P10/P50/P90 Explained: Confidence Intervals for Amazon Sellers

You receive a product analysis report. The financials section shows P10 = $1,200, P50 = $8,400, P90 = $22,000. You stare at three numbers instead of one, and you are not sure which one to believe. This article explains exactly what each number means, why all three matter, and how to use them to make sharper product decisions.

What Are Percentiles?

A percentile tells you what percentage of outcomes fall below a given value. If you run 10,000 simulated scenarios for a product and sort the results from worst to best:

• P10 = the value that 10% of outcomes fall below. 90% of scenarios produced a better result than this. Think of it as the "bad but realistic" scenario.
• P50 = the value where exactly half of outcomes are below and half are above. This is the median -- your most likely outcome.
• P90 = the value that 90% of outcomes fall below. Only 10% of scenarios beat this. Think of it as the "good but realistic" scenario.

Together, P10 and P90 form an 80% confidence interval. There is an 80% chance your actual outcome will land between these two numbers. The remaining 20% is split between the extreme tails: 10% chance of doing worse than P10, 10% chance of doing better than P90.

Why P50 Is Not the Average

This is the most misunderstood concept in probabilistic analysis. The median (P50) and the mean (average) are the same only when the distribution is perfectly symmetrical. Amazon FBA profit distributions are almost never symmetrical.

Here is why. Your downside is bounded: the worst that can happen is you lose your entire investment. But your upside is theoretically unbounded: if the product goes viral, sales can spike to multiples of your baseline estimate. This creates a right-skewed distribution where a few outstanding outcomes pull the mean above the median.

Practical example:

If you plan your finances around the mean ($9,600), you will be disappointed more than half the time, because the median -- the outcome you are most likely to experience -- is $8,400. The mean is inflated by the small number of scenarios where everything goes right and you earn $25,000+.

Rule of thumb: Plan your expenses around P50. Budget your reserves around P10. Celebrate if you hit P90.

Reading the Spread: What P10-to-P90 Width Tells You

The distance between P10 and P90 is the inter-decile range, and it measures how uncertain the outcome is. A narrow spread means the product is predictable. A wide spread means it is a gamble.

Product A has a spread of $19,400 -- the outcome ranges from losing money to making $18,500. This is a volatile opportunity. Product C has a spread of $6,400, with even the worst case being solidly profitable. Product C is a "safer bet" even though its P90 is lower than Product A's.

Which is better depends on your risk tolerance and portfolio strategy. A verkoper with $50,000 in capital and five products can afford one volatile Product A in the mix. A verkoper with $10,000 going all-in on one product should strongly prefer Product C.

The Decision Framework

Here is how to use P10/P50/P90 for practical product decisions:

Check 1: Can you survive the P10?

The P10 outcome is not your worst case -- it is the boundary of the bottom 10%. Ask yourself: if this scenario plays out, can I absorb the loss without serious financial consequences? If P10 = -$5,000 and that would wipe out your reserves, the product is too risky regardless of how attractive the P50 and P90 look.

Check 2: Is the P50 worth your time?

The median outcome is what you should expect to earn. If P50 = $3,000 for a product that requires $12,000 in capital and six months of work, the return-on-effort may not justify the opportunity cost. Compare the P50 to what you could earn with the same capital deployed elsewhere -- including keeping your job.

Check 3: Does the P90 compensate for the risk?

The P90 represents your upside scenario. If P10 is negative (you might lose money), the P90 should be substantial enough to compensate for that risk. A product with P10 = -$2,000 and P90 = $3,000 offers terrible risk-reward. A product with P10 = -$2,000 and P90 = $25,000 might be worth the gamble.

The Asymmetry Ratio

Berekenen the ratio of upside to downside:

Asymmetry = (P90 - P50) / (P50 - P10)

For Product A: Asymmetry = ($18,500 - $4,200) / ($4,200 - (-$900)) = $14,300 / $5,100 = 2.8. This is strongly positively skewed -- the upside is nearly 3x the downside. That is an attractive asymmetry despite the risk of loss.

Common Mistakes When Using Percentiles

Anchoring on P90. Optimism bias causes verkopers to treat the P90 as "what will probably happen." It will not. Only 10% of scenarios reach P90 or above. If you base your financial plans on P90, you will be disappointed 90% of the time.

Ignoring P10 because "it probably won't happen." A 10% probability is not negligible. If you launch 10 products in your Amazon career, on average one of them will perform at or below P10. That is not a black swan -- it is a statistical certainty over a portfolio.

Comparing products on P50 alone. Two products might both have P50 = $8,000, but one has P10 = $4,000 while the other has P10 = -$6,000. They are fundamentally different risk profiles. Always compare the full triplet.

Confusing percentiles with time periods. P10/P50/P90 describe probability, not timelines. P50 does not mean "what you earn in your 50th month." It means the outcome that half of all scenarios exceed.

Percentiles in Portfolio Context

Experienced Amazon verkopers do not evaluate products in isolation. They build portfolios -- typically 3-8 products across different categories. In a portfolio context, percentiles work differently:

If you launch 5 products, each with P50 = $6,000 and independent outcomes, your portfolio P50 is approximately 5 x $6,000 = $30,000. But the portfolio P10 is much better than 5 x individual P10, because the chances of all five products simultaneously hitting their worst-case scenario are extremely low. Diversification narrows the confidence interval around your total earnings.

This is why understanding Monte Carlo methodology matters not just for single products but for building a resilient Amazon business. The probability of profitability for a well-diversified portfolio is substantially higher than for any individual product within it.

Practical Percentile Benchmarks

Based on thousands of product analyses, here are general guidelines for what the P10/P50/P90 triplet signals about a product opportunity:

These benchmarks assume a $10,000-$15,000 initial investment. Scale accordingly for larger or smaller launches.

From Numbers to Action

P10, P50, and P90 are not abstractions. They are the three numbers that replace the dangerous illusion of a single "expected profit" figure. When someone asks "how much will this product make?", the honest answer is always a range. Percentiles give that range structure, discipline, and actionability.

Use P10 to stress-test your finances. Use P50 to set realistic expectations. Use P90 to evaluate whether the upside justifies the risk. And use the tornado chart from your sensitivity analysis to identify which variables you can influence to shift the entire distribution in your favor.