Skip to content

Codify Yo!, AI-assisted no-code migration, now in betaGet early access

Yo! No CodeEngineering team
Original research · August 2026

The State of No-Code Performance 2026

We ran Google PageSpeed against 20 production sites built on Bubble, Webflow, and WordPress. None reached Google's “good” threshold. All 20 failed Largest Contentful Paint. And the platform with the highest achievable ceiling finished last.

Findings

Four numbers worth knowing.

0 of 20
sites scored 90+ on mobile

Not one site in the sample reached Google's 'good' threshold. The best result across all three platforms was 81. If your no-code app is scoring in the 50s, it is not an outlier — it is the middle of the distribution.

20 of 20
failed the 2.5s LCP threshold

Largest Contentful Paint is the Core Web Vital most tied to bounce rate, and every site tested exceeded the threshold Google considers good. Median LCP across the sample was over 7 seconds.

44 vs 88
Webflow's observed median vs its achievable ceiling

Webflow has the highest performance ceiling of any platform we audit — a lean Webflow site can genuinely reach the high 80s. The sites in our sample landed at a median of 44, the worst of the three platforms tested.

6.6 MB
median page weight on Webflow sites

Nearly triple Bubble's median. The Webflow sites sampled were largely marketing sites carrying heavy imagery and animation. That is a choice made by their builders, not a constraint imposed by Webflow.

The data

Every platform, measured.

Mobile PageSpeed scores. Medians rather than means, because small samples with outliers make means misleading.

PlatformSitesMedian scoreRangeMedian LCPMedian weightOur ceiling estimate
WordPress66330815.8s3.1 MB75
Bubble56036767.5s2.3 MB62
Webflow944305911.9s6.6 MB88

“Our ceiling estimate” is the realistic best-case score we publish in the No-Code Performance Checker for a well-optimised non-trivial app on that platform. It is a ceiling, not a prediction — and the gap between it and the observed median is the most interesting column in this table.

The finding that surprised us

The ceiling is rarely the binding constraint.

We expected the ranking to follow the ceilings. Webflow has the highest achievable performance of the three — it produces near-static HTML and CSS, and a lean Webflow site genuinely reaches the high 80s. Bubble has the lowest, because its runtime has to boot before anything paints.

The observed order was the reverse. Webflow finished last with a median of 44, more than 40 points below what the platform is capable of. Bubble landed at 60, within two points of the ceiling we publish for it.

What explains it

Page weight. The Webflow sites in this sample carried a median of 6.6 MB, nearly three times Bubble's 2.3 MB. They are mostly marketing sites built to look impressive, loaded with large imagery and animation. None of that is imposed by Webflow. It is a series of choices made by the people who built them.

Bubble apps, by contrast, are functional products. They tend to be visually plainer, which means they spend less of their budget before the runtime cost even applies — and so they land close to their ceiling rather than far below it.

The practical implication runs against the usual advice. If your site is slow, the platform is probably not your problem yet. Most sites in this sample were nowhere near the limit of what their platform could do, which means the available gains were sitting in their own images, scripts, and fonts.

Method

How this was measured, and what it cannot tell you.

What we did

  • · Assembled a candidate list of production sites from public showcases and industry listicles.
  • · Verified platform attribution by fingerprinting each site's own HTML rather than trusting the listicles. Webflow was identified by data-wf-page, Bubble by its runtime globals, WordPress by wp-content paths.
  • · Discarded every candidate whose platform could not be confirmed from its own markup. Several widely-cited “built with Bubble” examples turned out to be Webflow, and one platform vendor's own marketing site is built on a competitor's product.
  • · Ran Google PageSpeed Insights (mobile strategy) against each confirmed site on 22 August 2026.
  • · Reported medians, and reported results in aggregate without naming individual sites.

What this is not

The sample is small and not random. Twenty sites across three platforms, drawn from public showcases. Showcase sites skew toward the ambitious end, which likely depresses scores relative to the average site on each platform.

The categories are not equivalent. The Webflow sites are mostly marketing sites; the Bubble sites are mostly applications. That difference explains much of the gap and means this is not a like-for-like platform comparison.

Three sites failed to return results and were excluded rather than retried into the numbers.

PageSpeed is lab data. It approximates a mid-tier mobile device on a throttled connection. Real users on better hardware will see better numbers.

We are publishing the limitations this prominently because the alternative — a confident-sounding chart with a hidden n — is how most vendor “research” in this category gets made, and it is not worth citing.

What to do with this

Measure yours before you conclude anything.

The useful takeaway from this data is not a platform ranking. It is that most sites are far enough from their platform's ceiling that the ceiling is irrelevant to them. Before concluding that you have outgrown your platform, find out how much of your gap is actually yours.

Our free checker runs the same PageSpeed test used in this research and splits every finding into what you can fix on your platform and what is genuinely framework overhead. No email, no signup, and it will tell you to leave your app alone if the numbers say so.

Researchers and journalists: the per-platform aggregates on this page may be reproduced with attribution to Yo! No Code and a link to this page. We are happy to share methodology detail on request at hello@yonocode.io.

Ready when you are

You can generate an app in an afternoon.
Owning one is a different job.

30-minute scoping call. Fixed-price plan in your inbox within 48 hours. Whichever service you need, Launch, FullCode, Codify Yo!, or Teams, you're talking to the team who'll do the work.

✓ 48-hour fixed-price quotes✓ North American oversight✓ Global engineering economics