The Three Businesses at the Top of Google Load Slower Than the Ones at the Bottom

You know the call. Somebody has run a free scan of your website. There is a score out of 100 and it is not good. Your site is slow, that is why you are not showing up on Google, and for a monthly retainer they can fix it.

The scan is real. The score is usually real. The site probably is slow.

The conclusion is wrong, and we can now show that with numbers.

The myth, stated plainly#

The pitch goes like this: Google measures how fast your website loads, so a faster website ranks higher, so pay us to make your website faster and you will climb the map.

Two of those three clauses are true enough. Google does measure page speed. Page speed is a real ranking signal in web search. The failure is in the jump to the local pack, that box of three businesses with the map beside it, which for a category like auto repair sits above the ordinary web results and takes the tap before anybody scrolls.

So we tested it.

The method#

Between July 13 and August 8 of this year we ran the same measurement pass across 21 category and town combinations on the South Coast and just over the Rhode Island line: auto adjacent trades and everything around them, in Fall River, New Bedford, Somerset, Swansea, Dartmouth, Westport, Fairhaven, Tiverton and Providence. Plumbers, HVAC, electricians, movers, restaurants, salons, day cares, florists, print shops, vets, chiropractors, optometrists, furniture stores, jewelers, cleaners, driving schools.

For each one we pulled the businesses Google returns for "category in town," kept them in the order Google returned them, and measured every website with Google's own PageSpeed Insights on mobile. That gives 280 businesses with a recorded position and 205 with a measured Largest Contentful Paint, which is the moment the customer actually sees something rather than a white screen. Google calls under 2.5 seconds good and over 4.0 seconds poor. Those are Google's thresholds, not ours.

Then we simply asked whether position and speed move together.

What came back#

Read that column again. The businesses at the top are the slowest ones in the set. Position eight and below is 2.4 seconds faster at the median than the top three.

Run it as a correlation and the relationship between position and load time comes out at -0.18, meaning that if anything the better-ranked business is slightly slower. Restrict it to the 53 sites where Google had enough real visitor traffic to report field data rather than a lab estimate and the sign flips to +0.10. A relationship that reverses direction depending on which honest subset you look at is not a relationship. It is noise.

And in the 16 categories where we measured at least five sites, the single fastest website in the category held the number one spot exactly once. Its median position was eighth.

The cleanest disproof is the simplest one. Thirty six of the 280 businesses have no website at all. Six of them rank in the top three. You cannot have a page speed problem when you have no page. Whatever is putting those six at the top of the list is not on their website, because there is no website to put it on.

Google says this out loud#

None of this is secret. It is in the documentation, in plain English, and almost nobody in the "free website scan" business quotes it.

Google Search always seeks to show the most relevant content, even if the page experience is sub-par.

The same page goes further, warning that good scores in speed tools do not guarantee ranking and that chasing a perfect score for SEO reasons may not be the best use of your time. Google's own Business Profile help page on local ranking lists three factors: relevance, distance, and prominence. Page speed does not appear on it.

Nor does it appear in Whitespark's 2026 Local Search Ranking Factors, in which 47 practitioners scored 187 separate local pack factors. Not one of the 187 is page speed, load time, or Core Web Vitals. The factors that score highest there are your primary Google category, your proximity to the person searching, having a physical address in the city being searched, and being open at the moment of the search. Review signals cluster just underneath.

That is the honest shape of the thing. A Somerset shop ranks for a Somerset search largely because it is in Somerset, it is categorised correctly, and people have reviewed it.

What we cannot tell you#

Our own data is good at disproving one claim and bad at proving the replacement. The positions we recorded come from Google's text search for "category in town," which is the order a customer sees for that query, not a personalised map pack measured from one specific pin on one specific phone. Distance from the searcher is the one factor we could not hold still, and it is probably the biggest one in the box.

Most of our load times are lab estimates rather than field data, because Google only reports real visitor data for sites with enough traffic, and small local sites usually do not have it. And when we checked our own numbers for a review effect, we found no usable correlation between position and review count either, which we take as further evidence that proximity is swamping everything at this sample size rather than as evidence that reviews do not matter.

We are publishing the limits because you can reproduce all of it. Paste any local business URL into pagespeed.web.dev and you will get the same kind of numbers we did, with the caveat that a lab estimate moves a little run to run and field data shifts over its 28 day window.

The Somerset shops, measured today#

These shops are not careless, and this is not a report card. Several of the independent garages in town have been open for decades. They built their reputation the hard way, one bay at a time, and that reputation is what puts them in front of a customer. The website is not what got them there.

With that said, here is the state of the ground, measured on August 9 and deliberately reported in aggregate. Nine independent Somerset repair shops, chains and dealership service departments excluded. Fewer than half have a working website at all; the rest have a Facebook page, a directory stub, or nothing.

Every one of the sites that does exist fails Google's 2.5 second threshold on mobile. The quickest of them showed something in 3.7 seconds. The slowest took 19.2. Most of them sit in Google's poor band. Across the pages we could read, not one carries LocalBusiness schema, the small block of code that tells Google your hours and address as fact rather than as a guess, and there are still basics missing that cost calls: a phone number that is text rather than a link, a page served without a certificate.

One of those sites we could measure for speed but not for structure. It arrives as a one kilobyte shell and assembles itself in the browser, so our structural checks returned nothing rather than a fair reading. That is recorded as unmeasured, not as a failure.

We are not naming any of them and we are not pairing any shop to any number, because none of them asked to be graded in public.

But 19.2 seconds is a real cost. Just not the cost the sales call claimed.

Where speed actually earns its keep#

The other thing to keep in mind with core web vitals is that it's more than a random ranking factor, it's also something that affects your site's usability after it ranks (when people actually visit). If you get more traffic (from other SEO efforts) and your conversion rate is low, that traffic is not going to be as useful as when you have a higher conversion rate (assuming UX/speed affects your conversion rate, which it usually does).

After it ranks. That is the whole argument in two words. In fairness, the article that quote comes from is making the case that Core Web Vitals matter more than people assume, and we are not disputing that. We are disputing where in the queue it belongs.

A customer with a check engine light on Route 6 finds you in the map pack because you are the closest shop with decent reviews. Then they tap your website. If nothing appears for nineteen seconds, they go back and tap the shop underneath you. The slow site did not cost you the ranking. It cost you the call the ranking had already won.

Speed is a conversion problem wearing a ranking problem's clothes. Fix it for the right reason and you will value it correctly, which usually means fixing it fourth rather than first.

The free fix, today, in twenty minutes#

Do these in this order. None of them costs anything.

One. Make your phone number tappable. Find your phone number on your homepage and check on an actual phone whether tapping it starts a call. If it does not, whoever maintains your site changes this:

508-555-0100

into this:

<a href="tel:+15085550100">508-555-0100</a>

That is a five minute edit and it is the highest return five minutes in this entire article. On a mobile page, a phone number you cannot tap is a phone number most people will not dial.

Two. Open your Google Business Profile and check your primary category. Not your secondary ones, the primary. It is the single highest scored factor in Whitespark's survey of 187 local pack factors. "Auto repair shop" and "Auto body shop" and "Tire shop" are different categories and Google treats them as different businesses. If yours is wrong, or generic, fix it. Five minutes.

Three. Confirm your hours are current and your address is displayed. Being open at the moment somebody searches is one of the highest scored factors in Whitespark's list. A profile showing last year's holiday hours is quietly losing you the searches that happen at 7am on a Tuesday. Five minutes.

Four. Ask three customers this week for a review. Not thirty. Three, then three more next week. A steady flow of reviews scores better in that same list than a single burst does. Five minutes to write the text message you will reuse.

None of that touches your website. That is the point.

What we got wrong before#

Worth saying plainly, because we have made this argument the other way round in the past. Our own earlier write ups led with load time as the headline number for local businesses, and that framing overstated its effect on ranking. Speed is still worth fixing. It is worth fixing because it converts the visitors you already earned, and it belongs after your profile, your categories, your hours, and your reviews. The order matters more than the list.

If you run a shop in Somerset and you want to know which of these actually applies to you, we will run the same measurement on your site and read you the result, whether or not it says anything is wrong. No score out of 100, no retainer attached.

Joshua Amado, Avalon Partner. 774.559.8992 · Joshua.Amado@AvalonPartner.com