AMP SEO

I rarely recommend AMP for most sites today, but it still has a narrow use case. AMP stands for Accelerated Mobile Pages, an open-source framework originally created by Google to compete with Facebook Instant Articles and Apple News. It strips a page down to a strict HTML structure, bans most custom JavaScript, and often delivers the page from Google's own cache. The result can be a load time under one second, but the same is possible with a well-optimised standard page.

The catch is that AMP sacrifices custom interactivity. If your site relies on complex JavaScript for navigation, forms, or dynamic content, AMP will force you to use its own limited component library. I have seen projects abandoned because the design system could not fit AMP's constraints. For most sites, the juice is not worth the squeeze. A [solid technical SEO setup](/technical-seo/) with deferred JavaScript and a content delivery network can match or beat AMP speeds without the limitations.

That said, AMP has one genuine strength: it forces you to audit every script and style on the page. You end up asking 'does this tracking pixel actually need to block rendering?' That discipline is valuable, but you can apply it to your standard web pages too. I would only reach for AMP if your mobile pages consistently load above three seconds and you have no immediate way to prune third-party scripts.

AMP matters in two specific scenarios, and I have seen both pay off. The first is news publishing. Google's Top Stories carousel still uses AMP in many regions, and a validated AMP page can boost your visibility there. A news site I audited saw a 20% click-through rate increase from Top Stories after implementing AMP, but only because they fixed every validation error. The second scenario is sites with heavy, unchangeable ad or tracking scripts. If your business demands dozens of third-party tags and you cannot negotiate with the vendors, AMP acts as a fast fallback that strips that bloat away.

But the counter-argument is strong: you are treating a symptom, not the cause. I have seen AMP pages load in 0.8 seconds while the canonical version loads in 3.5 seconds. That gap suggests poor baseline performance. If you clean up the canonical page, you get the same speed without the double maintenance. For e-commerce sites, AMP is rarely worth it because you lose dynamic cart updates, personalised product recommendations, and A/B testing tools. And AMP's cached version can interfere with analytics attribution — you need to configure your tracking carefully.

My decision rule: if you are not a news publisher and you cannot commit to keeping both AMP and canonical pages in sync, skip AMP. Invest that effort in [Core Web Vitals](/core-web-vitals/) and [mobile performance optimisation](/mobile-seo/) directly. The ranking benefit comes from speed, not the framework.

I used to believe AMP was a direct ranking factor because Google promoted it so heavily. I was wrong. Google has stated repeatedly that AMP is not a ranking signal; only page experience and speed matter. I wasted months converting an entire blog to AMP, only to see no organic uplift — and then had to maintain two versions of every page. That taught me to question any SEO tactic that adds complexity without a clear payoff.

Another mistake: I treated AMP as a permanent fix for slow pages. It is a workaround. The moment you add a custom script or a new ad tag to the canonical page but forget to update the AMP version, your mobile experience becomes inconsistent. I once deployed a newsletter sign-up form that broke AMP validation because I used a standard JavaScript library. Google stopped showing my AMP pages in search, and I lost traffic for two weeks before catching the error. Now I always run the AMP Validator after any content change.

I also ignored canonicalisation. I thought rel="amphtml" was enough, but I forgot the reverse: the AMP page must point rel="canonical" back to the standard page. That oversight created a cascade of [duplicate content issues](/canonical-tag/) until a client spotted discrepancies in Search Console. Today I see AMP as a legacy option for very specific use cases, not a universal SEO lever. If I could undo one career decision, it would be the year I spent evangelising AMP to every client.