1% CR Lift at €1M vs €5M vs €15M Revenue Bands
A side-by-side worked example of the annualized profit a single 1-point CR lift produces at €1M, €5M, and €15M store sizes — so you can size CRO investment before you hire.
Quick answer
A 1-point conversion rate lift (e.g. 2.0% → 3.0%) at 35% contribution margin produces roughly €175k incremental annual profit at €1M revenue, €875k at €5M, and €2.6M at €15M. The lift scales linearly with revenue — but the business case for hiring a CRO specialist only clears around the €3M–€5M band.
1% CR Lift at €1M vs €5M vs €15M Revenue Bands
A side-by-side calculation of the annualized profit a single 1-point conversion rate lift produces across three common DTC revenue tiers.
This scenario quantifies the same experiment outcome — a 1 percentage point uplift in site conversion rate — at three store sizes online retailers typically sit in: €1M, €5M, and €15M in annual revenue. The point is not the math itself (it's simple multiplication) but the operational gap between bands. At €1M, a 1pp lift pays for a freelance CRO sprint. At €5M, it funds a full-time specialist plus tooling. At €15M, it funds a small experimentation team and still leaves margin on the table. Use this page to decide what to commit before you run the test, not after.
The framing matters because a 1-point lift sounds identical at every size. It isn't. The dollar value scales with revenue, but the decision to invest in CRO infrastructure — headcount, testing tools, a heatmap stack — has fixed costs that don't.
We hold three variables constant across the three scenarios so the comparison is clean: a starting CR of 2.0% lifting to 3.0%, an AOV of €70, and a 35% contribution margin after COGS, payment fees, shipping, and returns. Swap your own numbers in if your beauty or apparel store runs hotter or thinner.
The shared assumptions
Annual revenue defines the band. From that we back out sessions: at 2.0% baseline CR and €70 AOV, €1M of revenue implies roughly 714k annual sessions, €5M implies 3.57M sessions, and €15M implies 10.7M sessions.
After the 1pp lift, sessions stay flat — that's the whole CRO premise — but transactions rise 50% (from 2.0% to 3.0% is a 50% relative gain). New revenue equals incremental orders × AOV. Profit is that incremental revenue × 35% contribution margin.
Why we use contribution margin, not gross margin
Incremental orders carry incremental costs — pick-pack, payment processing, shipping, and the predictable slice of returns. Gross margin overstates the windfall. Contribution margin (typically 30–40% for Shopify apparel and beauty stores) is the honest number to multiply against.
Working the three bands
At €1M revenue, lifting CR from 2.0% to 3.0% adds roughly 7,140 incremental orders per year. At €70 AOV that's €500k of incremental revenue and, at 35% contribution margin, about €175k of incremental annual profit.
At €5M, the same lift produces 35,700 extra orders, €2.5M of new revenue, and €875k of profit. At €15M: 107k extra orders, €7.5M of revenue, and roughly €2.6M of incremental profit annually.
Read the table below as a cost-of-inaction signal. Every month the €15M store delays a working test, it's leaving about €218k in profit on the floor. The €1M store leaves about €14.5k — meaningful, but not screaming for a six-figure tool stack.
Profit impact by revenue band
Annualized profit from a 1pp CR lift (2.0% → 3.0%), AOV €70, 35% contribution margin
| Revenue band | Annual sessions | Extra orders/yr | Extra revenue/yr | Extra profit/yr | Profit/month |
|---|---|---|---|---|---|
| €1M store | ~714k | ~7,140 | €500k | €175k | €14.6k |
| €5M store | ~3.57M | ~35,700 | €2.5M | €875k | €72.9k |
| €15M store | ~10.7M | ~107,000 | €7.5M | €2.62M | €218.5k |
The profit column scales linearly with revenue because the lift is a percentage — that's expected. What's not obvious is how the headcount math flips. The €1M store can't justify a €70k/year CRO hire on €175k of upside; the €5M store comfortably can; the €15M store should already have one.
What this means for headcount and tooling
At €1M, the answer is usually a focused 6–8 week engagement: a CRO freelancer or agency sprint targeting one specific funnel leak (typically product page → cart, or cart → checkout). One landed 1pp lift pays the engagement back inside a quarter and leaves margin for the next one.
At €5M, an in-house CRO specialist plus a lightweight testing stack pays back in 6–10 weeks on the first compounding win. This is the band where consolidating GA4, a heatmap tool, and a separate A/B test platform into one snippet stops being a nice-to-have — page weight directly costs you conversions you're trying to win back.
Time-to-significance changes the calculus
The €1M store has a hidden problem: with 714k sessions a year, a fair-powered A/B test on a 2% baseline detecting a 10% relative lift takes 6–10 weeks per test. You get 5–8 shots a year. The €15M store gets the same significance in under a week and can run 30+ tests annually.
That's why the €15M profit number understates the gap. Bigger stores don't just earn more per win — they get more wins per year, and the compounding curve pulls away. Historical GA4 import helps smaller stores claw some of this back by surfacing the highest-EV test ideas first, rather than burning two months testing a low-impact CTA color.
Frequently asked questions
Going from 2.0% to 3.0% means 50% more visitors convert (3/2 = 1.5×). The 'percentage point' is the absolute move; the relative gain is what drives incremental orders. Both framings are correct — the relative gain is what matters for the profit math.
It's the middle of the typical range. Beauty SKUs often run 40–50% contribution margin; apparel with returns and free shipping runs closer to 25–35%; electronics resellers can be as thin as 15–20%. Recompute with your own number — the structure of the comparison holds.
Not in the way you'd expect. Revenue is sessions × CR × AOV, so doubling AOV at the same revenue means halving sessions — and your absolute profit lift per 1pp CR gain ends up the same. AOV matters for test-power velocity, not for the size of the win.
Interpolate. At €2.5M expect ~€437k incremental profit from a 1pp lift; at €8M expect ~€1.4M. The headcount inflection sits around €3M revenue — below that, freelance/agency is the right model; above, in-house starts paying.
For a store under €5M with a real funnel leak, 8–14 weeks is realistic — that's typically 2–3 sequential tests where one or two land. Above €10M, the same 1pp gain is harder because the easy wins are usually already taken; expect 4–6 months and a portfolio of smaller compounding tests.
No — the lift is on existing traffic you've already paid for. That's CRO's structural advantage over paid: the cost is the test program, not the click. The €175k–€2.6M profit numbers are after COGS and order-level costs, before fixed CRO headcount and tooling.
Around €1.5M–€2M annual revenue is the floor for a paid testing stack to pay for itself within a year, assuming you actually run tests monthly. Below that, focus on funnel analytics and qualitative session replay first — A/B test infrastructure is wasted if you can't reach significance.
Yes, indirectly. Adding a heatmap + test tool + recording tool typically costs 200–600ms of LCP, which itself depresses CR by 1–3%. At €15M that performance tax can cost €450k–€1.4M annually — often more than the test program earns. Consolidating to one lightweight snippet matters at scale.
It's realistic as a portfolio outcome over 6–12 months, not as a single test. The honest pattern: 70% of tests are flat or losing, 20% are small wins (0.1–0.3pp), 10% are bigger wins (0.5pp+). Summed over a year of disciplined testing, 1pp is a normal annual result.
The parent page walks through the formula and the general logic. This page is the operational view: same math, applied at three specific revenue bands, with the headcount and time-to-significance implications that fall out. Read this one when you're sizing the investment.
Test ideas before you ship them
Run unlimited A/B tests, attach hypotheses to outcomes, and build a searchable archive of what works — and what doesn't.