Your Conversion Rate Is Lying to You
“What’s a good conversion rate?”
Every e-commerce founder asks it at some point, and it has no useful answer. A £20 impulse product selling off bottom-of-funnel Google search will happily convert at 6%. A £200 considered purchase, sold cold to strangers on Meta, might convert at under 1% and still out-earn the first brand several times over on contribution. Is either website better than the other? You can’t tell from those numbers.
That’s the problem with conversion rate as a metric. It isn’t a lever you can pull directly; it’s an output of dozens of inputs, most of which aren’t on the page at all. Category, price point, traffic mix, how much your creative educates before the click, the share of returning customers in your traffic. Comparing your conversion rate to another brand’s tells you almost nothing, and optimising towards it in isolation can actively lead you astray.
Most CRO advice skips this bit and goes straight to button colours. So let’s do it properly: measure the right number first, then work back to the page.
Measure the right number
Here’s how gameable conversion rate is. If a client asked me to double their conversion rate by Friday, I could do it: turn off Meta. Cold traffic drops to zero, you’re left with warm traffic that converts at multiples of the blended rate, and the dashboard looks fantastic. You’d also have quietly torched the growth engine of the business.
The same trick works on nearly every e-commerce metric. Want sessions up? Run traffic campaigns. Want AOV up? Push expensive products, never mind the CAC. Any single metric, optimised in isolation, can be moved in the “right” direction while profitability goes the wrong way.
The better number is revenue per visitor:
RPV = conversion rate × average order value
Better still, multiply by gross margin and you get contribution per visitor — an actual money number, tied to actual profit. As a former accountant I’d push you towards the contribution version, because the closer your KPIs sit to financial reality, the better your decisions get. A conversion rate can drift up while the business gets worse. Contribution per visitor can’t.
This also explains why big brands tend to have lower conversion rates. As you scale cold acquisition, cold traffic (converting at, say, 1%) becomes a bigger share of the blend against warm traffic (converting at 4%), and the blended number sinks. Simple arithmetic, and no redesign will reverse it.
To be clear, none of this makes conversion rate useless. Measured like for like — same brand, same traffic mix, same channel — it’s a perfectly good way to judge whether a change to the site worked, and it’s exactly how we report before-and-afters on our own client work. The trap is treating it as the business’s north star, or comparing your number to somebody else’s. Diagnose with conversion rate; steer by contribution.
The order of operations
None of this means you shouldn’t work on your pages. It means working on them in the right order. Ranked by impact, confidence, and ease combined:
Clarity and hierarchy — can a stranger instantly understand what this is, above the fold?
Copy — does it remove doubt and speak to the actual desire?
Offer — is the thing on offer worth converting on at all?
Proof — do they believe your claims?
Friction — is anything physically in the way of the purchase?
One caveat worth being honest about: the biggest lever here is the offer, not clarity. A bad offer keeps you at zero no matter how beautiful the page; a great offer converts despite an ugly one. But good offer curation is genuinely hard, whereas fixing clarity and hierarchy is quick, cheap, and needs no split testing. Hence the order above. Start where the effort-to-impact ratio is best, then earn your way down the list.
Usefully, this list is also roughly the structure of a good landing page, top to bottom: clarity first, persuasion next, offer, proof, then remove every remaining obstacle.
Four questions above the fold
Pull up your product page on your phone (not your laptop — the overwhelming majority of D2C traffic is mobile). Above the fold, a stranger should be able to answer:
What is it?
Why this one? What’s the mechanism that makes it different from the ten alternatives they’ve seen this week?
Why now? Why not bounce and think about it, or wait for your sale?
Why should I trust you?
The default Shopify PDP — product name, price, add to cart — fails all four. “Face Serum 30ml” answers nothing. “The only overnight retinol buffered to work while you sleep, so you wake up firmer, not red” answers the first two in one line. Add a reason to act now (small-batch stock, instalment payments) and visible reviews, and you’ve covered all four before anyone scrolls.
The other half of this is continuity. If your ad sells collagen for joint pain and the click lands on a generic page listing nine benefits, of which “joint support” is number seven, you’ve broken the story. The customer watched an ad about their specific problem and landed on a page claiming to solve everything, which reads as solving nothing. One persona, one angle, one page. It’s more work than sending everything to the same PDP, and it’s worth it.
Copy: write about the person, not the product
The most common copywriting mistake on landing pages is writing about what the product is rather than why it matters. Every line should survive a simple “so what?” test. “Triple peptide complex” — so what? “Firmer skin in four weeks, using a triple peptide complex” — better. Outcome first, feature as justification.
The harder truth: your product probably isn’t unique. Most aren’t. The job of positioning is to make a commoditised product genuinely feel one-of-one, and there are honest ways to do it. Mine your reviews for language no competitor can borrow (“the first retinol that didn’t make me peel” is worth more than any headline you’ll write in a brainstorm). Or find the audience nobody is speaking to: Bloom launched a greens powder into the most saturated corner of supplements and won, largely by positioning it for women when every competitor was shouting at everyone.
What doesn’t make you unique: free shipping, great service, fast delivery. Those are table stakes. Nobody has ever converted because of a “fast dispatch” badge.
Offer: positioning beats discounting
When people hear “improve the offer”, they reach for 20% off. That’s the lazy version, and it trains customers to wait for sales while eroding margin.
The better version restructures value. A three-product routine bundle at £95 instead of £105 apart looks like a £10 discount, but if shipping the items together saves you that £10 in duplicated fulfilment cost, margin is untouched and AOV has roughly tripled. Add a gift with purchase that costs you £4 and is credibly worth £25, a free-shipping threshold set deliberately above your current AOV, and a one-click post-purchase upsell, and you’ve raised revenue per visitor several ways without a single blanket discount.
Rule of thumb on shipping: keep total extra costs at checkout under about 10% of order value and they stop being a conversion problem.
One filter on all of the above: run every offer tactic through your brand positioning before it goes near the page. A cart-slider upsell and a free-gift threshold suit a supplement brand; they’d look desperate wrapped around a £400 cashmere jumper. More on this below.
Proof that matches the claim
If you have no reviews, fix that before anything else in this post. Going from zero to somewhere around 80 reviews is the single biggest trust jump a young brand gets; beyond a couple of hundred, the marginal review does very little. Sort image and video reviews first, and incentivise customers towards leaving them (a credit on the next order, not payment for the review — one is marketing, the other is illegal in most markets).
Then make the proof specific. Five reviews that directly back the claim your ad made are worth more than four hundred saying “great product, fast delivery”. If the promise is firmer skin in four weeks, the page needs before-and-afters about firmness, not a wall of generic five stars. And trust badges — the little padlock seals in the footer? A relic of the dropshipping era. They close nobody.
Friction: where carts go to die
Around 70% of carts are abandoned. Some of those people were never buying. The rest left for reasons Baymard Institute has documented exhaustively: surprise extra costs at checkout (the biggest, at 39%), slow delivery, forced account creation, checkouts that ask for too much.
The fixes are unglamorous: guest checkout by default, express payment options (Shop Pay, Apple Pay) at the top, a lean form, full cost visible up front, returns policy stated plainly rather than buried in a footer link. No single one of these is exciting. Together they’re worth more than most redesigns.
A word on A/B tests
Finally, the maths nobody selling CRO services wants to run you through.
To trust a 10–20% conversion lift, you need roughly 1,000 conversions per variant. At a 2.5% conversion rate, that’s about 80,000 sessions through a single test. Most seven-figure brands can’t push that through one product page in a month. So the winning test your agency shows you after two weeks and a hundred conversions is, in most cases, just noise with a green tick on it.
It gets worse: most tests are checked daily and stopped the moment they look significant, which inflates the false-positive rate further. Zoom out over two years of this and revenue per visitor is often exactly where it started.
Split testing has its place — at eight or nine figures, with proper test design and enough volume, a real programme is high leverage. Below that, skip the testing theatre. Make the obvious, high-confidence changes above, and judge success on revenue per visitor over months, not on a dashboard that tells you what you want to hear.
For brands who value brand
Everything above assumes the job of the page is to convert as hard as possible. For premium brands, that assumption needs handling with care.
Part of what a premium customer pays for is the feeling around the product, and that feeling is built (or eroded) by every touchpoint. You won’t find countdown timers or “only 3 left” badges on Aesop or Rapha, and it isn’t because nobody there has read a CRO blog. Restraint is the selling effort. A quieter, more confident page might convert a touch worse this quarter and still win comfortably over time, because it protects the pricing power and the repeat behaviour that premium positioning exists to create.
The fundamentals in this post cost a premium brand nothing: clarity above the fold, copy about the customer, proof that matches the claim, a checkout that doesn’t get in the way. It’s the pressure tactics — urgency banners, aggressive pop-ups, discount-anchored bundles — that need to pass through your brand strategy before they pass into your theme. Applied in a vacuum, they’ll lift a dashboard while quietly cheapening the thing that lets you charge what you charge.
The recap
Conversion rate is an output, not a north star. Judge the business on revenue (or contribution) per visitor; save conversion rate for like-for-like before-and-afters.
Expect blended conversion rate to fall as you scale cold traffic. The mix is doing that, and it’s fine.
Work the levers in order: clarity, copy, offer, proof, friction.
Answer the four questions above the fold, and keep the ad’s promise on the page.
Reposition before you discount.
Proof only works when it matches the specific claim being made.
Most brands don’t have the traffic to split test honestly. Judge the trend, not the test.
If you’re premium, filter every tactic through your positioning. Restraint on the page is sometimes the highest-earning choice over a customer’s lifetime.
If you’d like a second pair of eyes on where your own funnel leaks, get in touch.
