Conversion makes every other lever pay
This is lever 3 of 5 in How To Optimise: Traffic, Content, Conversion rate, Average order value and Retention. Last time we covered The Content Lever. If you're new here, start with Why Optimise?
You've built a brand that sells. You have customers, reviews and years of data. The next stage of growth comes from optimising what you already do, with smarter use of data, technology and the human touch.
You and your team have already paid for every visitor on your site. Conversion decides how many of them you get a return on.
It doesn't take much. Take a £1m store with 1,000,000 visitors a year, converting at 2% with a £50 average order. That's 20,000 orders. Move conversion to 2.25% and you get 22,500. That's 2,500 extra orders, worth £125k a year.
It's one extra buyer in every 400 visitors. No discounting. No extra ad spend.
And every gain here multiplies the others. Better conversion makes every pound of traffic, every email and every repeat customer work harder.
Signs conversion is holding you back
Plenty of people add to basket, but far fewer reach checkout.
Mobile converts well below desktop.
Some traffic sources or landing pages convert at a fraction of the rest.
Customer service answers the same pre-purchase questions every day.
Returns often come back as "not what I expected".
Sales only spike when you run a discount.
Your ads send everyone to the homepage or a generic collection.
Pages get changed on instinct, and nobody can say which changes worked.
How to find the leak
Conversion is about learning how people buy. The numbers show you where they drop out. Your customers tell you why.
Start by splitting conversion rate by device, traffic source and landing page. One sitewide number hides most of the story. A mobile shopper from paid social landing on a product page behaves very differently from a desktop shopper clicking through from your email.
Where do visitors drop out? Shopify's conversion funnel: sessions, add to basket, checkout, purchase. Split it by device and source.
Is the drop-out data accurate? Littledata, so GA4 and your ad platforms see every real purchase.
Which pages get traffic but don't sell? Shopify or GA4 landing page reports, sorted by visits, with conversion rate alongside.
What are shoppers doing on the page? Session recordings and heatmaps of your key product pages. Watch where they scroll, hesitate and leave.
What do customers need to know before they buy? Support emails, live chat, and reviews in Junip.
Why did buyers choose you? Fairing's post-purchase survey answers.
How to optimise it
The heart of this lever is product page A/B testing done properly. Plenty of brands run a test, see a wobble in the numbers and call it. Here's how to do it so the results hold up.
Find the biggest drop-off. Pick one step in the journey, on one device. If the data is clear, narrow it to one traffic source too. That's your starting point.
Gather the evidence. Watch session recordings of that step. Read the reviews, survey answers and support questions. Note the exact words customers use and the doubts they raise.
Write the hypothesis down. For example: "Mobile shoppers from Instagram don't trust our sizing. If we answer the sizing question near the top of the page, in customers' own words, more of them will add to basket." One problem, one change, one number you expect to move.
Check you have enough traffic. Before you start, work out how many visitors each version needs to call a winner. A page with a few hundred visits a week may need many weeks. If the numbers are too small, test on a busier page or test a bigger change.
Answer objections in your customers' words. Your strongest test ideas usually come from the questions people keep asking. Delivery costs. Sizing. How long it lasts. What it feels like. Put the answer where the doubt appears, written the way customers say it.
Run the test and leave it alone. Let it run its full course. Then check the result by device and source as well as overall. A winner on desktop can lose on mobile.
Keep a running test log. Record the hypothesis, the evidence behind it, the dates, the traffic, the result and what you learned. A losing test still teaches you how your customers buy, and the log stops your team retesting old ideas.
Ship the winner and move to the next drop-off. Small, steady gains, one after another.
Even the "obvious" fixes you read about, like a shorter mobile checkout or a sticky add-to-basket button, are worth testing. What lifts one brand can do nothing for yours.
The tools from the toolbox
The full set lives in The Optimisation Toolbox. For this lever, three tools do the heavy lifting.
Shoplift: A/B and split tests on product pages, prices, landing pages and whole themes, launched from your theme customiser with no developers.
Jurni: landing pages and funnels that publish into Shopify, so each campaign lands on a page made for that offer.
Junip: more reviews and customer photos, shown in fast, on-brand widgets right where shoppers decide.
Behind them sits Fairing, which tells you why people bought, in their own words. That's the raw material for every page you test.
The tools do the measuring. Deciding what to test, and why, is the human part. That's the job for you and your team.
Do it yourself this week
Pull conversion rate for your five most-visited product pages, split by device and traffic source.
Copy the text from those five pages.
Gather 50 recent reviews, survey answers and pre-purchase support questions. Watch ten session recordings on the weakest page.
Paste the page copy and the customer comments into Claude or ChatGPT with the prompt below.
Pick the page with the most traffic. Write one hypothesis from what you find and add it to your test log.
Test the new version against the old one. Give it enough visitors to call a winner, record the result, then move to the next page.
The prompt:
"Here is the copy from my five most-visited product pages, and 50 things customers have said or asked. List the questions customers need answered before they buy. Show me which pages fail to answer them. Then rewrite the top section of each page to answer those questions, using the customers' own words. Finally, suggest one testable hypothesis for each page."
Read next
A store that converts well is ready for the next gain: making each order worth a little more.
The full process: The Optimisation Toolbox
The mindset behind it: Why Optimise?
Next lever: The Order Value Lever
Want a hand putting this into practice? Work with Ian.
Not sure which lever to start with? Take the free Optimisation Audit to find your biggest marginal gain.