The Readiness Teardown: Are the Sites That Already Rank Machine-Ready?

The measurement frame, and this cut of it

Over August 2026, Raveneye Global measured eight local service trades across six ordinary, non-metro Indian cities: Jaipur, Indore, Lucknow, Surat, Kochi, and Nagpur. The trades were chartered accountants, dentists, gyms, interior designers, digital marketing agencies, coaching institutes, physiotherapists, and wedding photographers. The underlying data came from live APIs pulled and verified during that window: Google Search results and AI Overview data, Google Places, Google Ads search volume, the Google Knowledge Graph, Google PageSpeed Insights, and DNS-over-HTTPS lookups. This article is one cut of that dataset, and it asks a narrower question than the frame as a whole: once a business already shows up when someone searches for one of these eight trades in one of these six cities, what does its website look like when Google's own mobile speed test is pointed at it?

For this cut, eighteen websites belonging to businesses that surfaced in the underlying local searches were run through Google PageSpeed Insights, mobile report, on 2026-08-10. Two further sites belonging to national chains, a multi-city dental chain and a large home-interiors platform, were tested alongside them purely as a size-and-budget comparison; their figures are reported separately below and are not folded into the 18-site averages. All 18 local sites returned a usable performance score; none failed to load or errored out of the test. Only the performance category is reported in this article. PageSpeed Insights also scores accessibility, best practices, and search-engine fundamentals, and those results sit outside the scope of this piece.

A median score of 58, a median load of 12.4 seconds

Lighthouse grades its own 0-to-100 performance score into three bands: 0 to 49 is marked red, poor; 50 to 89 is marked orange, needs improvement; 90 to 100 is marked green, good. Across the 18 local business sites tested, the median mobile performance score is 58 out of 100, placing the typical site for these eight trades in the middle of the orange band, closer to failing than to a clean pass. Six of the 18 sites, roughly a third of the sample, scored under 50 and landed in the red band outright.

The score is a composite, and the composite hides where the failure actually sits. Under Lighthouse's own current weighting, Total Blocking Time carries 30 percent of the performance score, Largest Contentful Paint and Cumulative Layout Shift carry 25 percent each, and First Contentful Paint and Speed Index carry 10 percent each. Median Total Blocking Time across the 18 sites is about 180 milliseconds, which sits inside Lighthouse's own green, fast band for mobile (under 200 milliseconds). Once a page for these sites starts rendering, it tends to respond to a tap or a scroll reasonably well. The failure is concentrated earlier, in getting anything meaningful onto the screen in the first place.

That earlier failure is stark. The median lab Largest Contentful Paint, the time it takes the largest visible element in the main viewport to render under the simulated mobile conditions described above, is 12.4 seconds across the 18 sites. Google's own definition of a good LCP is 2.5 seconds or less, measured at the 75th percentile of real visits. Seventeen of the eighteen sites measured have an LCP over that 2.5-second line; only one clears it. A visitor on a typical mobile connection is being asked to wait roughly five times longer than Google's own good threshold before the page they searched for shows them anything, on all but one of the sites currently winning these searches.

Lighthouse's own diagnostics typically trace a slow Largest Contentful Paint to a recurring, well-documented set of causes: unoptimized hero images and video posters, render-blocking CSS and custom web fonts loaded before anything else, and third-party scripts, chat widgets, and marketing tags firing ahead of the content a visitor actually came for. This teardown did not audit each site's individual cause; it measured the outcome, not the diagnosis. A median LCP running close to five times the good threshold is nonetheless the signature of accumulated, unaddressed technical debt across a site's build, not a single missing setting.

The spread: what 91 proves, and what 2 to 24 confront

A median obscures a range, and the range here does real work. The best-scoring site among the 18, the interior-design firm decoruss.com, reached 91 out of 100, the single highest score in the set and the only one that clears into Lighthouse's green band. Its Largest Contentful Paint measured 2.5 seconds, landing exactly at the edge of Google's own good threshold rather than comfortably inside it. Even the strongest performer in this competitive set is not resting on a fast page. It got there and stopped at the line.

That single data point matters more than its rank suggests, because it settles a question this teardown would otherwise leave open: is a fast page simply unreachable for a small local business site in these categories, or is it a matter of will and maintenance? decoruss.com answers it. It is not a national chain and not an outlier category; it is one interior designer's website, competing in the same searches and under the same constraints as every other site in the sample, and it reached a score most of its direct competitors did not attempt.

The other end of the range is where most of the sample actually sits. The four lowest scores measured in this teardown were:

• kosmicdentalclinic.com, score 2: a dental practice site, effectively failing the mobile test outright.
• dreamsdecorinterior.in, score 15: an interior-design competitor to the study's own best performer.
• dlifeinteriors.in, score 19: a second interior-design site in the same low band.
• citydentalnagpur.in, score 24: a dental practice site in Nagpur.

Budget does not appear to buy readiness

A reasonable objection to reading too much into 18 small-business sites is that scale and budget would fix this quickly if it mattered enough. Two national chains were tested alongside the local sample specifically to check that objection: a multi-city dental chain, clovedental.in, and a large home-interiors platform, livspace.com, both operating well beyond any single city.

Neither cleared the bar the local sites were missing. clovedental.in scored 56, inside the same orange band as the local median, with a Largest Contentful Paint of 35.6 seconds, the single slowest reading anywhere in this teardown, local or national. livspace.com scored 45, in the red band, with an LCP of 10.4 seconds. Both readings are worse on load time than the 18-site local median of 12.4 seconds would suggest a funded competitor should manage.

The finding this supports is narrow but useful: whatever is producing these mobile scores, it does not look like a simple resourcing gap that a bigger marketing budget or a national footprint automatically closes. decoruss.com, a single local interior-design site, outperformed a national interiors platform on the same test by 46 points and roughly eight seconds of load time. On this evidence, readiness looks decided by attention to the site, not by the size of the business behind it.

This also matches a pattern documented across the web performance industry more broadly, not a peculiarity of these two businesses: larger organizations tend to accumulate more third-party marketing tags, analytics scripts, and embedded tools over time, each one a further tax on load time, and size alone does not immunize against it. This teardown did not diagnose the specific cause on either chain's site; it measured the outcome. The outcome is consistent with that general, well-known pattern.

The businesses Google can barely see: the CrUX gap

There is a second measurement inside this same test that is easy to skip past, and it may be the more consequential one. PageSpeed Insights reports two kinds of data when it can: lab data, the simulated score and metrics discussed so far, and field data, real measurements from actual visitors collected in Google's Chrome UX Report (CrUX). Field data only appears for a site once it clears an undisclosed minimum-traffic threshold that Google sets specifically to keep the numbers statistically reliable; sites below that threshold are simply excluded from the dataset, with no field data reported at all.

Only 4 of the 18 local sites in this teardown had any CrUX field data whatsoever. For the other 14, Google does not have enough real-world Chrome visits to report a single field number, not a good one and not a bad one. For three-quarters of this sample, the lab score is the only performance read that exists anywhere in Google's systems.

Read one way, that is a measurement limit: it means the lab scores above cannot be cross-checked against how real visitors actually experienced 14 of the 18 sites. Read the other way, it is itself a finding. A site absent from CrUX is a site Google's own infrastructure has not seen enough live human traffic to characterize. For a local business, that is a second, quieter readiness problem sitting underneath the speed problem. For most of this sample, the site is slow, and it is also largely unmeasured by the system that increasingly decides what gets shown.

This also reframes what a business owner might reasonably believe about their own site from casual use. A site can feel fine to the person who built it, opened on a fast office connection, on a desktop, already cached from a dozen prior visits. That subjective impression and Google's own measurement can diverge sharply, and for 14 of these 18 sites, there is currently no field data at all to settle the disagreement one way or the other. The lab score is not a second opinion here. For most of this sample, it is the only opinion that exists.

Why this is a fragile position, not a comfortable one

Put the last five sections together and two readings follow from the same numbers, and both are true at once. The first is that the bar being cleared today is low enough that a genuinely fast site would stand out immediately: decoruss.com's 91 proves the ceiling is reachable by a business with no more resources than its competitors, most of whom are not attempting it. For a business willing to fix its loading speed, that is a real, available advantage, not a theoretical one. The second reading is that the position of the businesses ranking today is fragile, not settled, because the mechanisms that reward speed are documented, not speculative, and they are not static.

What this establishes, and what it does not

This is a lab measurement of 18 sites, on one date, inside one field study covering eight trades across six Indian cities. It should be read as exactly that: a snapshot of a moving surface in August 2026, not a permanent verdict on any business named here or on the Indian small-business web at large. A site's PageSpeed score can change with the next plugin installed, the next image uploaded, or the next redesign, and Google's own thresholds and weightings have moved before and will move again.

Inside that limit, the finding is still a clear one. The sites already winning these eight kinds of local searches in these six cities are, on the whole, not machine-ready on speed. A median score of 58, a median load of 12.4 seconds, and 17 of 18 sites over Google's own good threshold describe a competitive field where almost nobody has cleared a bar Google itself calibrated to be reachable by most of the real audience. The one site that did clear it, and the two funded national chains that did not, together rule out the easiest excuses: it is not that the bar is unfair, and it is not that only a large budget can clear it.

What is left is closer to an opportunity than an indictment. The businesses ranking today are exposed on a documented mechanism they have not yet been visibly punished by, since they are still the ones showing up in the search results this study captured. That is precisely what makes the position fragile rather than safe: the businesses have not been penalized yet, on a signal Google has already named and is already weighting into both its ranking systems and the eligibility chain feeding its AI Overviews and AI Mode answers.

Set beside the rest of this field study, the pattern fits a broader shape. These businesses are, on the whole, winning their current visibility on signals other than technical readiness, whether that is review volume, listing longevity, or category relevance, while the newer, machine-facing layer of the web, loading speed among it, sits comparatively untouched. That is not a criticism of any single business named in this teardown. It is a description of where the remaining, uncontested ground currently sits.