7 Website 'Best Practices' Your Designer Still Swears By That Stopped Working Years Ago

Someone is buying a hero carousel right now. Somewhere, a designer is telling a client that a rotating banner adds "engagement," and the client is nodding, because it sounds reasonable and the designer has a portfolio.

The data on carousels has been public since 2013. Erik Runyon measured the University of Notre Dame homepage and found that about 1 percent of visitors clicked the carousel at all, and 84 percent of those clicks went to slide one. Slides two through five split the leftovers. That study is old enough to be in middle school, and carousels still show up in half the competitor quotes my clients forward me.

This is a list of seven outdated web design best practices that agencies actively sell in 2026. Not the ancient stuff. You don't need another article warning you about Flash intros and hit counters. These are the ones that look professional, cost real money, and quietly hurt your conversions.

Why "Best Practices" Calcify

Here's the mechanism nobody talks about: a best practice is a snapshot of what worked when your designer learned their craft.

Around 2018, exit popups genuinely grew email lists. Chatbots were novel enough that people actually used them. Video backgrounds signaled "premium" because bandwidth had just caught up. All of it worked, for a while, and it went into the slide decks.

Then users adapted. Google changed the rules. Attention patterns shifted. The slide decks did not. Agencies sell what they can produce efficiently, and what they produce efficiently is what they've built two hundred times before. That's how outdated web design best practices survive: not because they work, but because they're efficient to sell.

There's a simple test: ask your designer why a feature is on the proposal. If the answer is "it's best practice," that's not a reason. That's the absence of one.

1. Hero Carousels

Start with the numbers from the intro, because they haven't improved in over a decade: roughly 1 percent interaction, with the overwhelming majority of clicks going to the first slide. Slides two through five are where content goes to be technically present.

The reason is boring and human. The Nielsen Norman Group has documented banner blindness for over two decades: anything that moves and rotates on a website pattern-matches to advertising, and people have trained themselves not to look at ads. Your carousel isn't showcasing five messages. It's marking one region of the page as ignorable.

And there's a deeper problem than the click rate. A carousel is what a homepage gets when a company can't decide what it's about. Five departments each wanted the top spot, so the designer gave it to all of them, which means nobody got it. Good web design serves the user's decision-making, and a rotating banner asks the user to make sense of five pitches on a timer.

What to do instead: one static hero. One headline, one sentence, one button. If a message doesn't earn that spot, it gets a section further down the page, where people can actually read it.

2. Autoplay Video Backgrounds

Open the network tab on a site with a video background and look at the numbers. That looping drone shot is routinely the single heaviest asset on the page, often several megabytes even when compressed, and it starts competing for bandwidth exactly when the browser is trying to render your headline and your Largest Contentful Paint element.

So you pay twice. Once in load time, once in attention: the human eye is hardwired to track motion, and your headline sits on top of the one thing on the page that's moving.

On mobile, where most of your traffic lives, the payoff often doesn't even ship. Browsers restrict autoplay, connections choke on the file, and users get a janky poster frame that took four seconds to settle.

If the brand genuinely needs motion, there are honest versions: a short, aggressively compressed loop, lazy loaded, with a proper poster image. But in most cases the premium feeling clients want comes from typography, spacing, and one strong image. Those cost nothing to load.

3. Exit-Intent Popups

Google put a date on this one: since January 2017, intrusive interstitials on mobile are an explicit negative ranking signal. That's how long the industry has known, officially, from the search engine itself, that aggressive popups are a liability. They still show up in conversion audits as a "quick win."

The technical cost shows up in Core Web Vitals. Popups that inject late cause layout shift, and the JavaScript that watches your mouse velocity to detect "exit intent" is one more script running on every single page view, taxing interactivity for the 97 percent of visitors who were never going to subscribe anyway.

The trust cost is worse and harder to measure. When your popup fires, you've taught the visitor that reading your site requires defending themselves. People now hunt for the X before they process the offer. A high bounce rate rarely has one single cause, but I've never once seen an aggressive popup make it better.

The alternative isn't abandoning email capture. An inline form at the end of a good post converts fewer people per view, but it converts them because the content earned it. That list actually opens your emails.

4. Chatbot Widgets Nobody Staffs

Before we talk about staffing, weigh the widget itself: popular third-party chat scripts ship hundreds of kilobytes of JavaScript, and they load on every page whether the visitor chats or not. That's a permanent tax on interactivity for a feature most visitors never touch.

But the real damage isn't the payload. A chat bubble is a promise: "someone is here." On a small company site, that promise is almost always false. The bubble says "usually replies within a few hours." The visitor asks a question, waits, gets nothing, and leaves with a worse impression than if the button had never existed. The silence gets blamed on your business, not on the widget.

Here's the honest question: do you have a person whose job includes answering that chat within minutes, during stated hours? If yes, live chat is genuinely great. If no, a contact form with a truthful response time beats a fake pulse. "We reply within one business day," kept, beats "online now," broken.

5. Cramming Everything Above the Fold

"The fold" is a newspaper term. It made sense when there was exactly one broadsheet format. On the web there is no fold: there are hundreds of viewport sizes, and the line your designer optimizes for on a 27-inch monitor doesn't exist on the phone where your customer actually is.

Meanwhile, scrolling is the cheapest action in computing. Every user does it, on every device, without thinking. Smartphones turned it into a reflex years ago.

But the belief persists, and it has a signature look: shrunken type, four competing buttons, a logo bar, a mission statement, and, inevitably, a carousel, because rotating content is how you "fit everything" into one viewport. The result isn't a page that communicates everything. It's a page that communicates nothing, loudly. Professional web design is mostly the discipline of deciding what matters most and letting everything else breathe further down.

One clear message per screen. Let the page be long. Nobody has ever left a website because they had to scroll. They leave because scrolling didn't reward them.

6. Chasing a Perfect PageSpeed 100

I'll say something that sounds strange coming from someone who sells performance work: the last ten points of a PageSpeed score are usually procrastination.

The distinction that matters is lab versus field. Lighthouse gives you a lab score, a synthetic number from a simulated device. Google's ranking systems care about field data: real users passing real thresholds, LCP under 2.5 seconds, INP under 200 milliseconds, CLS under 0.1. That's a pass/fail bar, not a leaderboard. Testing your real-world load time tells you which side of the bar you're on, and that's the number worth paying for.

Taking a site from 45 to 90 is real work with real returns. I've done it plenty of times and watched rankings and conversions move. Taking a site from 92 to 100 mostly produces screenshots for the agency's own marketing. I've watched developers spend days inlining critical CSS for eight lab points while the contact form on the same site was silently failing.

Fast enough is a real threshold. Past it, the next conversion win is in your message, your offer, your proof. Boring, fast, one clear message per page beats a trophy score on a page nobody understands.

7. Stock Photos as "Professional Polish"

The handshake photo. The headset woman. The diverse team laughing at a laptop that isn't turned on. For years, designers added these because they made small companies look bigger, and honestly, it used to work.

In 2026 the polish signal has flipped. AI image generators have flooded the web with exactly this kind of glossy, generic imagery, and visitors have recalibrated. When every site looks the same, polished-generic reads as "possibly not a real company." The photo that used to say "established" now says "template."

Real photos win, even mediocre ones. Your actual workshop, your actual face, your actual slightly cluttered desk. They're proof of existence, and proof of existence is the scarcest asset on the modern web. A phone photo with decent light beats a $50 stock license, and it definitely beats the AI-generated office where the plants have too many leaves.

So, Is Your Website Outdated?

Not because it's three years old. A three-year-old site with one clear message, fast pages, and honest photos is fine. Your website is outdated when it's still running 2018's playbook: the carousel, the popup ambush, the ghost chatbot, the trophy score.

The frustrating part is that the fix is less work, not more. Delete the carousel and pick your one message. Delete the popup and earn the signup. Delete the chatbot and tell the truth about response times. Replacing outdated web design best practices almost always means removing things, which is exactly why it's hard to sell: nobody feels good paying for fewer features. I still catch myself over-explaining this to clients, because "I'll remove things" is a strange pitch.

If your designer swears by any of the seven above, ask them for evidence newer than the practice itself. And if you want a second pair of eyes on what your site actually needs, let's talk. Removing things is cheap. Keeping the wrong things isn't.