Ecommerce Optimization: What to Fix First

Halftone stack of five layers with the bottom layer in pink, ecommerce optimization

Ecommerce optimization is the work of making an online store earn more from the traffic it already has. Not more visitors: more revenue from each one. It covers five layers, and the order you tackle them in matters more than any single tactic.

Most guides list twenty tactics and stop. The question store owners ask me is simpler: where do I start? So this is an order of operations, with the one layer I'm not the expert on marked plainly.

In short: measure revenue per visitor first. Fix the bottom of the funnel before the top. Conversion, order value and repeat purchase make the money; speed and findability let them work.

What ecommerce optimization actually covers

Five layers, in two groups.

The three that make money:

Conversion rate: how many visitors buy.
Average order value: how much each order is worth.
Repeat purchase: whether they come back.

The two that let those work:

Speed: a slow page drags down every layer above it.
Findability: whether the right people reach the store at all.

The number that ties the first two together is revenue per visitor: conversion rate multiplied by average order value. It's the one I watch, because it catches the change that lifts one metric while sinking the other.

The order that pays back fastest

Work from the bottom of the funnel up. A visitor you stop losing at checkout is worth more than a new visitor at the top, because they had already decided to buy.

Measure a clean baseline. Take revenue per visitor over a normal 30-day window before you touch anything. Without it you can't tell a lift from a good month.
Fix the checkout and the cart. This is where people who chose to buy still leave.
Fix the product and collection pages. This is where the buying decision happens.
Raise order value. Once people buy reliably, make each order worth more.
Fix speed where it's hurting. Usually on mobile, and often caused by overlapping apps.
Build repeat purchase. Start with capture and the first follow-up flows.
Improve findability. Do it once the store converts the traffic it already gets.

The homepage is missing from that list on purpose. In the stores I audit, it's rarely where the money leaks.

Layer by layer

Checkout and cart: where decided buyers still leave

Baymard puts average cart abandonment at 70.22%, across 50 studies. Some of that is window shopping. A lot of it is friction the store added: surprise costs, forced account creation, a checkout button pushed below the fold.

These fixes are small, specific and close to the money, which is why they come first. I cover them in the checkout conversion guide.

Product and collection pages: say the benefit, not just the spec

Buyers decide on the collection grid as much as on the product page. On the grid, three benefit bullets per product card lifted conversion 10 to 20% across the stores I've tested it on. The full story is in ecommerce conversion rate optimization. On the product page itself, lead with what the product does for the buyer and put proof next to the decision, as in product page optimization.

Average order value: the layer most stores leave empty

Once people buy reliably, make each order worth more. A free-shipping threshold with a progress bar set above your current order value typically lifts it 12 to 18% over time. A well-chosen cross-sell after checkout adds to an order without risking it.

Work out your starting point with the AOV calculator, then use how to increase average order value for the playbook.

Speed only matters where buyers feel it

Speed is rarely the first problem, but it multiplies every other one. It usually shows up as a store that feels fine on office wifi and slow on a phone on mobile data. Most of your buyers see it that way, which is why mobile conversion rate is the number to watch.

The first check is simple: open your product page and cart on a real phone, on mobile data, and wait. Then list your apps. On Shopify the most expensive pattern I see is overlap: several apps doing the same job, each loading its own scripts on every page.

It's worth fixing when the slowness sits where buyers decide. It isn't worth chasing a lab score for its own sake. I go through where Shopify speed work goes wrong separately.

Repeat purchase starts at the signup form

The form that captures the address matters more than the email platform behind it. A bare "subscribe" box underperforms a headline plus a reason, and the offer belongs on the button as well as in the copy.

Then build the three flows that earn their keep, in this order: welcome, abandoned cart and one post-purchase flow. Get those live before comparing tools. Often what holds a store back is flows that don't exist, not the platform they would run on.

For consumables, don't push subscriptions at first-time buyers. Sell them later by email, once the customer knows they want more. If you're choosing a platform, Klaviyo vs Omnisend vs Shopify Email covers what actually decides it.

Findability: the layer I'm not the expert on

I'm a conversion specialist, not an SEO agency, and I'd rather say so than bluff. What I can give you are the basics a store owner controls:

Titles in buyers' words. Product and collection titles should use the words people search, not internal names or SKU codes.
Your own product descriptions. Copy pasted from the manufacturer says what every other stockist says. Write what your buyer needs to know.
Collection pages with a short paragraph of real text, placed or collapsed so it never pushes the products down. The grid is the page's job.
Links between collections and products, so both visitors and crawlers can move through the catalogue.
Descriptive image alt text that says what the image shows.
Fast pages, which help conversion whatever they do for rankings.

Beyond that, hire an SEO specialist. But fix conversion first: a store that converts poorly gets less from every visitor SEO brings in.

What doesn't count as optimization

A redesign without a diagnosis. It's the most expensive way to move the least. The store looks better, the numbers stay where they were, and nobody can say why. I've written about why redesigns fail to move results.

Discounting. A discount lifts conversion by giving away margin. That's a pricing decision, not optimization.

Stacking apps. Every app promises a lift in isolation. Together they slow the store and crowd the page.

How to tell whether it's working

Watch revenue per visitor over clean before-and-after windows you set in advance, not day to day.

Expect some results to be real but hard to prove. A lift in add-to-cart clicks can arrive while the store is in a downtrend, and then the revenue never shows it. Ecommerce conversion rate optimization has a real example. That's why the baseline in step one matters.

And read the evidence, don't just count it. A small click share for a "buy once" option isn't the same as small demand for buying once.

Where to start

If you want your ecommerce optimization done in that order, that's what my conversion rate optimization service does. It starts with an audit of where your store actually leaks. For scale: across Baseline™ projects, conversion rate lifts of 20 to 50%+ are typical, with outliers of 100%+ conversion rate improvements.

Not sure which layer is costing you most? Book a free 30-minute call and we'll find it.

Frequently asked questions

Ecommerce optimization is the work of making an online store earn more from the traffic it already has. It covers five layers: conversion rate, average order value, repeat purchase, site speed and findability. The first three make the money. The last two let the first three work.

Measure first, then work from the bottom of the funnel up. Fix the checkout and the cart before the product and collection pages, and those before the homepage. A visitor you stop losing at checkout is worth more than a new visitor at the top, because they had already decided to buy.

No. SEO brings visitors in; ecommerce optimization decides what they're worth once they arrive. They overlap on speed and on clear product and collection pages. A store with weak conversion gets less from every SEO gain, so conversion work usually pays back first.

Usually not. Most stores lose money in specific places: a leaky checkout, a product card that never states the benefit, no second offer after purchase. Fixing those is cheaper and faster than a redesign, and a redesign that skips the diagnosis often looks better without selling more.

Track revenue per visitor, which is conversion rate multiplied by average order value. It catches changes that win one metric while losing the other. Measure a clean window before you change anything, and compare the same window after, rather than reading day-to-day swings.

Darjan HrenCRO Strategist & Shopify Consultant

Co-founder of Skuology and creator of the Baseline Framework. Darjan has worked across 90+ Shopify stores, with a focus on conversion rate optimisation and AOV growth.

About Darjan Book a call

Want this applied to your store?

See how to fix your store.

View services Book a call

Ecommerce Optimization: What to Fix First, and in What Order