Landing page speed costs money twice in paid search. Visitors leave before the page renders, so you paid for a click that produced nothing. And page experience feeds into ad ranking, so a slow page can raise what you pay for every subsequent click as well.
The second cost is the one people miss, and it compounds quietly across an account.
Measure on the connection your visitors actually have
A page tested on office fibre from a desktop is not the page your customers experience. Most paid traffic for local services arrives on a phone, often on mobile data, sometimes with a weak signal. Test under throttled mobile conditions and the numbers usually look considerably worse, which is the point.
The metric worth caring about is when the page becomes useful: when the main content is visible and the primary action is tappable. Total load time includes a lot that happens after the visitor has already decided.
The four things that usually explain it
- Uncompressed images, especially a large hero shot served at full resolution to a phone.
- Third-party scripts. Chat widgets, heat maps, multiple analytics tags and social embeds, each adding delay.
- Web fonts loaded before text renders, which produces a page that looks empty while it is technically working.
- A page builder loading its whole framework for a layout that could have been three blocks.
That list covers the overwhelming majority of slow landing pages. Deeper technical work matters at scale and rarely explains an obviously sluggish page.
The biggest speed win in most accounts is deleting a tracking script somebody added for a two-week experiment three years ago and never removed.
Aisha Karim, Analytics & CRO Manager, Media @ Marsons
Auditing your own tags
List every script on the landing page and ask what each one is for and who reads its output. In most businesses several have no owner and no consumer, and removing them is free performance with no downside. The remainder can usually be deferred so they load after the content rather than in front of it.
Speed is a ceiling, not a strategy
A fast page that does not answer the query still fails. Speed determines how many visitors get the chance to read your offer; the offer determines what happens next. Fixing speed on a page with a weak headline and a buried phone number produces a faster version of the same disappointing result.
Where to spend the effort
Compress and correctly size images first, because it is the largest gain for the least risk. Remove unused scripts second. Defer what remains third. Only then consider anything involving templates or hosting, which carries more risk and usually returns less than the first three combined.
If your ads point at a general homepage rather than a page written for the campaign, that is a bigger problem than any of this. The PPC hub covers how the pages behind ads differ by sector.
Re-measure after the change, on a real device
Synthetic scores move in ways that do not always match what visitors experience. After making changes, load the page on an ordinary phone away from your office wifi and time how long it takes before you could tap the call button. That number is the one that correlates with your conversion rate.
Where speed stops being the bottleneck
There is a point past which further optimisation returns nothing. A page that becomes usable in under two seconds on mobile data is not losing visitors to load time, and pushing it lower is engineering effort spent where no customer will notice. The next constraint is almost always the offer, the clarity of the next step, or what happens after someone gets in touch.
Knowing where that line sits prevents the common pattern where a team spends a quarter on performance work and reports a faster page with an unchanged conversion rate.
Mobile and desktop are different problems
Desktop scores flatter almost every site, because the connection is better and the screen shows more before anyone scrolls. If most of your paid traffic is mobile, a desktop score is close to irrelevant, and the fix list changes: image weight and script count dominate on a phone, while layout shift matters far more because a mistapped button on a moving page is a lost enquiry.
Layout shift is worth singling out. A page where the content jumps as images and fonts arrive will lose people who tapped the wrong thing, and it registers as a bounce rather than as the interface fault it actually is.
Paying for clicks that never see the page?
We will measure your landing pages on a real mobile connection and tell you what is costing conversions and what is safe to ignore.
