Skip to content 
If you have been investing in SEO for any meaningful amount of time and the needle is not moving, I can almost guarantee the problem is not that SEO has stopped working. SEO works. I have built businesses on it for more than twenty years. I have also watched a remarkable number of intelligent, well-funded companies pour budget into it and get back almost nothing. The question is rarely whether the discipline still delivers. The question is whether the work being done is actually SEO, or whether it is a collection of tactics that look like SEO but never quite add up to a strategy.
In this article I want to walk through the seven mistakes I see most often. None of them are obscure. None of them require a technical audit to spot. They are mistakes that smart founders, experienced marketers and well-meaning agencies make every day, often because the playbook they are using was correct five years ago and quietly stopped being correct since. If your SEO programme is underperforming, it is almost certainly making at least three of the seven.
Mistake 1: Treating SEO as a project rather than a system
The first and most expensive mistake is the most foundational one. SEO is not a project that you commission, complete and move on from. It is a continuous system, with inputs, outputs and feedback loops, and it has to be operated as such. The most common version of this mistake is the agency engagement that runs for six months, gets paused while the team “evaluates results”, and never restarts at the same cadence. Six months in, all of the work is in the early-compounding phase — indexed but not yet ranking, ranking but not yet driving revenue. Pull the plug at that point and you have funded the cost of the asset without ever capturing the return.
In every successful SEO programme I have worked on or sold, the work was treated as ongoing infrastructure, not as a campaign. There was always something being published. There was always a backlog of internal links being added. There was always a small set of pages being refreshed. The compounding only happens when the system runs continuously. Every time you stop and restart, you reset the clock.
If you are about to commit to a new round of SEO spend, the first question to ask is not “who should we hire?”. It is “what is the smallest version of this system we can run forever?” Half the budget for twice as long will outperform full budget for half as long, almost without exception.
Mistake 2: Optimising for keywords instead of for search intent
The second mistake is the one that haunts the most experienced SEO teams, because it is the one the original playbook taught them to make. For years, the discipline was about keywords. You picked a target phrase, you wrote a page for that phrase, you stuffed the phrase into the title, the H1 and the first paragraph, and you crossed your fingers.
Google has been quietly moving away from rewarding that approach for at least a decade, and the move is now effectively complete. Modern search ranks pages on whether they satisfy the underlying intent behind a query — the actual job the searcher is trying to do — not on whether the words match. A page that targets the phrase “best CRM for small business” but does not actually compare CRMs, give clear recommendations and address the trade-offs that small business owners care about will lose, every time, to a page that does all of those things using slightly different language.
The fix is to start every brief with a single question: when somebody types this query, what are they actually trying to accomplish? If the answer is “decide between three options”, write a comparison. If it is “understand whether this applies to me”, write an explainer. If it is “do it themselves right now”, write a step-by-step. The keyword is a label for the intent, not a substitute for it.
Mistake 3: Obsessing over technical SEO at the expense of substance
This one tends to show up in companies with engineering DNA. The team gets enthusiastic about Core Web Vitals, structured data, crawl budgets and log file analysis, and over the course of a quarter the technical health of the site goes from a 78 to a 96. None of which translates into more traffic, because the underlying problem was never technical.
Technical SEO matters — of course it matters. A site that cannot be crawled cannot rank. A page that takes seven seconds to load on mobile will lose to one that takes two. But for the vast majority of businesses I look at, the technical foundations are already good enough. The reason the site is not ranking is not that Google cannot read it. It is that Google does not see a reason to choose it over the dozen other pages competing for the same query.
If you are spending more than about a quarter of your SEO time on technical work in any given month after the first ninety days, you have probably misallocated. The marginal hour goes further on stronger content, better internal linking and earned authority than on shaving another hundred milliseconds off Time to First Byte.
Mistake 4: Publishing thin, generic, or AI-slop content with no point of view
The volume of content being published every day has roughly tripled since generative AI tools became mainstream, and the average quality has fallen off a cliff. Most of what is now being put online is what I call AI-slop: technically grammatical, broadly accurate, and entirely interchangeable with what every competitor is also publishing. Google has noticed. The Helpful Content Update and the broader trajectory of search since make it very clear that pages with no original perspective, no first-hand experience and no genuine expertise are no longer competitive.
This does not mean you cannot use AI in your content workflow. I use it constantly. It means that AI-generated text without a human point of view layered into it is not content, it is filler, and filler does not rank in 2026. The pages that win now are the ones that include data the business actually owns, opinions the author is willing to put their name to, examples that nobody else could have given, and structure that reflects how the topic actually plays out in practice rather than how it looks on a content brief.
If you cannot read a page back to yourself and identify a single sentence that nobody else in your category could have written, that page has nothing to say. Take it down or rewrite it. Empty pages are not neutral — they actively pull down the perceived quality of the rest of your site.
Mistake 5: No internal linking strategy
Internal linking is the most under-leveraged tool in SEO, and the gap between businesses that take it seriously and businesses that do not is enormous. Every page on your site is a vote you cast in favour of other pages on your site. The pages you link to most often, with the clearest anchor text, in the most relevant context, are the pages you are signalling matter. Most companies cast those votes essentially at random.
The pattern I see again and again in audits is this: a thin homepage links to a few feature pages, the feature pages link only back to the homepage, the blog publishes content that links nowhere except to the homepage, and the company’s actual money pages — the pages they would pay tens of pounds per click to drive traffic to — are functionally orphaned. Three or four internal links is all most of them have. No wonder they do not rank.
A working internal linking strategy is not complicated. Identify your money pages. Make a deliberate plan that every piece of supporting content includes a contextually relevant link to one of them, with anchor text that describes what the destination page actually offers. Audit your existing content and retro-fit those links. Done well, this single discipline will outperform months of new content production.
Mistake 6: No external authority — nothing the rest of the internet says about you
Internal signals will only take you so far. Beyond a certain point, you cannot out-write your way past competitors who are being talked about on the rest of the internet while you are not. Google has been clear about this for years through the E-E-A-T framework, and the recent generation of AI-driven search has made it more important still: the systems answering queries are not just looking at your page, they are looking at what other respected sources say about your domain, your authors and your business. If the answer is “not much”, you are not in the consideration set.
This is where most SEO programmes silently underperform, and it is the bridge between SEO and digital PR. You can have technically excellent content with thoughtful internal linking and still be invisible if no third-party publishers, no recognisable authors and no industry communities have ever cited you. The fix is not buying links — that strategy is now actively dangerous — but earning them: original research the industry will reference, expert commentary that journalists will pick up, founder content that gets discussed.
If your SEO agency does not have a credible answer to the question “how are we earning external authority over the next twelve months?”, you are running half an SEO programme. The other half is what is keeping you off the first page.
Mistake 7: Measuring the wrong things
The final mistake is the one that masks all of the others. SEO has a measurement problem because the industry got attached to the wrong metrics early and never quite let them go. Average position. Total indexed pages. “Keywords ranking in the top 10.” Pageviews. None of these are useless, but none of them are what your business is actually trying to do.
The metrics that matter are the ones that connect organic search to revenue. Branded search volume — are more people Googling you by name month over month? Rankings on commercially decisive queries — the handful of phrases that, if you owned them, would change your business. Conversion rate from organic visitors. Pipeline sourced. Customer acquisition cost compared to paid alternatives. If your dashboard is not built around those numbers, you cannot tell whether the work is succeeding, and you cannot make a credible case to renew the budget.
The hardest version of this mistake is when teams confuse traffic with progress. A blog post that pulls in twenty thousand visitors a month from a vaguely related informational query is, in revenue terms, often worth less than a single page that ranks third for a buying-intent query and gets two hundred visits. Volume is the easiest metric to grow and the easiest to fake yourself into believing things are working. Resist.
How to tell which of these is hurting you most
If you are reading this and recognising more than one of the seven, you are in good company. Almost every SEO audit I run finds at least three. The order of operations matters, though. There is no point earning external authority for pages that are written for the wrong intent. There is no point publishing more content if the existing content is orphaned by your internal linking. There is no point measuring revenue impact if the system itself is being run as a stop-start project rather than as ongoing infrastructure.
The sequence I would suggest, if you are starting from a flat performance line, looks like this. First, fix how you are operating the system — commit to a smaller, sustainable cadence rather than a heroic short burst. Second, audit your existing content for intent fit and rewrite the pages that are closest to converting. Third, get your internal linking right; this is almost free and almost always undervalued. Fourth, layer in earned authority through a real digital PR programme. Only then does it pay to invest heavily in new content production or technical refinement.
If you would like a frank, outside view on which of these is the bottleneck for your business, the SEO & Content Marketing service page on this site explains how I work with clients on exactly this question. Get in touch and we can take a look. There are very few SEO problems that are genuinely as bad as they feel — most of them are one or two missing pieces away from working properly.
