More sales from the traffic you already have
Merchants ask which theme is fastest as if speed were a fixed property of the code, but tracking five themes across six months of CrUX history shows the pass rate for a single unchanged theme can move 10 points in a month, which means the question itself is broken.
Merchants ask “is theme X fast” as if speed were stamped into the code once and for all, the way a file size is. It isn’t. Core Web Vitals pass rates are measured on real visitors to real stores, and the population of stores running any given theme changes every month, which means the number changes every month too, without a single line of the theme’s code being touched.
Across five themes with monthly CrUX history from March 2026 to August 2026, the share of stores passing all Core Web Vitals moved within a single theme by 10.0 points at the median, and by as much as 12.3 points for the theme that moved the most. That range sits inside one theme, not between two different themes, so the spread is not explained by one codebase being better engineered than another. It is explained by the fact that “passing Core Web Vitals” is a snapshot of a shifting group of stores, and the group shifts every month.
One theme in the tracked set moved from a 76.6% pass rate to an 88.9% pass rate across the six month window. Nobody edited that theme’s code during that time. What moved was the base of stores running it: which merchants were live that month, what they’d added to their product pages, what apps sat on top of the theme, and how their traffic mixed between fast and slow connections. The theme’s files stayed still. The number describing it did not.
Compare that to the smallest range in the same set of five themes, which was 3.0 points over the same six months. So “how fast is this theme” only has a real answer if you also say “compared across what range of time,” because without that qualifier the honest answer is a spread, not a single figure.
| Metric | Value |
|---|---|
| Median swing within a theme (6 months) | 10.0 points |
| Largest swing within a theme (6 months) | 12.3 points |
| Smallest swing within a theme (6 months) | 3.0 points |
| Widest single theme range (min to max) | 76.6% to 88.9% |
| Median month over month move | 3.1 points |
| Largest month over month move | 7.5 points |
| Median swing, desktop | 8.7 points |
| Median swing, mobile | 10.0 points |
The month over month move across the five tracked themes sits at 3.1 points at the median, and the largest single month over month jump recorded was 7.5 points. That is not a slow drift. It means picking last month’s number and this month’s number and calling the difference an improvement, or a regression, can produce a swing bigger than what a lot of page level speed work claims to deliver. If a merchant reads a 5 point gain after a code change and the same theme’s baseline noise alone can move 7.5 points between two ordinary months with no change at all, the 5 point gain is inside the noise, not proof of anything.
The numbers behind this rest on monthly CrUX field data pulled for five themes across the March 2026 to August 2026 window, checked September 4, 2026, which reflects real visitor sessions rather than a lab test run against a fixed page.
The same five themes showed an 8.7 point median swing on desktop against a 10.0 point median swing on mobile, measured over the same six months. Mobile traffic is more varied in device and connection quality than desktop traffic tends to be, and that variance shows up directly in how much the pass rate moves month to month. So a desktop number from one month sitting next to a mobile number from a different month is not a comparison, it’s two different instruments being read as if they were one.
None of this means monthly measurement is useless, it means a single month’s number needs to be read as a point inside a known range rather than as a fixed verdict on a theme. In broader theme rankings that include speed as one factor among several, UTD’s themes appear in the listings the same as any other vendor’s, and merchants comparing options across that catalog can look at the full listing of UTD themes alongside the rest of the market rather than treating any single vendor’s number as settled.
Pull at least three months of CrUX history for any theme before deciding it’s fast or slow, and treat a one month reading as a point inside a 10 point range rather than as the theme’s final score.
Because the score is not a property of the theme's code, it is a property of the stores running that theme in a given month. CrUX field data for five themes tracked from March 2026 to August 2026 shows the passing share moving by 10.0 points at the median with no code change involved. Stores get added, removed, get heavier product images, add apps, or change hosting configuration, and all of that shifts the aggregate even though the theme file itself sits untouched in its repository. So a single snapshot tells you about that month's fleet of stores, not about the theme's ceiling.
The facts here do not support ranking themes by a single speed number, because the same theme can sit anywhere inside a range as wide as 12.3 points depending on the month you check. The narrowest range among the five tracked themes was still 3.0 points. So instead of asking which theme is fastest, the more answerable question is how wide a given theme's range has been over six months, and that range is the more honest number to compare.
No. The same themes showed an 8.7 point median swing on desktop against a 10.0 point median swing on mobile, tracked over the same six months. That gap means desktop and mobile are different measurement surfaces with different volatility, so a desktop pass rate from one month is not a fair stand-in for a mobile pass rate from another month, or vice versa.
Less than it looks like. The median month over month move across the five tracked themes was 3.1 points, and the largest single month over month move was 7.5 points. That means picking two adjacent months and reading a difference between them can produce a number bigger than what many code level speed fixes report as their gain, so a two month comparison without a wider window can mislead more than it informs.
The theme's own files did not change during the six month window covered by this data, so the swing between 76.6% and 88.9% sits entirely on the store side: which merchants were running the theme that month, what products and images they added, what apps they installed, and how their traffic mix shifted. CrUX aggregates real visitor experience per origin, so a theme's number is really an average of whatever stores happen to be live under it that month.