How I find where a website loses people

While I was working full time as a marketing manager for an e-commerce brand on Shopify, we ran a big sale. The kind of campaign where you expect the products to sell out, so you prep the emails and the stock and everything around it. After our announcement email the traffic showed up too, around a thousand people on the site. But what was weird was that almost none of them were actually getting to the check out.

Looking just at the numbers alone you couldn't tell why. The site looked perfectly fine. When we ran some tests the products were there, the prices were there, the cart worked. So we decided to turned on Lucky Orange, which is basically a plugin that records how real people click and move around the site, and there it was. A second plugin we used for product bundles was glitching right before checkout, catching people at the very last step and making the experience lag for them. Nobody could have guessed it was happening just by looking at the page, or the UI alone.

So we fixed the flow, sent another email letting people know it was sorted and inviting anyone who had left their card to try again, and recovered around 80% of the abandoned carts. For a problem we didn't even know existed an hour earlier, that felt like a success.

But the whole experience taught me something I keep coming back to. Just because something seems right to you, it might be completely broken for a different user, with a different profile, on a different device, in a different mood, across the globe. So putting in the effort to figure out where people actually get stuck is, in my view, one of the most important parts of the job of a designer.

Why you can't see it from your own POV

Here's the uncomfortable truth that most business owners ignore. You might be the worst possible person to judge your own website. The reason is pretty strange: you just know your own website too well. You built the thing, or you've stared at it for years, so your eyes skip over the exact spots a first-time visitor would trip on. For example you already know where the contact form is, so you never feel the three seconds someone else spends hunting for it.

Designers have a name for this, the curse of knowledge, basically the idea that the more familiar you are with something, the harder it gets to imagine not knowing it. Just imagine someone landing on your site without any context. They've never seen your menu, never read your copy, don't know your story, literally nothing. They scan, then decide in a few seconds whether to trust you, and most of the time they leave without telling you why. You can't see it happen, which is exactly why it keeps happening.

My process for evaluating a client's website

So when I sit down to actually evaluate a site, I don't start usually on the homepage. I start where a real visitor would start, which is through social media channels or a quick Google search of the brand name. This approach puts me through the same door a potential client comes through. Because the moment you arrive from a specific platform, you're already carrying expectations. If their Instagram looks one way but the website feels like it belongs to a completely different brand, that gap between the brand voice, the visual identity, and how the actual service is presented is already a important problem worth fixing.

From there I do the thing that's harder to explain but is basically the core of how I work. I stop being myself for a short while and embody the visitor. I set myself a goal a real person would have, trying to find a solution to a problem, and I try to reach that goal. Recently we worked with an HR agency, so I would switch between two types of people: a company trying to outsource their hiring, and an individual looking for help polishing their resume and prepping for interviews. Walking the site as each of them, with their goals in my head, the misalignments and the redundant bits showed up fast. A few questions came up, such as can I tell which service is actually for me? Is it easy to reach out through the contact form, can I find it easy, and does the form even ask for the right things for that kind of client?

It helps that I do a bit of competitive research first, looking at how other brands in the same niche approach their digital presence, combined with pulling references off Pinterest and Behance. So by the time I'm checking the client's site, I'm not just a random visitor anymore. I'm something closer to a power user within their niche who can also diagnose the specific issues, due to my background and experience in both marketing and design. Is it the font, the colors, the layout, the images or does the branding feel outdated for the market they're trying to compete in? Those are the questions I'm holding while I navigate through their website as someone else.

Points where websites lose people most

If I had to name the single most common reason sites underperform, it's that the website or the branding was done too quickly, or without much care, and was never really researched or tested. So a lot of the key aspects a business actually needs to hit its goals end up missing. Sometimes it's the copy, the words doing none of the work they should be doing in order to inform or sell. Sometimes it's the forms, laid out in a way that asks too much or inspires no trust at the exact moment you're asking someone to hand over their details, here having what I like to call a comfortable layout helps.

The common denominator underneath all is a missing eye for detail. People don't realize how fast visitors scan a brand to decide whether they want to do business with it, and it's mostly unconscious. Sure, obsessing over whether the padding around your menu items is evenly centered sounds like a strange thing to lose sleep over during a build, but the visitor scans so quickly that any little inconsistency chips away at their trust, usually without them ever being able to tell you that's what happened. If you have enough of those small things wrong and the site just feels off, even to someone who couldn't explain why.

What I do with my findings

Identifying the friction is one thing. And turning it into a fix without just redesigning for looks is the part that actually matters. What I usually do is I make a mental list of everything that needs to change, then prioritize it by urgency and impact. Not everything carries the same weight in impact for how a site communicates and converts, so the things that matter most for that get handled first.

In an ideal world we'd hand every client a full audit, and walk them through every detail. In practice most of them don't want one, and in reality it can be super overwhelming. They want to see our vision turned into something real and quick, so they can judge for themselves whether it fits what they had in mind in terms of vision. So I tend not to bury them in hyper-specific notes. I'd rather jump in, fix what's broken, and deliver something they can react to fast so we can see if my ideas actually hold.

But this can go wrong though, because it's not a clean process. Left to my own taste, my evaluation tends to build a best-case scenario unfortunately, where I lift the brand up to the industry standard the way I would want it to perform. But the client is the one who actually knows their audience, and more importantly how they want to present themselves to it which is important. So my first pass might be a nice fairy tale to look at, but then the real product comes out of a long back and forth with them, round after round, until it's something their customers will genuinely respond to. The diagnosis is the part I carry. The final shape always comes out of that back and forth, iterating together to find the best solution.

But what if you just want it to look more modern

Every so often a founder comes to me convinced the site just needs to "look more modern," and honestly that's a perfectly fine place to start, it sums up a vibe you're after. A designer can usually take a brief that vague and figure out where things can genuinely be elevated and brought up to date and improved. But I'd push back on one thing first. Those decisions need to be informed, and fully articulated. Start by actually looking at your own metrics, the bounce rate, the click-through, the conversion rate, and see what they're already telling you.

Because "modern" on its own can work against you. If your audience is used to a more familiar way of navigating, and your site is already established, a dramatic overhaul can lose the people you already have. In that case you're better off rolling the changes out in stages. What actually matters is whether the experience works for the user. Get that right, and looking current tends to follow on its own, as a side effect of being clear.

If there's one habit underneath all of this, it's the willingness to stop being yourself for a few minutes. To step out of your own head, your own taste, your own knowledge of where everything is, and stand inside someone who has never seen your site and owes you nothing. That's the whole move, really. The rest is just paying close attention to what they would feel, and being honest about what you find.