AND higher Variance.
Decision: If your Q3 plan depends on that $80K generating at least $15K to hit your P&L target, Option A delivers $12K+ in 90% of scenarios. Option B only clears $12K+ in 75% of scenarios. The 2.75 percentage point Expected Return advantage of Option B might not be worth the extra 15 percentage points of downside exposure.
Insight: Expected Return ranks Option B higher. But when you look at the probability of meeting your minimum threshold, Option A is more reliable. Expected Return starts the conversation - it doesn't finish it.
Your team needs a data integration layer. Two paths, each requiring $50K in Implementation Cost:
Build in-house:
Buy a vendor tool ($50K annual license):
E[R_build] = (0.50 × 60%) + (0.30 × 20%) + (0.20 × -15%) = 30% + 6% + (-3%) = 33%
E[R_buy] = (0.70 × 25%) + (0.20 × 10%) + (0.10 × -20%) = 17.5% + 2% + (-2%) = 17.5%
Expected Return strongly favors building: 33% vs 17.5%. But the build option has a 20% chance of a negative return, while the buy option has only a 10% chance.
If your organization has the engineering capacity and a longer Investment Horizon, the build path's higher Expected Return makes sense. If you're in a Turnaround situation where you can't afford a miss this quarter, the buy path's tighter distribution might be worth the lower Expected Return.
Insight: Expected Return differences of nearly 2x can still be a close call when you factor in your operational context. A solo number never captures the full decision - it captures the center of gravity.
Expected Return = Expected Value applied to Returns. It's the probability-weighted average of all possible return outcomes and gives you a single number to compare unlike Capital Investments.
Two investments with the same Expected Return are NOT equivalent if their return distributions differ - one might range from -20% to +60% while the other sits between +5% and +25%. The Expected Return hides that difference entirely.
Your probability estimates matter more than the formula. Garbage probabilities produce garbage Expected Returns. Be honest about downside scenarios - especially the ones that feel uncomfortable to write down.
Treating Expected Return as a guaranteed outcome. A 20% Expected Return does not mean you will get 20%. It means that across many repetitions you'd average 20%. You only get to run this project once, and the actual return will be one specific draw from the distribution. Operators who Budget as if Expected Return is certain will blow their operating plan when variance hits.
Ignoring asymmetry in the outcomes. If Project A has a 50% chance of +40% and a 50% chance of -20%, the Expected Return is +10%. But losing 20% of a $200K investment ($40K loss) might cause real operational damage - a missed payroll cycle, a blown Budget - while gaining 40% ($80K) is nice but not transformational. Expected Return weighs both sides equally. Your P&L does not.
You have $150K to allocate. Three options:
Calculate the Expected Return for each. Then answer: which do you pick if your P&L absolutely cannot absorb a negative return this quarter?
Hint: Calculate E[R] for each using the formula. Then look at which options have any probability of a negative outcome.
E[R_X] = (0.70 × 25%) + (0.30 × 5%) = 17.5% + 1.5% = 19.0%
E[R_Y] = (0.50 × 35%) + (0.30 × 10%) + (0.20 × -10%) = 17.5% + 3% + (-2%) = 18.5%
E[R_Z] = (0.90 × 18%) + (0.10 × 0%) = 16.2% + 0% = 16.2%
By Expected Return alone, Option X wins at 19.0%. But if your P&L cannot absorb a negative return, Option Y is eliminated (20% chance of -10%). Between X and Z, both have zero probability of going negative. X has the higher Expected Return (19.0% vs 16.2%) with a tighter worst case (5% vs 0%), so Option X is the clear pick under this constraint.
Your team proposes two versions of a feature launch. The fast version costs $60K with a 40% chance of 50% return and a 60% chance of 10% return. The careful version costs $60K with a 80% chance of 25% return and a 20% chance of 15% return. Calculate Expected Return for each. Then calculate the expected dollar profit for each. Which version would you fund, and why?
Hint: Expected Return is a percentage. Multiply by the investment amount to get expected dollar value. Think about what the spread of outcomes means for your operating plan.
Fast version: E[R] = (0.40 × 50%) + (0.60 × 10%) = 20% + 6% = 26%. Expected dollar value: $60K × 26% = $15,600.
Careful version: E[R] = (0.80 × 25%) + (0.20 × 15%) = 20% + 3% = 23%. Expected dollar value: $60K × 23% = $13,800.
The fast version has a higher Expected Return (26% vs 23%) and higher expected dollar value ($15,600 vs $13,800). However, the fast version ranges from 10% to 50% while the careful version ranges from 15% to 25%.
If you need predictability (say, you're committing to a revenue number), the careful version delivers at least $9,000 (15% of $60K) with certainty. The fast version might deliver only $6,000 (10% of $60K) in the majority of scenarios. The $1,800 Expected Return advantage of the fast version may not justify the wider spread, depending on your risk appetite and what other commitments ride on this quarter's numbers.
Expected Return is the direct child of two concepts you've already learned. Expected Value gave you the machinery - probability-weighted averaging - and Returns gave you the unit of measurement - economic value relative to invested capital. Expected Return simply applies the first to the second.
But you've probably noticed the gap this lesson keeps pointing at: Expected Return tells you the center of the distribution but nothing about its shape. That gap leads directly to Variance and Standard Deviation, which measure how spread out the possible outcomes are. From there, you'll encounter Risk-Adjusted Return and the Sharpe Ratio, which penalize Expected Return by the amount of Variance you had to accept. And when you're comparing an entire set of Capital Investments - not just two - Portfolio Construction and the Efficient Frontier show you how to combine investments so that the Variance of the whole collection is less than the Variance of the parts. Expected Return is the starting number for all of that machinery. Get comfortable with it, but never mistake it for the full picture.
Disclaimer: This content is for educational and informational purposes only and does not constitute financial, investment, tax, or legal advice. It is not a recommendation to buy, sell, or hold any security or financial product. You should consult a qualified financial advisor, tax professional, or attorney before making financial decisions. Past performance is not indicative of future results. The author is not a registered investment advisor, broker-dealer, or financial planner.