If we had to name one thing that slows down more small business websites than anything else, it would be images. Not code, not hosting. Photos.

It makes sense. Phones now take enormous photos, and it’s easy to upload them straight to a website. A gallery of beautiful 4 MB images looks great on fibre and feels painfully slow on a phone in a mall basement.

The good news: images are also one of the easiest things to fix.

Why do images slow websites down?

Every image is a file the visitor’s device has to download. Big files take longer, especially on mobile data. And the largest image at the top of the page is usually your Largest Contentful Paint element, the moment Google uses to judge how fast your page loads.

Three things make images heavier than they need to be:

• Wrong dimensions. A 4000-pixel-wide photo shown in an 800-pixel space.
• No compression. Files saved at maximum quality with extra data the eye can’t see.
• Old formats. JPEG and PNG are less efficient than modern alternatives.

Step 1: Resize to the size you actually need

Find out how wide the image displays on your site. A full-width hero might need around 1600 to 2000 pixels for sharp results on large screens. A blog image in a content column might need 800 to 1200. A team headshot might need 400.

Resize before uploading. It’s the single biggest saving, and it costs nothing.

WordPress automatically creates several smaller copies of each upload and uses responsive image code (srcset) so phones can download a smaller version. That only works well if the original isn’t absurdly large and your theme uses the feature properly.

Step 2: Compress

Compression removes data you won’t notice. Most photos can be compressed significantly before any visible difference appears. Free tools like Squoosh (from Google) let you compare quality side by side, and WordPress plugins can compress automatically on upload.

Step 3: Use WebP or AVIF

WordPress supports uploading WebP natively, and AVIF in recent versions. Many optimisation plugins can also convert your existing library automatically and serve the modern format to browsers that support it.

Logos and icons deserve special mention. If your logo is a large PNG, ask your designer for an SVG version. It’ll be a fraction of the size and crisp on every screen.

Step 4: Prevent layout shift

Images should have width and height attributes so the browser reserves space before the image arrives. Without them, content jumps as images load, which hurts your Cumulative Layout Shift score and annoys visitors. Most modern WordPress themes handle this; some page builders don’t.

Step 5: Lazy-load correctly

Images below the first screen should lazy-load, so they only download as people scroll. WordPress does this by default for many images.

The image at the top of the page should not lazy-load. It should load immediately, ideally with a high priority hint. This one mistake is behind many slow LCP scores we see.

A real-world example

Imagine a boutique wine farm near Paarl. Their homepage has a slider with five landscape photos, each uploaded straight from a professional camera. The gallery page has thirty more.

The fix: replace the slider with one hero image, resize it to around 1920 pixels wide, convert it to WebP or AVIF, and make sure it isn’t lazy-loaded. Bulk-optimise the gallery and let it lazy-load. Nothing about the look of the site changes. The mobile load time usually drops noticeably.

Common image mistakes

Using PNG for photos. PNG photos can be many times larger than necessary.

Uploading, then shrinking with CSS. The browser still downloads the full-size file.

Background images for important content. CSS backgrounds are discovered later by the browser, which can slow LCP.

Forgetting alt text. Not a speed issue, but descriptive alt text helps accessibility, image search and AI systems understand your content.

Optimising once and forgetting. New uploads by staff can undo everything. Automate it.

Quick checklist

• Resize to display size before uploading
• Compress every image
• Use WebP or AVIF for photos, SVG for logos and icons
• Set width and height
• Lazy-load below the fold, never the hero
• Write descriptive alt text
• Automate with a plugin so new uploads stay optimised

Then re-test with our foundation checker. Image fixes often show up immediately in your mobile score. For the rest of the speed picture, see our WordPress speed fixes.

Frequently asked questions

What is the best image format for websites?

For photos, WebP or AVIF. For logos and icons, SVG. Use PNG only when you need transparency and a modern format isn’t suitable.

Does WordPress support WebP and AVIF?

Yes. WordPress supports WebP uploads, and recent versions support AVIF. Plugins can convert existing images automatically.

What size should website images be?

Roughly the largest size they’re displayed. Around 1600 to 2000 pixels wide for full-width hero images, smaller for content images and thumbnails.

Should I lazy-load all images?

No. Lazy-load images below the first screen, but load the main above-the-fold image immediately.

The post Images Are Killing Your Load Time: A Guide to WebP, AVIF and Sizing appeared first on Outview.

Hosting is the ground your website is built on. You can have a beautifully designed, perfectly optimised site, but if the server underneath is overloaded or far away, every page starts with a delay you can’t fix from inside WordPress.

The tricky part is that hosting is also the most over-blamed thing on the internet. So let’s look at how to tell whether it really is the problem.

What does hosting actually affect?

Your host controls what happens before the browser receives the first byte of your page: how quickly the server wakes up, runs WordPress (or serves a cached page) and sends the HTML back. This is measured as Time to First Byte (TTFB).

Everything after that (downloading images, running scripts, painting the page) depends mostly on how the site is built. That’s why a fast host can’t rescue a bloated site, and a lean site can’t fully overcome a slow host.

Google’s web.dev guidance suggests aiming for a TTFB of roughly 0.8 seconds or less as a rough guide. Slow TTFB delays every other metric, including Largest Contentful Paint.

Signs your hosting is the bottleneck

“Reduce initial server response time” appears in PageSpeed Insights. In the full report on pagespeed.web.dev, this diagnostic (or a slow “document request latency” warning) points directly at the server.

It’s slow even with caching on. If a cached page still takes over a second to start arriving, the server or network is the issue.

Speed varies a lot by time of day. Cheap shared hosting puts many websites on one server. When neighbours get busy, you slow down.

The admin area crawls. A sluggish WordPress dashboard, especially when saving pages, often reflects limited server resources.

Frequent downtime or errors. “Error establishing a database connection” or 500 and 502 errors are classic warning signs.

Your server is far from your customers. If your customers are in Cape Town and your server is in another hemisphere with no CDN, every request travels a long way.

How to test it yourself

1. Make sure page caching is active. (If you’re not sure, see our WordPress speed fixes.)
2. Run your homepage through our foundation checker a few times at different times of day.
3. Open the full Google report and look at server response time in the diagnostics.
4. Test a very simple page, such as your privacy policy. If even that is slow to start, the problem is below the page level.

Does server location matter for South African websites?

It can. Data has to physically travel between the server and the visitor. For a business whose customers are mostly in South Africa, hosting in a local data centre (Johannesburg and Cape Town both have well-connected facilities) or using a CDN with local edge locations reduces that distance.

If you serve both South African and international customers, for example a Cape Town tour operator with US and European clients, a CDN is usually the best of both worlds. It serves cached content from whichever location is closest to each visitor.

What to look for in better hosting

Is cheap hosting always bad?

No. A small brochure site with light traffic can run perfectly well on modest shared hosting, especially with a CDN in front. The problem is cheap hosting combined with a heavy site, growing traffic or an online store.

Think about cost differently. If a slow site costs you even one or two enquiries a month, or makes your Google Ads more expensive, the difference between bargain and decent hosting pays for itself quickly.

Common hosting mistakes

Moving hosts without fixing the site. If images are huge and plugins are piled high, a new host gives a small improvement at best.

Choosing on price alone. The cheapest plan often means crowded servers and slow support.

Never checking PHP versions. Hosts don’t always upgrade you automatically.

Running a WooCommerce store on entry-level shared hosting. Stores need more resources because carts and checkouts can’t be fully cached.

How Outview can help

When we build or rebuild websites, we set them up on hosting that suits the site’s size and audience, with caching and CDN configured properly from the start. It’s part of treating speed as foundation work rather than an afterthought. See our website design and e-commerce build services.

Frequently asked questions

How do I know if my web hosting is slow?

Check server response time in PageSpeed Insights. If it’s consistently slow even with caching enabled, or varies a lot by time of day, your hosting is likely the bottleneck.

What is a good Time to First Byte?

Google’s guidance suggests roughly 0.8 seconds or less as a reasonable target.

Should a South African business host its website locally?

If most customers are in South Africa, local hosting or a CDN with local edge locations reduces delays. For mixed audiences, a CDN is often the best option.

Will better hosting improve my SEO?

Indirectly. Faster server response helps Core Web Vitals and user experience, which supports SEO, but content and relevance matter more.

The post Is Your Hosting Slowing Your Website Down? How to Tell appeared first on Outview.

When a WordPress site is slow, the hosting company usually gets the blame first. Sometimes that’s fair. More often, the site is carrying years of extra weight: plugins nobody uses, photos straight off a phone, a page builder doing ten times the work it needs to.

Changing hosts without fixing those things is like moving house without unpacking the boxes. Everything still feels cluttered.

Before you start, run a test on our foundation checker and note your mobile scores. You’ll want a “before” to compare against. And back up your site; some of these changes touch things that matter.

1. Optimise your images

On most small business sites we audit, oversized images are the single biggest problem. A 5 MB photo displayed at 800 pixels wide is pure waste.

Resize images to the largest size they’re actually displayed, compress them, and convert to WebP or AVIF. An image optimisation plugin can do this automatically for new uploads and bulk-process old ones. Our image optimisation guide goes deeper.

2. Turn on page caching

Without caching, WordPress builds every page from scratch for every visitor: running PHP, querying the database, assembling HTML. Page caching stores a ready-made copy and serves that instead.

Many managed hosts include server-level caching. If yours doesn’t, a reputable caching plugin is one of the cheapest speed wins available. Just use one caching plugin, not three; they conflict.

3. Remove plugins you don’t need

Go to your plugins list and ask of each one: “What does this do, and would anyone notice if it disappeared?”

Typical candidates for removal: old slider plugins, social share buttons nobody clicks, duplicate SEO or security plugins, contact form plugins for forms you no longer use, and “temporary” plugins installed for a one-off task years ago. Deactivate first, test the site, then delete.

It’s not purely about the number of plugins. One badly built plugin can be heavier than twenty light ones. But fewer moving parts almost always helps.

4. Question your theme and page builder

Some themes and page builders load large amounts of CSS and JavaScript on every page, whether you use those features or not. If your scores stay poor after the easier fixes, this is often the underlying cause.

Switching themes is a bigger job, but a lightweight, well-coded theme can transform performance. It’s often the point where a rebuild becomes better value than more patching.

5. Update PHP

WordPress runs on PHP, and newer PHP versions are generally faster and more secure. Many older sites are still running outdated versions because nobody changed the setting. Check your hosting control panel. Test on a staging copy first if you can, because very old plugins may not be compatible.

6. Limit third-party scripts

Chat widgets, heatmaps, multiple analytics tools, social feeds, embedded reviews, ad pixels. Each adds JavaScript that slows mobile loading and responsiveness. Our guide to Interaction to Next Paint explains why this hits phones hardest.

Keep the tools that earn their place, delay the rest until after the page loads, and remove anything you don’t actually look at.

7. Lazy-load below-the-fold media (but not the hero)

Lazy-loading means images and videos further down the page only load as people scroll. WordPress does this by default for many images. The exception: your main above-the-fold image should load immediately, or it will hurt your Largest Contentful Paint.

For YouTube embeds, use a lightweight “click to play” placeholder rather than loading the full player upfront.

8. Clean up the database

Over time, WordPress accumulates post revisions, spam comments, expired temporary data and leftovers from deleted plugins. A cleanup won’t usually transform front-end speed on its own, but it can help admin performance and server response on larger sites. Back up first.

9. Use a CDN

A content delivery network stores copies of your files on servers around the world and serves them from the closest location. If your visitors are spread across South Africa, or you also serve customers in the US or UK, a CDN reduces the distance data has to travel. Many offer free tiers.

10. Fix the hero image specifically

We’ve mentioned it twice already, and it deserves its own spot. The large image at the top of your homepage and landing pages is usually your LCP element. Make sure it’s compressed, correctly sized, in a modern format, not lazy-loaded, and ideally marked as high priority. If you have a slider, replace it with one strong image.

When is it actually the hosting?

After working through the list, look at the “Reduce initial server response time” diagnostic in PageSpeed Insights, or check your Time to First Byte. If the server still takes a long time to respond even with caching on, the host is likely the bottleneck. We cover the signs in is your hosting slowing your website down?

Impact ratings are general guidance from our experience, not guarantees; every site is different.

Frequently asked questions

Why is my WordPress site so slow?

The most common causes are oversized images, no caching, too many or poorly built plugins, a heavy theme or page builder, outdated PHP and too many third-party scripts.

How many plugins is too many?

There’s no magic number. Quality matters more than quantity, but every plugin you don’t need is a risk to speed and security.

Will a caching plugin fix my speed problems?

It helps a lot with server response time, but it can’t fix oversized images, heavy JavaScript or a bloated theme on its own.

Should I change hosting to speed up WordPress?

Only after fixing on-site issues. If server response is still slow with caching enabled, better hosting is worth considering.

The post 10 WordPress Speed Fixes to Try Before You Change Hosts appeared first on Outview.

Every business pauses ads at some point. A quiet month, a cash-flow squeeze, a change of agency, or simply wanting to see what happens. It’s a reasonable thing to do. It’s also revealing.

Here’s what typically changes when the ads go dark, what stays the same, and how to set things up so a pause doesn’t feel like a cliff.

What happens immediately

Paid clicks stop. Within hours of pausing a campaign, your ads stop showing. Every visitor who came only through those ads stops arriving.

Leads from those clicks stop. If 70% of your enquiries came through ads, expect a drop close to that, fairly quickly. Some leads lag (people who clicked last week may still phone today), so the full effect usually shows within a few weeks.

Your share of attention shrinks. Competitors who keep advertising pick up the searches you were catching. That’s not a penalty; it’s just an empty slot being filled.

What doesn’t change

Your organic rankings. Google is clear that running or pausing ads doesn’t affect organic search rankings. If you ranked third organically for “bookkeeper Claremont” before, you’ll rank about the same after, unless something unrelated changes.

Your Google Business Profile. Map listings, reviews and profile visibility carry on. For local businesses, this is often a bigger lead source than people realise.

Your account history. Paused campaigns keep their settings and historical data, so you can restart later. Automated bidding may need a short period to recalibrate after a long break, so the first week or two back can be a little less efficient.

What fades more slowly

Retargeting audiences. Lists of past website visitors shrink as people drop off after the list’s membership period. A long pause means smaller lists when you restart, unless organic traffic keeps topping them up.

Brand awareness. Ads don’t just drive clicks; they put your name in front of people repeatedly. Some later search for you by name or walk in. That effect fades gradually, and it’s hard to measure precisely.

The real question a pause answers

Pausing ads shows you, very clearly, how strong your foundation is.

Picture two hair salons in Durbanville, spending the same on ads. One has a fast website, hundreds of genuine Google reviews and helpful blog posts about hair care. The other has a slow five-page site and a handful of reviews. Both pause their ads for January.

The first salon notices a dip. The second salon’s phone goes quiet.

That’s the difference between rented and owned traffic in real life.

How to see your dependence before you pause

In Google Analytics, look at where your conversions come from, by channel or source. Roughly:

These aren’t official benchmarks, just a practical way to think about risk. If you don’t have conversion tracking set up, that’s step one.

How to prepare for a pause

Strengthen your Google Business Profile. Complete every section, add real photos, post updates and ask happy customers for reviews. Our Google Business Profile service exists because this channel is so often under-used.

Make sure your site is fast and clear. Organic visitors need a good experience too. Check it with our foundation checker.

Publish content around your best-converting keywords. Your ad data shows which searches turn into customers. Create genuinely helpful pages for those topics so you can rank for them organically.

Build your email list. An email list is the most “owned” audience there is. A quiet month is a good time to send a useful newsletter or offer.

Reduce, don’t stop. Instead of switching everything off, keep a small budget on your highest-converting terms and your brand name.

Should you pause your ads?

Sometimes it’s the right call: seasonal lulls, a broken website that’s wasting clicks, or a campaign that simply isn’t profitable. Pausing isn’t a failure.

But if the thought of pausing makes you nervous, that’s useful information. It means your growth rests on something you rent rather than something you own. The fix isn’t to never pause; it’s to build a foundation that keeps working when the ads don’t. That’s what our Growth Package is designed to do.

Frequently asked questions

Will pausing Google Ads hurt my SEO?

No. Google doesn’t use ad spend as an organic ranking factor, so pausing ads doesn’t directly affect organic rankings.

Do I lose my Quality Score if I pause ads?

Your account and campaign history remain. Performance can take a short while to settle when you restart, especially with automated bidding.

How quickly do leads drop after pausing ads?

Paid clicks stop within hours. Leads from those clicks usually tail off over a few days to a few weeks as delayed enquiries come through.

Is it better to pause or reduce ad spend?

For most businesses, reducing spend to the best-performing campaigns is safer than switching everything off.

The post What Happens to Your Leads When You Pause Your Ads? appeared first on Outview.

Open any PageSpeed Insights report and you’ll see two sections that look like they’re measuring the same thing. Sometimes they agree. Often they don’t. And for many small business sites, the top section just says “No Data”.

Here’s what’s going on, and which numbers deserve your attention.

What is field data?

Field data comes from the Chrome User Experience Report (often called CrUX). Chrome collects anonymous performance data from real users who have opted in to sharing usage statistics. Google aggregates it over a rolling 28-day window.

In PageSpeed Insights, it appears at the top under “Discover what your real users are experiencing”. It shows the Core Web Vitals (LCP, INP and CLS) plus a few supporting metrics, and a simple verdict: the Core Web Vitals assessment passed or failed.

Each metric is reported at the 75th percentile. That means three out of four real visits were at least that good. It’s a deliberately tough standard: your fastest visitors don’t get to hide your slowest ones.

What is lab data?

Lab data comes from Lighthouse, which loads your page once in a controlled environment: an emulated device, a set network speed, a fixed processor speed. It appears in the lower section, “Diagnose performance issues”, with the four familiar scores.

Because it’s a controlled test, it’s repeatable and it comes with diagnostics: a specific list of what’s slowing the page down. That makes it the best tool for fixing things.

Lab data vs field data: the key differences

Why does PageSpeed Insights say “No Data”?

Google only shows field data when there are enough real visits from Chrome users who share usage statistics. If a page doesn’t meet that threshold, there’s simply nothing reliable to show.

That’s common for:

• Smaller local businesses
• New websites and new pages
• Individual blog posts and deep service pages
• Sites with mostly desktop or non-Chrome visitors

Sometimes PageSpeed Insights falls back to “origin” data, which is an average across your whole website instead of the specific page. If you see that, remember it’s site-wide, not page-specific.

“No Data” isn’t a penalty and it isn’t a sign something is wrong. Our own homepage showed “No Data” for field data on the day we tested it, alongside perfect 100 lab scores.

Why don’t my lab and field numbers match?

Because they’re measuring different things. A few common reasons for the gap:

Your real visitors have different devices. If most of your customers use older phones on mobile data, field data can be worse than the lab test. If they’re mostly office workers on fast connections, it can be better.

Timing. Field data covers the last 28 days. If you fixed something yesterday, the lab test shows the improvement immediately; field data catches up gradually.

Interactions after loading. Field data includes things that happen while people use the page, like layout shifts when a pop-up appears or slow responses when someone taps a menu. The lab test mostly measures the loading phase.

Caching. Returning visitors often load pages faster thanks to their browser cache. Lab tests simulate a fresh visit.

Which one should you trust?

Both, for different jobs.

Use field data to answer: “How are real customers experiencing my site, and am I passing Google’s Core Web Vitals?”

Use lab data to answer: “What’s slowing this page down, and did my fix work?”

If you have no field data, lab data is your best guide. Aim for good lab scores on mobile and you’ll usually be in good shape when field data does appear.

Where else can you see field data?

Google Search Console has a Core Web Vitals report that groups your pages into good, needs improvement and poor, based on the same real-user data. It’s useful because it shows patterns across many pages at once. We cover it in our Search Console beginner’s guide.

Common mistakes

Panicking about “No Data”. It just means not enough Chrome traffic yet.

Ignoring field data when you have it. If real users are failing Core Web Vitals, that matters more than a good lab score.

Expecting field data to update instantly. Give it a few weeks after a fix.

Only looking at the score. The real value of the lab test is in the diagnostics underneath. Our guide to reading the full report explains how to use them.

Frequently asked questions

What does “No Data” mean in PageSpeed Insights?

Your page doesn’t have enough real visits from Chrome users for Google to report field data. The lab test below still works.

Does Google use lab data or field data for rankings?

Google’s page experience signals use real-user Core Web Vitals, which is field data. Lab data is a diagnostic tool.

How long does field data take to update?

It covers a rolling 28-day period, so improvements show up gradually over several weeks.

Why is my field data worse than my lab data?

Your real visitors may use slower phones or networks than the lab test assumes, or problems may occur after the page loads, such as pop-ups or slow interactions.

The post Lab Data vs Field Data: What “No Data” in PageSpeed Insights Means appeared first on Outview.

It’s one of the first things people notice: desktop says 92, mobile says 58. Same website, same page, same moment. Is something broken?

Usually not. The two tests are measuring very different situations, on purpose.

How does PageSpeed Insights test mobile vs desktop?

PageSpeed Insights runs Google’s Lighthouse tool twice, with different settings:

So the mobile test is essentially asking: “What happens when an ordinary person on an ordinary phone, on mobile data, opens this page?” That’s a much harder exam.

Why the slowed-down processor matters so much

Most people assume the slower network is the main reason. It’s a big part, but the processor is often the bigger one.

Modern websites run a lot of JavaScript: page builders, sliders, tracking tags, chat widgets. A powerful laptop chews through it in moments. A mid-range phone takes several times longer. That’s why Total Blocking Time and responsiveness often look fine on desktop and poor on mobile.

Heavy images hurt too. A 2 MB hero image downloads quickly on fibre. On a throttled mobile connection, it can add seconds to your Largest Contentful Paint.

How big a gap is normal?

There’s no official rule, but in our experience:

• A gap of 10 to 30 points is very common, even on well-built sites.
• A gap of 40+ points usually means the page leans heavily on JavaScript or large images.
• A small gap with both scores high is a sign of a genuinely lean site. It’s achievable: our own homepage scores 100 on both.

Which score should you care about?

Mobile. Almost always.

First, Google uses mobile-first indexing, which means it primarily uses the mobile version of your site for indexing and ranking.

Second, that’s where your customers are. Across South Africa, a large share of people browse mainly on their phones, often on prepaid data and patchy signal. Think of someone in a taxi on the N2 searching for a nearby panel beater. They’re not on fibre.

Desktop still matters for some audiences, particularly B2B buyers researching at work. But if you can only improve one, improve mobile. Desktop usually improves along the way.

How to close the gap

Shrink images for mobile. Serve smaller image sizes to smaller screens and use modern formats like WebP or AVIF. See our image optimisation guide.

Cut JavaScript. Remove unused plugins, delay non-essential scripts, and question every third-party widget.

Simplify the mobile layout. Keep everything important, but drop decorative extras: autoplay videos, big sliders, animation-heavy sections.

Improve hosting and caching. Server delays hurt both scores, but they compound on mobile.

Test on a real phone. Turn off Wi-Fi and load your site on mobile data. If it feels slow to you, it feels slow to customers.

Common misunderstandings

“Google is penalising my mobile site.” The score itself isn’t a penalty; it’s a diagnostic. For rankings, what counts is real-user Core Web Vitals and, above all, content relevance.

“My phone loads it fine.” You probably have a newer phone, good signal, and the site cached from previous visits. Your customers may not.

“A responsive theme means mobile is handled.” Responsive means the layout adapts to the screen. It doesn’t mean the page is light enough for a phone.

For a walkthrough of every number in the report, see how to read a PageSpeed Insights report.

When to get help

If your mobile score stays in the red after compressing images and removing plugins, the problem is often structural: a heavy theme or page builder. That’s when a performance-focused rebuild becomes the better investment, because every marketing channel you pay for sends people to that mobile experience.

Frequently asked questions

Why is my mobile PageSpeed score so low?

The mobile test simulates a mid-range phone with a slower processor on a throttled connection, so heavy images and JavaScript hurt much more than on desktop.

Is a mobile score of 70 good?

It’s in the “needs improvement” range, but a reasonable result for many business sites. Aim for 90+, and treat anything under 50 as urgent.

Does Google rank sites based on mobile or desktop?

Google uses mobile-first indexing, so the mobile version of your site is primarily used for indexing and ranking.

Can a site score 100 on mobile?

Yes. It takes a lean build, optimised images and minimal JavaScript, but it’s achievable.

The post Why Is My Mobile PageSpeed Score Lower Than Desktop? appeared first on Outview.

For about two years, business owners have been told that AI search needs a whole new playbook. New file formats. New ways of writing. New agency services with three-letter names.

Now Google has put its position in writing. Its Search Central documentation includes a guide called Optimizing for AI features in Google Search, and a large part of it is a list of things you can stop worrying about.

We offer an AI Visibility (GEO) service, so you might expect us to downplay that. We’d rather do the opposite. We read the guide line by line, compared it with what we’d published on our own blog, and corrected the places where we’d overstated things. This article is the plain-English version of what we found, and what it means if you run a business in Cape Town, Johannesburg, Austin or anywhere else.

What does Google’s AI optimisation guide actually say?

The guide covers Google’s AI features in Search, mainly AI Overviews and AI Mode. Its core argument is that these features are built on the same foundations as normal search results. If your pages are in Google’s index and can appear with a snippet, they can be used in an AI answer. If they can’t, no amount of “AI optimisation” will change that.

Google frames it in a way that’s worth repeating word for word: optimising for generative AI search is still SEO. The guide tells site owners to focus on what visitors would enjoy, find helpful and feel satisfied with.

That sounds obvious. It isn’t how much of the industry has been talking.

The myths Google has now addressed directly

Here’s what the guide says you don’t need to do, with our honest take on each.

1. You don’t need an llms.txt file

llms.txt is a proposed text file that summarises a website for large language models. It spread quickly in SEO circles, and some providers started charging to set one up. Google’s guide says Google Search doesn’t use it. It won’t help or hurt your rankings.

Could another AI company read one someday? Possibly. But if someone is selling it to you as a way into Google’s AI Overviews, that’s not what Google says.

2. You don’t need to “chunk” your content

A popular tactic has been to break pages into short, self-contained fragments so AI systems can supposedly lift answers more easily. Google says its systems can understand the nuance of multiple topics on a page. You don’t need to chop a good guide into disconnected pieces.

This doesn’t mean structure is useless. Clear headings, a short summary at the top and readable paragraphs help people. They just aren’t a secret AI lever.

3. You don’t need to rewrite your copy for machines

Google’s systems understand synonyms and general meaning. Writing awkward, keyword-heavy sentences “so the AI understands” tends to make pages worse for the people reading them, and Google says it doesn’t need it.

4. You don’t need a page for every way a question can be asked

Imagine a plumbing business in Durbanville that publishes separate pages for “geyser burst what to do”, “burst geyser help”, “geyser leaking emergency” and so on. That’s the pattern Google warns against. The guide says not to create separate content for every variation of a query, and that a high number of pages doesn’t make a site higher quality. One thorough, genuinely useful page on burst geysers will do more.

5. Structured data isn’t required for AI features

Schema markup has been sold as the foundation of AI visibility. Google’s guide says structured data isn’t required for generative AI search.

We still add schema for clients, because it removes doubt about business details and can make pages eligible for rich results. We’ve written about that in our schema markup guide for business owners. But it’s a supporting detail, not the reason anyone gets cited.

6. Inauthentic mentions don’t work

Paying for mentions on low-quality sites, seeding your name in forums, or spinning up fake “best of” lists won’t get you into AI answers, according to Google. Genuine reviews and real coverage are a different thing entirely, and still worth earning.

What Google says you should do instead

Not one of those rows is new. That’s rather the point.

Why “non-commodity content” is the part to take seriously

If there’s one phrase in the guide worth sticking on the wall, it’s this idea of content that goes beyond common knowledge.

AI Overviews summarise. If your page says exactly what twenty other pages say, there’s very little reason for Google to point anyone to you specifically. Generic “10 tips for choosing an accountant” articles were already struggling. They’ll struggle more now.

Compare two pages from a Cape Town accounting firm.

The first explains what a tax practitioner does, in general terms. It’s accurate, and it’s on thousands of other websites.

The second explains what the firm actually sees when small businesses miss provisional tax deadlines, what SARS correspondence typically looks like afterwards, and how they handle it, with a realistic fee range. Nobody else can write that page, because nobody else has that experience.

The second page is what Google is describing. It’s also the page a prospective client is more likely to trust.

Does this mean GEO is a scam?

No, but it does mean some of what’s sold under that name is.

Here’s how we’d split it. For Google’s own AI features, Google is clear: do good SEO. There’s no hidden layer. Anyone selling a separate “Google AI Overviews optimisation” package built on llms.txt files, chunking and schema bundles is selling things Google has said it doesn’t use or need.

ChatGPT, Perplexity, Claude and other assistants are different companies with their own systems. Google’s guide doesn’t govern them, and none of them publishes a full rulebook. What we can say from testing is that the same foundations tend to carry over: clear information about what you do and where, consistent business details across the web, and genuine reviews and coverage.

So the useful part of AI visibility work isn’t a set of tricks. It’s two things. First, solid SEO, done properly. Second, regularly checking what these AI tools actually say about your business, and fixing the inaccurate or missing information at the source. That second part is what ranking reports don’t show you, and it’s where our AI Visibility service spends most of its time.

What we changed on our own site after reading it

We’d be in no position to write this if we hadn’t checked our own work first.

Going back through our blog, we found a few places where we’d leaned too hard on the “structure your content for AI extraction” idea, and where we’d described schema as central to AI visibility. Those weren’t outrageous claims at the time, but they don’t match Google’s published position. We’ve updated them, including our articles on staying visible as AI Overviews take clicks and getting mentioned by ChatGPT.

We’d suggest every business do the same with its own site and with any agency proposal on the table. If a plan leans on things Google has explicitly ruled out, ask why.

A practical checklist for the next 30 days

1. Confirm your important pages are indexed. In Google Search Console, inspect your homepage, main service pages and top blog posts. If one isn’t indexed, nothing else on this list matters for that page.
2. Look at the generative AI performance report. It sits under Performance in Search Console and shows how often your pages appeared in AI Overviews and AI Mode, by page, country and device. It counts impressions only, not clicks, and newer or smaller sites may see no data until they pass Google’s threshold.
3. Find your commodity pages. Pick the five pages that could have been written by anyone. Add what only you know: prices, process, local detail, mistakes you see customers make.
4. Merge near-duplicates. If you have several thin pages chasing variations of one question, combine them into one strong page and redirect the rest.
5. Fix your Google Business Profile. Hours, categories, services, photos. Our Google Business Profile optimisation service covers this if you’d rather hand it over.
6. Check speed on mobile. Run your site through our free foundation checker and fix the biggest problem first.
7. Ask the AI tools about yourself. Put your customers’ real questions to Google, ChatGPT, Gemini and Perplexity. Write down what’s wrong or missing, then correct it on your site and profiles.
8. Cancel anything built on myths. llms.txt setup fees, chunking rewrites, mass page generation. Google says these don’t help in its search.

Common mistakes we expect to keep seeing

Treating AI search as separate from SEO. Businesses that split budgets between “SEO” and “GEO” often end up paying twice for overlapping work, with neither done well.

Publishing more instead of better. AI writing tools make it cheap to publish hundreds of pages. Google allows AI-assisted content as long as it meets its spam policies and is genuinely helpful, but volume on its own does nothing for quality.

Ignoring the basics because they’re boring. A noindex tag left over from a website rebuild can quietly remove a page from both normal results and AI Overviews. We see this more often than you’d think.

Blocking Google by accident. Security plugins, aggressive bot rules and misconfigured robots.txt files can stop Googlebot reaching pages. If Google can’t crawl it, it can’t use it.

Where this leaves South African and US businesses

For local businesses in South Africa, the guide is good news. Many SA sites still have basic problems: slow mobile pages, thin service descriptions, unclaimed or outdated Business Profiles. Fixing those is affordable, and it now counts for AI features as well as standard results.

In the US, competition for AI visibility is fiercer and the temptation to buy shortcuts is stronger. The guide gives you a clear filter for agency pitches: if a tactic depends on something Google says it doesn’t use, it isn’t a strategy for Google.

Either way, the businesses that come out ahead will be the ones whose websites are actually the most useful in their field. That was true before AI Overviews. It’s still true.

Frequently asked questions

Does Google use llms.txt?

No. Google’s AI optimisation guide says Google Search doesn’t use llms.txt or similar AI text files, and they don’t help or hurt rankings.

Do I need schema markup to appear in AI Overviews?

No. Google says structured data isn’t required for its generative AI features. It’s still useful for rich results and for stating business details clearly.

Is GEO different from SEO?

For Google’s AI features, Google says it’s the same thing: SEO. Other AI assistants run their own systems, so it’s still worth checking what they say about your business, but the foundations are shared.

Should I break my content into smaller chunks for AI?

No. Google says its systems understand pages that cover several topics. Write and structure pages for readers instead.

How do I see if my site appears in AI Overviews?

Google recommends the generative AI performance report in Search Console. You can also search your customers’ real questions and see which sources are shown.

Can AI-written content rank in AI Overviews?

Google allows content made with the help of AI tools if it meets Search Essentials and spam policies and is genuinely helpful. Mass-produced, low-value pages won’t help.

The post Google’s AI Search Guide: What Business Owners Should Actually Do appeared first on Outview.

Your page loaded quickly. Great. Now someone taps the menu button. Nothing happens. They tap again. Half a second later the menu opens, then closes, because it registered both taps.

That’s a responsiveness problem, and it’s measured by a metric most business owners have never heard of: Interaction to Next Paint.

What is Interaction to Next Paint?

INP looks at the interactions a visitor has with a page (taps, clicks and key presses) and measures how long it takes for the screen to visibly update after each one. It then reports a value close to the slowest interaction during the visit.

In other words, it’s not about how fast the page loads. It’s about how fast the page reacts once it’s there.

INP became one of Google’s three Core Web Vitals in March 2024, replacing First Input Delay, which only measured the very first interaction. INP is stricter because it considers the whole visit.

What is a good INP score?

Why doesn’t INP appear in my lab test?

Good question, and it confuses a lot of people. INP needs a real person interacting with the page, so it can only be measured from real visitors. It appears in the field data section at the top of PageSpeed Insights, if your site has enough traffic.

In the lab test, the closest stand-in is Total Blocking Time (TBT). TBT measures how long the page is too busy running code to respond during loading. High TBT is a strong hint that INP will be poor too. We explain the difference between these two kinds of data in lab data vs field data.

Why does responsiveness matter?

Because the moments when people interact are usually the moments that make you money. Opening the menu to find “Services”. Tapping “Add to cart”. Typing into an enquiry form. Selecting a date in a booking widget.

Imagine a customer booking a table at a restaurant in Stellenbosch. They tap a time slot and nothing seems to happen for a second. They tap again, accidentally selecting a different slot. Frustration, then a phone call, or a booking somewhere else.

Laggy interactions make a site feel broken even if it loaded quickly. And because INP is a Core Web Vital, it’s part of how Google assesses page experience.

What causes poor INP?

Your browser runs most of a page’s code on a single main thread. While it’s busy with a long task, it can’t respond to taps. Common reasons it gets busy:

• Too many plugins and page builder features, each loading its own scripts.
• Third-party scripts: chat widgets, heatmaps, social embeds, multiple analytics and ad tags.
• Heavy event handlers: code that does a lot of work every time someone clicks, such as filtering a large product list.
• Very large pages with thousands of elements, which take longer to update.
• Older, cheaper phones, which magnify every one of these problems. Many of your visitors use them.

How to improve INP

Audit your scripts. List every plugin, tag and widget on the site and ask, “Does this earn its place?” Removing two or three unnecessary scripts often makes a visible difference.

Load third-party tools later. Chat widgets and tracking tags can usually load after the page is interactive, or only when needed.

Use Google Tag Manager carefully. It’s convenient, but it makes it easy to pile up tags. Clean out old ones.

Simplify interactive features. Mega-menus with animations, complex sliders and live filters are frequent offenders.

Ask your developer about long tasks. Breaking heavy code into smaller chunks, so the browser can respond in between, is a standard fix.

Choose lightweight tools. When we built the Outview AI Chatbot, keeping it light on the page was a deliberate design goal, because a chat widget that makes the whole site laggy defeats the purpose.

Common mistakes

Only looking at the lab score. A green lab Performance score doesn’t guarantee good INP. Check the field data and TBT.

Testing on a fast laptop. Try your site on a mid-range Android phone. That’s the reality for a lot of customers.

Adding “one more” tool. Every widget feels small. Ten of them together are not.

When to get help

Removing plugins and old tags is something you can start today. Code-level fixes need a developer. If your site was assembled from many heavy plugins over the years, a cleaner rebuild is sometimes cheaper than patching. It’s one of the foundation checks we run in a website review.

Frequently asked questions

What does INP stand for?

Interaction to Next Paint, a Core Web Vitals metric for how quickly a page responds to user interactions.

What is a good INP score?

200 milliseconds or less is good, 200 to 500 milliseconds needs improvement, and above 500 milliseconds is poor.

What replaced First Input Delay?

INP replaced First Input Delay as a Core Web Vital in March 2024.

Why can’t I see INP in my PageSpeed lab results?

INP requires real user interactions, so it only appears in field data. Total Blocking Time in the lab test is the closest indicator.

Do chat widgets hurt INP?

Heavy ones can. Choose lightweight tools and load them after the main content is interactive.

The post Interaction to Next Paint (INP): The Responsiveness Metric Most Sites Ignore appeared first on Outview.

When Google Ads get expensive, most people’s first instinct is to look at the ads themselves: the bids, the keywords, the copy. Fair enough. But some of the most expensive problems live on the page people land on after they click.

Here’s how your website quietly affects what you pay, and what you can do about it.

How Google decides what you pay

Google Ads is an auction, but not a simple highest-bidder-wins auction. Google wants to show ads that searchers find useful, because that keeps people using Google. So every time your ad is eligible to show, Google weighs your bid against the quality of your ad and the page behind it.

That quality assessment is summarised for you as a Quality Score from 1 to 10 for each keyword. Google describes three components:

The practical effect: advertisers with better quality can often win better positions without bidding as much. Advertisers with poor quality pay more to compete, or don’t show at all.

Where speed comes into it

Google’s guidance on landing page experience mentions relevant, original content, clear navigation, transparency about your business, and a page that loads quickly and works well on mobile.

Speed isn’t the only factor, but it’s one you can measure and fix. And a page that’s slow to load on mobile tends to fail on several other fronts too: it’s frustrating to navigate, it shifts around, it hides the call to action below a half-loaded hero image.

In your Google Ads account, you can see each keyword’s landing page experience rated “Above average”, “Average” or “Below average”. If you see “Below average” across the board, your website deserves a hard look.

The hidden cost: visitors who leave before the page loads

Even if Quality Score didn’t exist, a slow page would cost you money.

You pay for the click the moment someone taps your ad. If your page takes six seconds to show anything useful on a mobile connection, a chunk of those paid visitors will give up and go back to the search results. You’ve paid for the click and got nothing.

Let’s walk through a simple illustration with made-up round numbers. Say you pay R20 per click and 5% of visitors become enquiries. That’s R400 per lead. If a faster page means 7% convert instead, the same clicks now cost about R286 per lead. You haven’t touched your bids or your ads. You’ve just stopped losing people at the door.

Your real numbers will be different, but the principle holds: anything that improves the landing page improves every rand you spend on clicks.

Signs your website is inflating your ad costs

• Landing page experience shows “Below average” in Google Ads.
• Lots of clicks, few enquiries, especially from mobile.
• High bounce rate or very short sessions from paid traffic in Google Analytics.
• Mobile PageSpeed score in the red or orange.
• Ads send people to your homepage instead of a page that matches the ad.

You can check the speed side in a minute with our free foundation checker. Test the exact page your ads link to, on mobile.

How to fix it

Start with mobile speed

Focus on Largest Contentful Paint, usually the hero image, and on removing heavy scripts. Compress images, cut unused plugins, and make sure the page is cached.

Build dedicated landing pages

A page built for one offer, with one clear action, is almost always faster and more relevant than a general homepage. If the ad says “Same-day aircon repairs in Somerset West”, the page should say exactly that above the fold.

Strip out distractions

Sliders, auto-playing videos and pop-ups slow pages down and pull attention away from the enquiry form.

Keep the action visible

Phone button, WhatsApp link or short form near the top, without the page jumping around as it loads. Our guide to layout shift covers why that matters.

Make sure tracking works

If conversions aren’t tracked, Google’s bidding can’t learn and you can’t see which pages are underperforming.

A note of honesty

A faster page won’t rescue a bad offer or poorly chosen keywords. And Quality Score is Google’s estimate, not something you can game. But in our experience, landing page fixes are among the most reliable improvements to ad performance, because they help on every click, from every keyword, every day.

This is exactly why we see the website as the foundation and ads as something you build on top. We explain the bigger picture in paid ads vs SEO.

How we approach it at Outview

When we take on Google Ads management, one of the first things we check is the landing page, not the bids. Where pages are slow or unfocused, we fix or rebuild them through our web design service before scaling spend.

Frequently asked questions

Does page speed affect Google Ads Quality Score?

Speed and mobile usability are part of what Google considers in landing page experience, which is one of the three Quality Score components.

How do I check my landing page experience in Google Ads?

Add the Quality Score and landing page experience columns to your keyword report. Each keyword is rated above average, average or below average.

Will a faster website lower my cost per click?

It can help, because better quality can let you compete without bidding as much. The more reliable gain is a lower cost per lead, since fewer paid visitors leave before converting.

Should my ads link to my homepage?

Usually not. A dedicated page that matches the ad’s message tends to be faster, more relevant and better at converting.

The post Why a Slow Website Makes Your Google Ads Cost More appeared first on Outview.

You’re reading an article on your phone. You reach for the “Next” button. Just as your thumb lands, an ad loads above it, everything slides down, and you’ve tapped something else entirely.

Annoying for a news site. On a business website, it can mean someone accidentally leaves, mis-taps a call button, or simply decides the site feels broken. That jumpiness has a name and a score: Cumulative Layout Shift.

What is Cumulative Layout Shift?

CLS is one of Google’s three Core Web Vitals. It measures visual stability: how much visible content moves without the user doing anything to cause it.

Each unexpected movement gets a score based on how much of the screen moved and how far. Google then takes the worst burst of shifts during a visit as your CLS. Movements that happen right after someone taps or clicks (like an accordion opening) don’t count, because the user expected them.

What is a good CLS score?

Unlike loading metrics, CLS isn’t measured in seconds. It’s a unitless number. A score of 0 means nothing moved at all, which is entirely achievable. Our own homepage scores 0 on both mobile and desktop.

Why does layout shift matter?

Visitors rarely complain about layout shift. They just trust the site a little less. A page that twitches while loading feels cheap, unfinished or even suspicious, especially when someone is about to enter card details or contact information.

It also causes real mistakes. Imagine a customer on a car dealership’s site in Montague Gardens trying to tap “Book a test drive”, and a late-loading finance banner shoves the button down so they hit “Back to listings” instead. They might not try again.

And like LCP, CLS feeds into Google’s page experience signals. It won’t outrank great content, but it’s part of how Google judges whether a page is pleasant to use.

What causes layout shift?

Images and videos without dimensions

If the browser doesn’t know how big an image will be, it reserves no space. When the image arrives, everything below gets pushed down. This is the most common cause on WordPress sites, especially with older themes or images inserted by page builders.

Ads, embeds and iframes

Google Maps embeds, YouTube videos, social feeds and ad slots often load late and expand. Without a reserved container, they shove content aside.

Content injected at the top

Cookie banners, promo bars (“Free delivery this weekend!”), newsletter sign-ups and chat prompts that appear above existing content push everything down.

Web fonts swapping in

The page first shows text in a fallback font, then switches to your custom font. If the two fonts are different sizes, lines reflow and content jumps.

Animations that move layout

Animations that change an element’s size or position in the layout (rather than using transforms) can register as shifts.

How to fix CLS

Set width and height on every image and video. Modern browsers use these to reserve the right space even on responsive layouts. Most modern WordPress themes do this automatically; older themes and some page builders don’t.

Reserve space for embeds and ads. Wrap maps, videos and ad slots in a container with a fixed minimum height or aspect ratio.

Overlay banners rather than insert them. Cookie notices and promo messages should float over content (for example, fixed to the bottom of the screen), not push it down.

Handle fonts carefully. Preload your main font, and choose a fallback font with similar sizing. Some font services offer size-adjusted fallbacks for exactly this reason.

Animate with transforms. Ask your developer to use CSS transform and opacity for movement, which don’t shift layout.

How to find what’s shifting

Run your page through our foundation checker or PageSpeed Insights. In the full Google report, the diagnostic “Avoid large layout shifts” lists the elements that moved. That’s your fix list.

You can also simply watch. Load your site on your phone over mobile data (not Wi-Fi), and pay attention to the first five seconds. Anything that jumps is a candidate.

Remember that lab tests only catch shifts during loading. Shifts that happen later, like a pop-up after 20 seconds, show up in real-user field data. Our article on lab data vs field data explains why your numbers can differ.

Common mistakes

Adding a “sticky” promo bar at the top of every page. If it loads after the content, it pushes everything down on every visit.

Embedding several social feeds on the homepage. They’re slow, they shift, and few people interact with them.

Assuming the theme handles it. Many do, but page builders and plugins can undo good theme behaviour.

When to get help

Setting image dimensions and moving a cookie banner are usually quick wins. Font loading and theme-level fixes may need a developer. If CLS problems come from the way the site was built, it’s often part of a broader website rebuild conversation, where we’d make stability part of the foundation from day one.

Frequently asked questions

What does CLS stand for?

Cumulative Layout Shift, a Core Web Vitals metric that measures how much page content unexpectedly moves.

What is a good CLS score?

0.1 or less is good, between 0.1 and 0.25 needs improvement, and above 0.25 is poor.

Does a cookie banner cause layout shift?

It can, if it’s inserted at the top and pushes content down. Banners that overlay the page, for example fixed to the bottom, avoid the problem.

Why is my CLS different in PageSpeed Insights field and lab data?

The lab test only measures shifts during page load. Real-user data includes shifts that happen later while people scroll and interact.

Keep reading

• Mobile-First Design: What It Actually Means for Your Website
• Website Accessibility: Why It Matters for SEO and Who You’re Excluding Without It
• How to Choose a Website Platform: WordPress vs Wix vs Squarespace

The post Cumulative Layout Shift (CLS): Why Buttons Jump and How to Fix It appeared first on Outview.