How the benchmark is built

How dealer sites are found and verified

Sites are found mainly through Google search queries that pair a trade with a US city and state. The trades searched are flooring, tile, cabinet, countertop, and kitchen and bath.

Discovery is not verification. Search surfaces many sites that are not dealers, so the two are separate steps. A site found by search is measured first, and only a site that measures successfully goes on to be classified. It joins the verified cohort only if that classification marks it as a dealer. Some cohort sites did not come from search: they came from a supplied dealer list, and they joined on that basis without the classification step. Each verified site counts once, using its most recent usable phone measurement.

In the first discovery frame, the benchmark covers over 7,700 verified dealer websites, flooring and related trades. Comparisons are made only against the verified cohort, never against everything search turned up.

Each discovery frame, meaning the rule for which cities and trades were searched, is recorded as its own population version. The first frame was frozen in August 2026, and every figure on the benchmark page is computed over it. The current frame reaches smaller towns, and its sites are not yet part of the verified cohort. Every cohort figure states which frame it was computed over, and a figure spanning more than one says so, so results from different frames are never silently mixed.

What we measure, and with what

Every measurement is a lab test run through Google PageSpeed Insights, which runs Lighthouse against the page under simulated phone or desktop conditions. Each test is a single run at one moment. From it we read:

• Performance score. Lighthouse's overall performance score for the page.
• Largest Contentful Paint (LCP). The time until the largest piece of content on the first screen has appeared. This is the load-time measure the snapshot ranks.
• Cumulative Layout Shift (CLS). How much the page moves around while it loads.
• The desktop score. The same test under the desktop profile, kept as context rather than ranked.

Because a lab test is a single reading, results can vary from one run to the next. A snapshot is a first signal about how the site tests on a phone, not a final verdict on the site.

What we deliberately do not measure

• Search rankings, traffic or visibility. Nothing here measures where a site ranks in Google or how many people find it.
• What visitors did. A lab test does not observe real visitors, so it says nothing about abandonment, leads, calls, conversions or revenue. We do not turn a load time into a claim about customer behaviour.
• Google's own Core Web Vitals assessment. Google decides whether a page passes using data from real visits recorded by Chrome. A lab test cannot produce that verdict, so we never say a site passes or fails it. We compare a lab measurement against the threshold Google publishes, and we say so each time.

How the comparison works

When you submit a site, we test it at that moment on both profiles and rank it against verified dealers. The comparison group is chosen in this order, stopping at the first group large enough to rank against:

1. Dealers in your trade, in your state.
2. All verified dealers in your state.
3. Dealers in your trade, in your census region.
4. Dealers in your trade, nationwide.
5. All verified dealers, nationwide.

The snapshot ranks two things against that group: the overall phone performance score, and phone load time (LCP). Your site is ranked as if it were one of the group, so ranks are shown out of the group's size plus one, and "outperforms" is the share of that total your site finished ahead of.

Where the group's spread is shown, it uses the fastest tenth, the median and the slowest tenth of load times rather than the single fastest or slowest site. That resists outliers and keeps any one dealer from being singled out.

The snapshot also reports whether layout shift is in the good range or needs attention, and shows the desktop score for context. Desktop is not ranked. It is shown for context, because a site can test well on the desktop profile and poorly on the phone profile.

See an anonymized example of a real snapshot.

What we will never do

• Sell the fix. We diagnose problems. We do not fix them. We do not offer development, design or marketing services, so there is no work to sell on the back of a finding.
• Name a dealer in published research. Benchmark figures describe the population. No dealer is identified in anything we publish, and the sample snapshot on this site is anonymized.

That separation keeps our advice unbiased. Take a snapshot to any developer or agency you trust, handle it in-house, or get a second opinion.

Data retention

• When you request a snapshot, we keep the website address and email you submit so we can deliver the result, for as long as we need them to provide the service.
• Measurements of a website may be kept as part of the benchmark.
• For routine tests we keep the measurements we extract and discard the full raw test output after a short rolling window.
• You can ask us to delete your information at any time by writing to hello@showroomstandard.com.

The privacy policy is the governing text.

How often figures are refreshed

Every published cohort figure carries the date of the measurement it comes from, and cohort figures are rechecked quarterly. When a recheck moves a figure, it is either corrected or kept with the date of the measurement it describes, rather than left to drift.

A snapshot of your own site is measured live when you submit it. The dealers it is compared against use their most recent measurement on file.