In Short
- Even web designers get it wrong: While rebuilding my own agency website, I discovered it had drifted into many of the same bad habits I warn clients about — proving that constant vigilance matters more than occasional redesigns.
- Decorative images do more harm than good: Eyetracking research consistently shows that users ignore stock photography and treat anything that looks promotional as an ad. Your hero image of a sunset is invisible.
- Ads and clickbait trained us to distrust the web: Decades of banner ads, sensational headlines, and generic stock imagery have conditioned users to filter out anything that looks like marketing — including your legitimate content.
- SEO gaming created a vicious cycle: Google's evolving algorithms forced content creators to optimise for machines rather than humans, producing oceans of thin, keyword-stuffed pages that nobody actually wants to read.
- The fix is old-fashioned: User-centred design, plain language, honest content, and genuine trust-building — principles that predate search engines — are what Google's AI-powered ranking systems now reward.
I am in the middle of rebuilding my own website. As someone who has been designing and building websites for small businesses since 2009, you would think my own site would be a shining example of best practice. It was not.
That is a genuinely uncomfortable thing to admit. But honesty is the whole point of this article, so here it is: despite regular visual refreshes over the years, my website had quietly drifted into the same patterns I find myself advising clients against. It lacked a coherent user-centred design focus. It did not build trust progressively. It talked about what I do rather than what my clients need. Some of the service descriptions were AI-generated filler that referenced "tax professionals" on a web design agency's homepage. In short, it was not practising what I preach.
The rebuild has been genuinely humbling — and unexpectedly educational. It forced me to confront not just what was wrong with my own site, but why so many websites across the internet have ended up in the same sorry state. The answer, I think, is a story about how advertising, search algorithms, and stock photography collectively broke the web — and why the fix looks a lot like going back to basics.
What I Found Wrong with My Own Website
When I sat down to audit my site objectively — the way I would assess a new client's site — the problems were obvious. The homepage led with a carousel (a design pattern that NNGroup research shows only 2% of users interact with). The service descriptions were written to sound impressive rather than to answer the questions visitors actually arrive with. The navigation was structured around my business categories, not around the tasks my visitors want to accomplish.
Most tellingly, the site was full of images that served no informational purpose. Stock-style photography, decorative backgrounds, promotional-looking feature sections — exactly the kind of visual noise that eyetracking research tells us people have learned to ignore completely.
I had fallen into the trap that catches everyone eventually: incremental additions over years that each seem reasonable in isolation but collectively push a site away from its users. A new service page here, a trendy design element there, an AI-generated industry description because seventeen other agencies had them. Death by a thousand well-intentioned cuts.
The Surprising Research on Images
The most eye-opening part of this process was revisiting the research on how users actually interact with images on the web. I had always understood, in theory, that images should serve a purpose. But the depth of the evidence surprised me.
Jakob Nielsen's eyetracking research at NNGroup found that users pay close attention to information-carrying images but ignore purely decorative ones. Real photographs of real people get scrutinised — users spent 10% more time looking at staff portrait photos than reading the biographies beneath them. But stock photos of generic people? Completely ignored. The eye simply skips over them as if they are not there.
It goes further than that. Kathryn Whitenton's research on image-focused design demonstrated that "devoting this much emphasis to any single design element necessarily means that users are less likely to notice other elements." Large hero images do not just fail to help — they actively push important content below the fold, forcing users to scroll past decoration to reach substance. The content they came for gets buried beneath the content you wanted to show off.
The implication for small business websites is stark: that beautiful hero image you paid a photographer for (or worse, licensed from a stock library) is not just failing to impress visitors. It is actively making your site harder to use. It is pushing your value proposition, your pricing, and your calls to action out of sight.
How We Got Here: The Ad-Poisoned Web
Here is the question that kept nagging me through this rebuild: how did we, as an industry and as web users, arrive at a point where images — the very thing that makes websites visually appealing — became something people have trained themselves to ignore?
My theory is straightforward: advertising broke the contract.
NNGroup has documented banner blindness across three decades of eyetracking research — from 1997 through to their most recent study in 2018. Users do not just ignore ads. They ignore anything that looks like an ad. Anything with fancy formatting, coloured backgrounds, or promotional-looking imagery gets filtered out through what researchers call selective attention. We simply do not have the cognitive capacity to evaluate every visual element on a page, so we have learned to skip the ones that pattern-match to "advertisement."
And it has only got worse. The "faux ad" effect that NNGroup identified on mobile — where users scroll past legitimate content because it visually resembles an inline ad — is a direct consequence of years of ad saturation. When every second element on a news article is a sponsored post disguised as content, users develop a shoot-first-ask-questions-never approach to anything that stands out visually.
Add clickbait to the mix and you have a perfect storm of distrust. How many times have you clicked on a headline promising "You Won't Believe What Happened" only to find three paragraphs of filler, twelve ads, and a conclusion that amounts to nothing? We have all been burned often enough that we now approach visually prominent web content with the same wariness we reserve for unsolicited emails from Nigerian princes. The blonde, curly-headed child running through a sunny wheat field in your website's hero section? To a modern web user's trained eye, that is not an aspirational image. It is a warning sign.
Google's Role in the Problem
Google deserves a significant share of responsibility for the state we are in, even though their stated intentions have always been to improve the web.
For years, Google's ranking algorithms rewarded specific, measurable signals: keyword density, backlink quantity, page speed, content length. In response, an entire industry grew up around gaming those signals. Not because practitioners were malicious, but because ranking determines survival. If your client's plumbing business does not appear on the first page of results for "emergency plumber Melbourne," that business loses work to the one that does. The stakes are existential.
I am guilty of this myself. Not for any cynical reason — I genuinely care about my clients' success — but because towing the Google line is a professional obligation when your clients' livelihoods depend on search visibility. I have optimised meta descriptions, written content to target specific keyword clusters, and structured pages around what Google's algorithms were known to reward rather than what would genuinely serve the visitor best.
The irony is not lost on me. At the same time, I have been advising clients to steer well clear of dedicated SEO services. The SEO industry is rife with overpriced packages that promise first-page rankings through "proprietary strategies" that are mostly just repackaged common sense — or worse, black-hat techniques that can get a site penalised. Google's own spam policies explicitly target the tactics many SEO agencies still peddle: keyword stuffing, link schemes, scaled content abuse, and cloaking.
So we ended up in a perverse cycle. Google wanted better content. Businesses hired SEO agencies to game the algorithms instead of creating better content. Google tightened the algorithms. SEO agencies adapted their gaming. The actual quality of web content — the stuff real humans read with real questions — became collateral damage in an arms race between a search engine and an optimisation industry.
Maybe AI Finally Breaks the Cycle
Here is where I allow myself some cautious optimism.
Google's recent core updates — particularly the helpful content guidelines — signal a genuine shift. The language Google uses now is strikingly different from five years ago. They are explicitly asking content creators: "Does your content clearly demonstrate first-hand expertise?" and "After reading your content, will someone leave feeling they've learned enough about a topic to help achieve their goal?"
These are not algorithmic questions. They are human questions. And I believe AI is what has finally given Google the tools to evaluate content this way — not by counting keywords or measuring backlinks, but by understanding whether a page actually helps someone.
Google's E-E-A-T framework (Experience, Expertise, Authoritativeness, Trustworthiness) is the clearest expression of this shift. The first "E" — Experience — was added in December 2022, and it specifically rewards content created by someone who has actually done the thing they are writing about. Not someone who has summarised five other articles on the topic. Not an AI that has been prompted to produce 2,000 words on "web design trends." Someone with genuine, demonstrable, first-hand experience.
This is, frankly, wonderful news for practitioners. And terrible news for content mills.
Going Back to What Actually Works
What strikes me most about this moment in web design is how much the "new" best practice resembles the principles I was taught when I was first learning to build interfaces over 30 years ago — long before Google existed, let alone dominated how we think about the web.
In those days, we designed kiosk interfaces and application UIs grounded in cognitive psychology. We studied affordance — the idea, popularised by Don Norman in The Design of Everyday Things, that a well-designed interface should communicate its function without instruction. We obsessed over reducing cognitive load, the total amount of mental effort required to use an interface. We built trust through consistency, clarity, and honest communication.
None of these principles had anything to do with search engine optimisation, because there was no search engine to optimise for. They were about one thing: making the interface work for the person using it. User-centred design was not a buzzword. It was the only design philosophy that made sense.
Then the web exploded, search engines arrived, and a generation of web designers (myself included) gradually shifted focus from "Does this interface serve the user?" to "Does this page rank on Google?" We did not abandon the fundamentals entirely, but we let them erode under the relentless pressure of SEO requirements, client demands for "more content," and the temptation of stock photography that made pages look polished without adding any real value.
Stanford's Persuasive Technology Lab published their web credibility guidelines back in 2002 — the product of three years of research with over 4,500 participants. Guideline 9 reads: "Use restraint with any promotional content. Avoid pop-up ads. As for writing style, try to be clear, direct, and sincere." Guideline 7: "Sites win credibility points by being both easy to use and useful." Twenty-four years later, these are essentially what Google is now using AI to enforce.
The wheel has come full circle.
What This Means for Your Website
If you are a small business owner reading this, here is the practical takeaway. And I want you to trust me on this, because I am someone who just had to eat his own humble pie about these exact issues on his own website.
Your website should not be full of clickbait. Sensational headlines that promise revelations and deliver filler are actively harmful. Google's March 2024 core update explicitly targets content "made to attract clicks" rather than to inform. Your visitors can smell it too.
Your website should not be full of technical jargon. Your customers do not care about your "scalable enterprise-grade CMS architecture." They care about whether their website will work on a phone, whether it will help people find their business, and how much it will cost. Talk to them in their language, not yours.
Your website should not be full of distracting background images. Every decorative image that does not directly answer a visitor's question is cognitive noise. It makes your site harder to use and, thanks to decades of ad conditioning, may actually cause visitors to skip the sections it appears in.
Your website should not be full of marketing speak. "Cutting-edge solutions," "industry-leading innovation," "synergistic approaches" — this language does the opposite of what you think it does. NNGroup research shows that promotional copy imposes cognitive burden and reduces comprehension. It makes visitors trust you less, not more.
Instead, your website should do one thing brilliantly: help the visitor understand that you genuinely provide what they are looking for.
That means clear, honest descriptions of your services and pricing. It means real examples of your work. It means straightforward answers to the questions they arrived with. It means building trust through transparency, not through polish. If your website helps someone understand that you can solve their problem — quickly, clearly, without making them wade through marketing noise — you have won. Both the visitor and the algorithm.
The Uncomfortable Lesson
Rebuilding my own site has taught me something I did not expect: the cobbler's children really do go barefoot. I knew the principles. I teach them to clients. I have written about transactional design and the IA matrix on this very blog. But knowing and doing are different things, and the slow drift of incremental compromises had moved my own site further from those principles than I had realised.
If it can happen to someone with over 30 years of interface design experience, it can happen to anyone. The good news is that the fix is not complicated. Strip away the noise. Lead with clarity. Talk to your visitors like they are intelligent people with a real question, not eyeballs to be captured with stock imagery and buzzwords.
The web does not need more clever marketing. It needs more honest communication. And for the first time in a long while, the algorithms seem to agree.
