The Marketing Funnel Is Dead. Here’s What Replaced It.
2026-02-02
March 2024. I walk into a client meeting expecting the usual monthly review. Instead the CEO is staring at a screen full of red numbers. Organic traffic down 40 percent. Revenue down 35 percent. The Google Search Console report says “Poor” in bold red letters next to “Core Web Vitals” for 73 percent of their pages. They have been paying an agency $5,000 a month for eighteen months. The agency’s response to the crash: “Google algorithm update. Nothing we can do about it. It will probably recover in a few months.”
I was not so sure about that “nothing we can do” part. I had seen this pattern before with other clients. A sudden traffic drop blamed on an algorithm update that turned out to be a technical problem that had been building for months. Algorithm updates get blamed for a lot of things that are not the algorithm’s fault. I asked for access to the site and started digging.
The Diagnosis
I opened Chrome DevTools and ran a Lighthouse audit. It took about thirty seconds and painted an immediate picture. The Largest Contentful Paint — which is Google’s fancy term for “how long until the main thing on the page actually shows up” — was 6.2 seconds. Google’s threshold for a passing grade is 2.5 seconds. They were not just over the limit. They were more than double it. On mobile it was even worse.
I dug into the specific causes and found three main problems, each one worse than the last.
The hero image. Some designer along the way had set the homepage hero image at 2400 pixels wide at full JPEG quality. The file size was 2.4 megabytes. That is not an image designed for the web. That is a print file. On a typical 4G mobile connection, that image alone takes about three seconds to download before the browser can show anything below it. Nobody on the design team was thinking about file size because that is not their job. But nobody on the development team was checking either.
The theme bloat. Their WordPress theme was loading 18 separate CSS files and 22 separate JavaScript files before rendering anything visible on screen. Eighteen style sheets. Twenty-two scripts. Total code payload before first paint: 1.8 megabytes. Most of those files were not needed on every page. The contact form plugin was loading its CSS site-wide even though the contact form was only on one page. The slider plugin was loading five different JavaScript files even though the slider only appeared on the homepage. It was the digital equivalent of leaving every light in your house on all the time because flipping switches is too much effort.
The fonts. Someone had loaded four Google Font families with eight font weights each. That is 32 individual font files. Many of these weights were not even used anywhere on the site. The browser was downloading fonts for text that did not exist. Each font file was small on its own — maybe 20 to 50 kilobytes — but added together they represented another several hundred kilobytes of unnecessary downloads.
The Fixes
I fixed all three problems over a weekend. Total time invested: about two hours. Total money spent: $59 on a caching plugin license.
The image. I downloaded the hero image, opened a free browser tool called Squoosh that Google built, resized the image from 2400 to 1200 pixels wide, and converted it from JPEG to WebP format. The file went from 2.4 megabytes to 89 kilobytes. You literally cannot see the difference on a standard screen. The image looks identical. It just loads in a fraction of the time.
The scripts. I installed WP Rocket, a $59 caching plugin. Deferred all JavaScript so it loaded after the visible content. Minified CSS and HTML to remove unnecessary characters. Combined CSS files where possible. Removed six plugins that were loading scripts on every page but only needed on specific pages. This part took the longest because I had to test each removal to make sure nothing broke.
The fonts. I removed three of the four font families that were barely used. Reduced from eight font weights to three: regular, medium, and bold. Added font-display swap so text stays visible while fonts load instead of showing invisible text. Total time: about twenty minutes.
The Before and After
LCP dropped from 6.2 seconds to 1.8 seconds. Cumulative Layout Shift — where page elements jump around while loading, which is incredibly annoying on mobile — dropped from 0.45 to 0.02. Google’s threshold is 0.1. They were 4.5 times over and are now well under. Total page size dropped from 5.8 megabytes to 1.2 megabytes. HTTP requests dropped from 87 to 34.
The recovery timeline: Week one after the fix: nothing visible. Week three: Search Console starts showing green instead of red on about 30 percent of pages. Week five: about 60 percent green. Traffic starts recovering. Week eight: traffic back to baseline. Week ten: traffic surpasses baseline by about 10 percent. Week twelve: new all-time high at 33,500 visitors, compared to the pre-drop baseline of 28,000.
The recovery took ten weeks. If I had given up after a month, I would have missed the turnaround.
The Part That Bothers Me
There were warnings for months before the crash. Google Search Console started showing “Core Web Vitals: Needs Improvement” in January — three full months before the traffic collapsed. I flagged it in a monthly report. The agency acknowledged receiving the report. Nothing happened. The problem did not feel urgent because the traffic was still there. Nobody cares about performance until the revenue graph turns red.
The entire fix took two hours and cost $59. The traffic loss cost the client over $60,000 in revenue over three months. If someone had run a Lighthouse audit — which takes thirty seconds and is completely free — in January, the whole thing could have been avoided.
Run a Lighthouse audit on your site right now. It takes less than a minute. If your LCP is over 2.5 seconds or your CLS is over 0.1, you have a problem that will eventually cost you money. The fix is almost always simpler than you think. And waiting will always cost more than fixing it ever will.
Related Articles
Mobile-First Design: Why Your WordPress Theme Is Losing Mobile Traffic
The Hidden Cost of Bad UX: What Slow Navigation Costs You Every Month
