Why Conversion Rate Is a Vanity Metric
Conversion rate is the most-cited metric in ecommerce. It shows up in every dashboard, every agency report, every A/B test readout. And it is, almost universally, the wrong thing to optimize for.
The chart above tells the story clearly. Stores with a $35 AOV routinely post 7% conversion rates. Stores with a $500 AOV often sit at 0.3–0.5%. If you looked at conversion rate alone, you'd conclude the $35 AOV store is crushing it. But run the math — and that "crushing it" store is generating $2.45 per session, while the $500 AOV store at 0.5% is generating $2.50 per session. Nearly identical revenue efficiency, wildly different conversion rates.
Conversion rate is inversely correlated with Average Order Value. Higher-priced products naturally convert at lower rates. Comparing CVR across stores — or even across product categories within a store — is comparing apples to aircraft carriers.
This is not a new observation in isolation. What is new is the scale of the data. Across 1,200+ real ecommerce stores, the pattern is unambiguous: CVR and AOV have a strong negative correlation. The higher your average order value, the lower your conversion rate will be — not because your store is broken, but because that is simply how consumer purchasing behavior works.
"A store selling $600 furniture will never convert at 5%. That's not a CRO problem. That's physics."
The danger is not just that CVR is an imperfect metric. The danger is that optimizing for CVR actively destroys revenue. Teams that chase conversion rate end up discounting products (raises CVR, tanks AOV), adding low-quality traffic sources (raises sessions, raises CVR numerically, destroys revenue quality), or cluttering the checkout with desperation tactics that cheapen the brand and erode trust.
The Data: What 1,200+ Stores Actually Show
The scatter plot above represents the raw output of a proprietary study spanning over 1,200 ecommerce stores across verticals — apparel, home goods, electronics, beauty, sporting goods, and more. Each data point is a single store. The X-axis is Average Order Value. The Y-axis is either Conversion Rate or Revenue Per Session, depending on which toggle you've selected.
Finding #1: CVR Tells You Almost Nothing About Store Health
The CVR chart looks like a shotgun blast. There is a visible downward trend — higher AOV correlates with lower CVR — but the variance is enormous. Two stores with identical $150 AOVs can have CVRs of 0.9% and 2.2% respectively. Does the 2.2% store have a better business? Not necessarily. It might be running heavy discounts, targeting bargain-hunter traffic, or selling lower-margin products that inflate unit volume.
CVR without context is noise. It tells you how many people clicked "buy" relative to how many visited. It says nothing about how much money each visitor was worth to you.
Finding #2: RPS Reveals the True Competitive Landscape
Toggle to the RPS view and the picture changes dramatically. Revenue Per Session — calculated simply as AOV × CVR — compresses the noise. Stores cluster more tightly. Outliers become genuinely interesting rather than artifacts of AOV differences.
In the RPS view, you can immediately identify which stores are genuinely outperforming their AOV tier. A $510 AOV store with a 1.1% CVR generates $5.61 per session — the highest in this dataset. A $580 AOV store at 0.9% generates $5.22 per session. These are the stores worth studying. These are the benchmarks that actually matter.
Contrast that with a $35 AOV store at 7% CVR — which looks spectacular on a CVR dashboard — but only generates $2.45 per session. It is not a bad business. But it is not the benchmark for optimization that its CVR would suggest.
Finding #3: The $200–$600 AOV Range Is Where RPS Gets Interesting
One of the most striking findings in this dataset is the RPS performance of mid-to-high AOV stores. Stores in the $200–$600 AOV range show the widest RPS variance — from under $1.50 per session to over $5.50. This is where store quality, traffic quality, and conversion optimization actually differentiate businesses.
In the low-AOV range ($0–$100), RPS is relatively compressed. Most stores cluster between $1.50 and $3.00 per session regardless of CVR. The math simply doesn't allow for dramatic outliers when AOV is low. But at $400+ AOV, a difference of 0.5 percentage points in CVR translates to $2+ per session in RPS — a massive gap in revenue efficiency.
| AOV Tier | Typical CVR Range | Typical RPS Range | What CVR Tells You | What RPS Tells You |
|---|---|---|---|---|
| $0–$100 | 1% – 7% | $0.55 – $3.06 | Almost nothing | Actual revenue efficiency |
| $100–$200 | 0.85% – 2.2% | $1.00 – $2.52 | Misleading without AOV context | Comparable across stores |
| $200–$400 | 0.3% – 2.3% | $1.14 – $4.60 | Highly misleading | Reveals true outliers |
| $400–$600+ | 0.1% – 1.1% | $0.60 – $5.61 | Completely useless alone | Most differentiated signal |
How to Actually Use Revenue Per Session
RPS is not just a diagnostic metric — it is an optimization target. When you shift your team's focus from CVR to RPS, the entire conversation about what to test, what to fix, and what to prioritize changes.
Use RPS to Benchmark Against Competitors
If you know your AOV and your CVR, you know your RPS. And if you know your category's typical AOV range, you can benchmark your RPS against the distribution in this dataset. A $200 AOV store generating $1.50 per session is underperforming — the data shows stores in that AOV range can reach $4.60 per session. That gap is your opportunity.
Use RPS to Evaluate Traffic Sources
Different traffic sources have different RPS profiles. Paid social might drive high volume with low RPS. Organic search might drive lower volume with higher RPS. CVR alone will mislead you — a traffic source that converts at 4% but has a $40 AOV is worth less than one that converts at 1% with a $250 AOV. RPS normalizes this instantly.
Use RPS to Evaluate A/B Tests
This is where the CVR obsession causes the most damage. Teams run A/B tests and declare winners based on conversion rate lift. But if the variant that "won" on CVR did so by attracting lower-intent buyers who spent less — the test was a net negative for revenue. Always evaluate A/B test results on RPS, not CVR.
Stop asking "what is our conversion rate?" Start asking "what is our revenue per session, and how does it compare to stores with similar AOVs?" That is the question that leads to real growth decisions.
Use RPS to Set Realistic Optimization Targets
One of the most valuable outputs of this dataset is the ability to set realistic RPS targets by AOV tier. If your store has a $150 AOV and you're generating $1.30 per session, the data shows that $2.00–$2.50 per session is achievable for stores in your tier. That is a concrete, data-backed goal — far more useful than "increase CVR by 20%."
The Methodology Behind This Study
This data was collected from a proprietary dataset of over 1,200 ecommerce stores. Stores span multiple verticals and revenue tiers. Data points represent actual session and transaction data, not modeled or synthetic estimates. CVR is calculated as transactions divided by sessions. AOV is calculated as revenue divided by transactions. RPS is derived as AOV × CVR — equivalent to revenue divided by sessions.
The dataset intentionally excludes stores with fewer than 1,000 monthly sessions to ensure statistical reliability of CVR calculations. Outliers at the extreme ends of the AOV distribution (above $1,000 AOV) are excluded from this visualization to maintain readability, though they follow the same pattern.
This is, to our knowledge, one of the largest publicly shared datasets on ecommerce CVR and RPS distributions. The goal is to give operators and growth teams a real benchmark — not the cherry-picked case studies that dominate most CRO content.
The Bottom Line
Conversion rate is not a useless number. It is a useful input. But as a standalone optimization target, it is a trap. It rewards the wrong behaviors, penalizes high-AOV stores unfairly, and gives teams a false sense of progress when they are actually destroying revenue quality.
Revenue Per Session is the metric that cuts through all of that. It is simple to calculate, easy to benchmark, and directly tied to what actually matters: how much money does each visitor to your store generate?
If you can only track one metric, track Revenue Per Session. Everything else is a proxy for it.
The data in this study makes the case better than any argument could. Toggle between CVR and RPS in the chart above and watch the picture transform. That transformation — from noise to signal — is what happens when you stop optimizing for vanity and start optimizing for revenue.