Google Named the Numbers It Cares About

What was published

Earlier this week Google set out three metrics under the name Web Vitals, intended as a common way of measuring how a page feels to use.

The framing is that speed scores have been numerous, technical and hard to act on, and that three focused measurements are more useful than fifty.

Each one corresponds to something a visitor notices rather than something a developer measures.

What has not been said, at least not yet, is that these will directly affect search ranking. That is worth being precise about, because a good deal of commentary is already assuming it.

The three metrics

Largest Contentful Paint. How long until the main thing on the page appears. Usually the big image or the headline. This is the closest to what people mean by loading time.

First Input Delay. How long between somebody tapping something and the page responding. A page that looks ready but ignores taps for two seconds fails this.

Cumulative Layout Shift. How much the page moves around while loading. The reason you tap the wrong thing because a banner appeared and pushed everything down.

Together they describe loading, responsiveness and stability, which is a reasonable summary of whether a page is pleasant to use.

Why the third one matters more than it sounds

Layout shift is the least discussed and the most immediately recognisable.

Everybody has tapped a link and hit an advertisement because the page rearranged itself half a second later.

It is caused by images without stated dimensions, content injected after loading, and fonts that change the size of text when they arrive.

The fixes are unglamorous and effective: state the dimensions of images so the space is reserved, and avoid inserting content above what is already visible.

For a small business site this is frequently the easiest of the three to improve.

A worked example

A local restaurant whose menu page scored badly on layout shift.

The cause was a header image with no dimensions specified. The text loaded first and sat at the top, then the image arrived and pushed everything down by several hundred pixels.

Somebody tapping a menu item in that first second frequently hit something else.

Adding width and height attributes to the image, so the browser reserved the space before it downloaded, removed the shift entirely.

That was a five-minute change and it fixed a genuine irritation nobody had reported, because people do not report mistapping a link.

What to do about them now

1. Measure your main pages using the tools that now report these.
2. Fix layout shift first, which is usually cheapest.
3. Then loading, which is usually images.
4. Then responsiveness, which is usually scripts.
5. Re-measure, because assumptions about causes are frequently wrong.

That order is by effort rather than importance. Layout shift is often a handful of attributes; responsiveness can mean removing something a business relies on.

The counter-case

Reasons not to reorganise everything around this.

These are measurements rather than an announced ranking change. Rebuilding a site now on the assumption that they will become one is acting on speculation.

The metrics also describe things that were already worth fixing. A page that loads slowly, ignores taps and jumps around was a poor page last month, and the value of fixing it does not depend on who is measuring.

And for a small local business, the factors that decide whether you are found remain relevance, prominence and distance. A perfect score on a site nobody finds changes nothing.

Which suggests treating this as a useful vocabulary for problems you already had rather than as a new obligation.

Why naming things matters

The significant part of this announcement is not the metrics themselves. Similar measurements have existed for years.

It is that three of them have been selected, named, and put into the tools everybody already uses.

That changes the conversation. A supplier can now be asked about a specific number rather than about speed in general, and two people can disagree about a fact rather than an impression.

For a small business commissioning work, that is genuinely useful. "The main image should appear within two and a half seconds" is a requirement somebody can be held to.

Previously the equivalent conversation was about whether a site felt fast, which nobody could settle.

Where the numbers come from

Worth understanding because it explains why scores disagree.

Some tools measure in a controlled test on a simulated connection. Others report what real visitors experienced, gathered from browsers over a period.

Those produce different numbers, and the second is more meaningful and slower to reflect changes, because it is an aggregate over weeks.

So a fix made today will show immediately in a test and take a month or more to appear in field data, which catches people out and looks like the fix not working.

What this does not measure

Worth stating, because a metric attracts effort disproportionate to its importance.

None of these says whether the page answers the question, whether the business is any good, or whether somebody can find what they came for.

A fast, stable, responsive page with nothing useful on it scores well and produces nothing.

The metrics measure the delivery rather than the content, which is exactly what they are meant to do and is only part of what makes a page work.

What a small business should actually do

Proportionate advice, since most of what is written about this is aimed at large sites with development teams.

Measure once, on your two or three busiest pages, and write the numbers down.

Fix the obvious causes: image dimensions, oversized files, and anything loading that nobody uses. Those three cover most of what a small site can improve.

Then leave it. Re-measure in a few months and only act if something has got worse.

What is not worth doing is chasing a perfect score, which for a site built on a general-purpose theme is frequently unachievable without a rebuild, and which would not repay the cost.

The metrics are a diagnostic rather than a target, and the businesses that treat them as a target spend a great deal of money on the last few points.

What to expect next

Reasonable inference rather than prediction.

Google has generally introduced measurement before consequence, and having named these publicly and built them into its tools, it seems likely they will feature more prominently over time.

That is speculation and the sensible response either way is the same: fix the things that make your pages unpleasant to use, because that is worth doing regardless.

The concrete work is covered in why your site feels slow on a phone.

Frequently asked questions

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.