Full disclosure: this analysis uses Stackra's own website corpus, and Stackra sells a website audit that reports on both of the things measured here. Weigh it accordingly. The numbers below are install counts and field measurements, not opinions. You can reproduce the pattern on your own site with any free speed testing service.
Across a million US websites, the ones carrying more layers are consistently better structured for search engines and AI answer engines. They are also, on average, slightly slower. Those two facts get bundled into advice about what a script costs you. The bundling is wrong, because only one of the two is a mechanism, and telling them apart changes what you should install.
What 464,817 WordPress sites look like, layer by layer
Each row compares WordPress sites that run a given layer against those that do not. Structured data is the share carrying JSON-LD markup. The speed column is the share with a confirmed-good Core Web Vitals result. These are two populations of sites, not the same site measured twice.
| Layer | Structured data | Meta descriptions | Contact forms | Confirmed-good CWV |
|---|---|---|---|---|
| SEO plugin (Yoast, RankMath) | 99% vs 28% | 86% vs 43% | 62% vs 61% | 61% vs 64% |
| Google Tag Manager | 93% vs 73% | 88% vs 67% | 66% vs 59% | 58% vs 66% |
| Google Analytics | 87% vs 59% | 81% vs 56% | 63% vs 56% | 61% vs 67% |
| Marketing automation | 88% vs 80% | 81% vs 74% | 74% vs 59% | 60% vs 62% |
Source: Stackra analysis of stackra_us_corpus, 464,817 US WordPress sites, May 2026 crawl vintage.
One column is a mechanism, the other is a population
This distinction is the reason most versions of this article are wrong, so take it before the numbers.
- •The structure column is mechanical. Yoast and RankMath write JSON-LD into the page. The 28% to 99% gap is the plugin doing the thing it exists to do, and installing one on your site will produce it.
- •The speed column is not. Sites that run more layers also tend to be larger, busier and more complex, and complexity slows sites down on its own.
- •When we tested a closely related question, holding total technology count constant removed most of a far more dramatic effect. Free tools dragged on responsiveness as hard as paid ones.
- •The speed figure also counts only sites with real-world field data, so it is a confirmed-good floor rather than a pass rate. Most small business sites never reach the traffic threshold to be measured.
So read the last column as a description of what those groups of sites look like, not as a price tag on a script. Nothing here establishes that a specific plugin slowed a specific site down.
The gaps are nowhere near equal
With that said, the rows separate sharply. Sites with an SEO plugin sit 71 points higher on structured data and 3 points lower on confirmed-good speed. Sites with Tag Manager sit 20 points higher on structure and 8 points lower on speed. Analytics lands between them at 28 and 6. Marketing automation is 8 and 2.
The SEO plugin is the outlier and it is the one you can act on with confidence. It has by far the largest structural gap, the smallest speed gap, and a known mechanism behind the structure. Tag Manager has the opposite shape: a modest structural gap next to the widest speed gap in the set. That is not proof the container is slowing those sites, but it is where the heaviest sites cluster, and containers do accumulate tags nobody removes.
Why structured data is worth having at all
Structured data is how a machine reads your business: who you are, what you sell, where you operate, who wrote the page. It is an E-E-A-T signal for search engines and it is what AI answer engines read when deciding whether to cite you. That second surface is the one growing. ChatGPT, Perplexity, Gemini and Google AI Overviews all read structured data. Google has confirmed that continues even for schema types that no longer produce a visible search result.
Do not add schema expecting visual search enhancements. Google removed FAQ, Article and HowTo rich results, FAQPage for every site type as of May 2026. The reason to add schema now is machine comprehension and AI citation, not decoration in the results page.
What to actually do with this
The practical reading is simple. Install the SEO plugin, since the structural gain is mechanical and you will get it. Keep analytics, because you cannot improve what you are not measuring. Audit what you load through Tag Manager, since that container is where weight accumulates unnoticed. Then measure your own site rather than trusting an average.
How much of your speed ceiling is set by your platform before you add anything.
Which structured data types actually matter, and how to add them.
Reading your own Core Web Vitals instead of relying on a benchmark.