How not to optimize an online store. Good vs bad store optimization.

This article originally appeared in issue 02/2023 of DIGITAL NOW! magazine.

If you run an e-commerce business, you’re certainly making changes in your store. You change the look of the homepage or the layout of the product page. You want to improve UX, write better copy, build a nicer interface. But are you certain the changes you’re shipping grow sales? What if changing the store’s appearance actually lowers conversion?

---

Language purists will say there’s no such thing as bad optimization. Optimization, by definition, has to be good. But it often happens that we only think we’re optimizing something while sales actually fall. That’s what I call bad optimization.

Bad optimization of a store can be avoided. And I’ll show you how. You’ll learn how to make sure the changes you ship genuinely lift sales. Let’s start with the mistakes, though. Let’s look at the causes of bad optimization.

THE MOST COMMON MISTAKES IN STORE OPTIMIZATION

Mistake 1. You trust “UX/UI best practices” too much

Heads up, I’m dropping a bomb: there’s no such thing as “UX/UI best practices” that work everywhere. Just because something works in one e-commerce business doesn’t mean it will work in yours too. UX/UI best practices are popular because they free specialists, managers and e-commerce business owners from the tedious work of analyzing their customers’ problems and challenges.

The problem with best practices is that they don’t account for the fact that every store is different – different business characteristics, audience, problems, and it needs different solutions to those problems. I’ve tested the same changes across different stores, and in some of them those changes worked well while in others the opposite was true.

The internet is full of checklists with “universal” solutions that will improve UX/UI and sales. Unfortunately, it often happens that adding a “best practice” lowers conversion.

Example:

Shipping buy-now-pay-later (Klarna / PayPo / Twisto) in a store is one of the best practices and you’d assume a move like that would lift sales. Well, no, not always. In one store, removing buy-now-pay-later was tested. The result? An 8% conversion lift. Solutions like Klarna can have a positive effect on sales, but as you can see, they can also cause losses. Best practices aren’t always best for your business. They’re worth testing.

Mistake 2. You copy the competition

Copying the competition is a trap pretty much everyone has fallen into at some point. Me included. The typical story: you check out a competitor’s online store, look around a bit, decide “Oh, that’s nice, I want one too”, ship that solution yourself and feel like you’re optimizing your store. Beautiful, right? Unfortunately, too beautiful to work.

There’s a strong chance your competitor isn’t optimizing their store for conversion, or that an element that works for them won’t work for you. Good advice: never copy solutions from a competitor’s store just like that. If you do want to take inspiration from a solution, remember to test it first.

Mistake 3. You focus too much on whether the store is “nice” and “on-trend”

A “nice-looking” store doesn’t have to mean high sales at all. Of course there’s nothing wrong with building modern, attractive designs. A store’s design is an important part of brand communication and does affect conversion, if only because it affects credibility.

But far too often, building a “nice” interface takes up much more time than other key elements that drive online store sales growth. Unfortunately, if:

• product information is insufficient or unclear,
• visitors don’t understand the value you’re giving them,
• the purchase path has elements causing friction, distraction and frustration

… then even the most beautiful design won’t get you sales growth in your store.

Mistake 4. You work off opinions, guesswork and intuition

A great many decisions about making changes in online stores are made on intuition and guesswork rather than on research and test results.

When you redesign a store, you usually listen to opinions. The opinions of designers, marketers, the boss or the agency. But the fact that you or your boss prefer a one-step checkout to a multi-step one shouldn’t be the basis for a final decision.

As the owner of a CRO agency, I often get asked: “You’re the expert, so look at my site and tell me what I should change to raise conversion.” And yes, I often spot potential sales killers that someone else might miss, but that doesn’t mean I’m right with 100% certainty. It’s only my opinion. And my opinion doesn’t actually matter until I’ve verified it through research and testing.

The only opinion that really counts is the customers’. Unfortunately, their voice often isn’t heard or taken into account when changes are being designed.

Mistake 5. You do a total redesign when you don’t need one

By “total redesign” I mean a thorough reworking of the visual template, changing the layout, content and copy on the site. Decisions about a radical store redesign are often made too hastily, and end up causing enormous losses. One of the best-known examples of a failed redesign is Marks&Spencer. The new Marks&Spencer store cost £150 million (quite a lot) and took 2 years to build.

The results of the new store? An 8% conversion drop against the previous one. Unfortunately Marks&Spencer isn’t an exception. The e-commerce world knows plenty of cases of radical redesigns that ended in falling sales. Doing a total redesign, you introduce too many changes at once, and when sales fall or rise after the redesign, you won’t actually know which changes caused the rise and which the fall. Let me use a short example.

Let’s say that for some reason you want to make a lot of changes to your store’s appearance at once. You simultaneously introduce: a new product page, a new cart and a new checkout. It turns out that after shipping the changes, sales fall by 20%. You’re now at a point where you don’t know what caused the drop – rebuilding the product page, the cart or the checkout? Maybe the new product page layout generates slightly more conversion, but rebuilding the checkout form significantly lowered sales. Or maybe the other way round? You don’t know.

If your current store works “well enough” and you have many returning customers who are used to the current interface, redesigning gradually is the better choice.

Gradual redesign is a process where you test individual changes:

• current product page vs. new product page,
• current cart vs. new cart,
• current checkout vs. new checkout,

and you observe what effect they have on sales. You only ship them if the test results show they lift conversion. That way you’ll be certain the changes you want to ship are making you money.

Of course, in some cases a more radical redesign is necessary. If:

• you’ve found plenty of conversion killers but your store traffic is so low that testing every change would take years,
• you’ve hit a local maximum on your current technology and can’t squeeze any more out of it,
• the technology your store runs on is outdated,
• the current design looks amateurish,

you won’t achieve much without radically rebuilding the store’s appearance.

But if you’re not in one of the situations above, don’t do a radical redesign. Redesigning your store gradually will be the better solution for you. That’s enough about mistakes. Let’s get to the heart of good optimization.

WANT TO MAKE CHANGES IN YOUR STORE? FIRST FIND OUT WHERE AND WHY YOUR CONVERSION IS LEAKING

All the mistakes above have a lot in common. Specifically, they come from the fact that:

• you don’t run research and analysis, so you have no data on your customers’ behavior and you don’t know where the conversion problems are,
• and/or you don’t run A/B tests, so you don’t verify whether the changes you’re making genuinely bring additional revenue to your business.

Without research and a diagnosis – without naming the problem – that problem is hard to beat. It’s like a doctor treating you before diagnosing the illness. You wouldn’t want to be treated that way, would you? Your store (and its conversion) doesn’t handle treatment without a diagnosis any better. So before you start changing (treating) your store, run research. Understand your store’s situation and identify the key problems holding back its sales growth.

It’s best if I use examples of bad and good optimization here.

Examples of bad store optimization:

1. Hey, the competition has a nice banner with video in the background on their homepage. Let’s do the same!

2. The agency that audited our site wrote that we should ship a one-step checkout. We asked why – they said it’s one of the best practices. I suppose we should trust their instincts… Let’s ship it.

An example of good store optimization:

I noticed in Google Analytics that a great many visitors go back from checkout to the cart. There’s obviously a problem here. So I checked whether our checkout was loading too slowly. According to Google Analytics the checkout loads in only 2 seconds, so that’s not why users are going back to the cart. So I ran a manual expert review of the checkout and cart. I noticed that the checkout summary has no product images and no shipping time information. Both of those are in the cart. I also looked at Hotjar and analyzed recordings of users who were on the checkout page. It turned out a large share of them fill in the form, choose payment and delivery methods, and then desperately hunt for some piece of information – scrolling up and down the page, then going back to the cart. In all likelihood the absence of product images and shipping time information in the checkout is lowering conversion in our store and we’re losing money because of it. The best thing now is to run an A/B test and verify whether that hypothesis holds. If it does, we ship the change to the store.

What the example above shows is that good optimization consists of:

1. Analyzing WHERE the problem is that’s leaking your money.
2. Analyzing WHY there’s a problem in that place.
3. Finding a potential solution to that problem.
4. Testing that solution and checking whether it actually increases revenue.

Which research methods are worth using to find conversion problems?

In the example above I mentioned:

• Google Analytics analysis (you can also use another quantitative analysis tool),
• expert review of the purchase path – manually going through the store looking for elements that may be lowering conversion,
• Hotjar and session recording analysis

What else can help you find the problems lowering sales?

1. chat message analysis or a conversation with customer support,
2. user testing,
3. scrollmap and heatmap analysis,
4. form analysis,
5. customer surveys,
6. customer interviews
7. card sorting,
8. tree testing,
9. review mining.

Once you’ve found the conversion problems and have potential solutions to them, it’s worth running A/B tests next.

WHY DO A/B TESTS MATTER SO MUCH?

Above all because you can back your decision with data. A test verifies whether making a specific change in your store is a good idea. From a test you can learn whether the changed version of the page sells better than the original. Take a look at a test we ran for one of our clients earlier this year.

Next to the stars on the product page, we added the specific number of product reviews. At first glance you’d think that’s a detail and a change like that won’t affect sales. And yet details like that do the work. That change lifted the store’s conversion by over 4%. For many e-commerce businesses, growth like that means millions in additional revenue.

SUMMARY

Store optimization is a many-layered topic deserving its own book. But I hope this short article has shown you the difference between good and bad optimization. Remember:

• Research and looking for the answers to WHERE and WHY you have conversion problems are the key to growing revenue (not a competitor’s site, a list of “best practices” or a designer’s instincts),
• Testing changes can help you save or make thousands.