Lightweight GA Connectors That Do Not Hurt Core Web Vitals
The analytics tag is often the single heaviest third-party request on an otherwise well-optimised site. It arrives from another domain, it executes on the main thread, and it does all of this on every page, including the ones you spent a fortnight making fast.
There are real ways to reduce that cost, and one obvious way to avoid a good deal of it: not installing a plugin whose main purpose is showing you a dashboard you could open in another tab. This article covers what the tag actually costs, how local hosting works and what it breaks, and how to measure the difference on your own pages rather than trusting a number from an article.
Verified August 2026. Script sizes and behaviours change; measure your own pages rather than relying on published figures.
What the tag actually costs
The cost has four components and only one of them is the file size everybody quotes.
A new connection. The browser must resolve a different domain and negotiate a secure connection before a single byte of the script arrives. On a mobile connection that handshake alone is a measurable fraction of a second, and it happens before any download begins.
The download itself, which is not trivial and is often followed by a second request as the initial script fetches more.
Parsing and execution on the main thread. While that runs, the browser cannot respond to taps. This is the component that hurts interaction responsiveness, and it is far more consequential on a mid-range phone than on the laptop you are testing with.
Ongoing activity. The tag does not stop when the page loads. It watches for events and sends data throughout the visit, competing with your own interface for the same thread.
Add a dashboard plugin on top and you are also loading its own scripts, its own admin assets and sometimes a second tracking snippet, for the privilege of seeing charts inside wp-admin.
The 6 options compared
1. GA Google Analytics
The minimal snippet plugin, free, and the lightest way to add tracking without hand-editing a theme.
Its entire job is placing the tracking code in the right place with sensible options about where and when, and it adds essentially nothing of its own to the front end. No dashboard, no admin widgets, no extra scripts, no second snippet. For a site whose owner reads reports in the analytics interface anyway, this is exactly the right amount of plugin, and the performance profile is as good as it gets while still using the standard tag. It has been maintained for a long time by a developer known for small, well-behaved plugins, which is a reasonable reliability signal for something you will install and forget.
- Adds front-end weight: Only the standard tag
- Local hosting: No
- Best for: Anyone who reads reports outside WordPress
- Watch out for: No dashboard at all, by design
2. Local hosting plugins
Plugins that download the analytics script and serve it from your own domain, which removes the external connection entirely.
The mechanism is straightforward: a scheduled task fetches the current script to your server, and your pages reference the local copy. You save the domain resolution and connection negotiation, the file benefits from your own caching and delivery network, and one fewer external party is involved in every pageview. The measurable win is real, particularly on mobile. The caveats are that the local copy must be refreshed regularly or you are running an outdated script, and that this arrangement is a workaround rather than an officially supported configuration, so verify your data is still arriving correctly after setting it up and again after any plugin update.
- Adds front-end weight: Reduces it
- Local hosting: Yes, that is the point
- Best for: Sites where the connection cost matters
- Watch out for: Keep the local copy refreshed
3. Site Kit
The official free plugin, with an admin dashboard and no licence fee at any scale.
It connects analytics, search performance data and other services, and shows summaries inside WordPress, which is the feature most people are actually paying commercial plugins for. Being free and first-party makes it the obvious comparison point before spending anything. On performance it is middling rather than bad: it installs the standard tag like everything else, and its admin components do not affect the front end for visitors. Where it earns its place is the search performance data, which the commercial analytics plugins generally do not provide and which is more useful to most site owners than another view of the same traffic numbers.
- Adds front-end weight: Standard tag
- Local hosting: No
- Best for: A free dashboard with search data
- Watch out for: Connects several services you may not need
4. Dashboard plugins
The commercial products whose value is the reporting interface, priced accordingly: from $99.50 in the best-known family, or $99 for the cheaper alternative.
It is worth being clear about what you are buying, because the tracking itself is free and identical. What the licence pays for is charts inside wp-admin, guided setup, event tracking configured without effort, and support. For a business owner who will genuinely look at those charts weekly and would never open the analytics interface, that is a reasonable purchase. For a developer who lives in the analytics platform anyway, it is a licence fee for a screen nobody visits. On performance they are neutral to slightly negative: the front-end cost is the standard tag plus whatever extra event tracking they add, and the admin weight only affects logged-in users.
- Adds front-end weight: Tag plus event tracking
- Local hosting: No
- Best for: Owners who will actually read the dashboard
- Watch out for: Paying for an interface you never open
5. Server-side and self-hosted analytics
The approach that removes the third-party script from the browser entirely, which is the only way to eliminate rather than reduce the cost.
Self-hosted analytics tools track visits from your own server or with a very small first-party script, typically a fraction of the size of the standard tag and with no external connection at all. Some record page views in your database with no client-side JavaScript whatsoever. The performance improvement is the largest available, and there is a privacy benefit in not sending visitor data to a third party. What you give up is the depth of the major platform, its integration with advertising products, and the ability to compare against a market everybody else uses. For a content site that wants to know which articles are read, this trade is frequently worth making.
- Adds front-end weight: Minimal or none
- Local hosting: Entirely
- Best for: Content sites wanting speed and privacy
- Watch out for: Less depth, and database growth
6. Script control and delayed loading
Not an analytics plugin, and the most effective single change available to most sites.
Performance plugins can delay third-party scripts until the visitor interacts with the page, which moves the analytics tag out of the critical loading path entirely. Load time improves, main-thread contention during the initial render disappears, and interaction responsiveness improves noticeably on mobile. The trade is genuine and worth stating: a delayed tag misses visitors who leave before interacting, so your session counts fall and your bounce figures change. Some site owners find that clarifying, since a visitor who left in two seconds without touching anything was arguably not a session. Others need the complete picture. Decide deliberately, and if you delay, note the date so you know why your numbers stepped down.
- Adds front-end weight: Defers existing weight
- Local hosting: Complementary
- Best for: Any site keeping the standard tag
- Watch out for: You will lose some very short visits
Comparison table
| Option | Price | External connection | Dashboard | Performance |
|---|---|---|---|---|
| GA Google Analytics | Free | Yes, standard tag | No | Best of the tag options |
| Local hosting plugins | Free plus paid | Removed | No | Better |
| Site Kit | Free | Yes | Yes, with search data | Neutral |
| Dashboard plugins | From $99 | Yes | Yes, polished | Neutral to worse |
| Self-hosted analytics | Free plus paid | None | Yes, in wp-admin | Best overall |
| Script delaying | Performance plugin | Deferred | Not applicable | Largest single win |
Measuring it properly
Everything above is reasoning about architecture. What matters is your site, and the test takes twenty minutes.
- Test with analytics disabled, then enabled. Same page, same conditions, three runs each, take the median. This is the only comparison that isolates the cost, and single runs vary too much to be useful.
- Throttle to a mid-range phone on a slow connection. On a fast desktop the difference is nearly invisible, which is why so many people conclude the tag is free. Your visitors are not on a fast desktop.
- Look at main-thread blocking time, not just page weight. Kilobytes understate the cost of a script that must be parsed and executed. Blocking time is the number that correlates with a page feeling slow.
- Count the requests in the network panel. Note every external domain the tag introduces, including any secondary requests it makes after loading.
- Compare field data before and after a change. Real user measurements from your search console tell you what visitors actually experience, and laboratory scores routinely flatter a site.
- Check your data still arrives after any optimisation. Local hosting, delaying and consent tooling can each break tracking silently. Verify in the realtime view the same day.
The decision that matters most is usually not which plugin, but whether you need the standard tag at all. A content site that wants to know which articles are read can often answer that with a self-hosted tool costing nothing in load time, and reserve the heavier platform for sites where advertising integration genuinely requires it.
Frequently asked questions
Does the analytics tag really hurt performance?
Measurably, especially on mobile. The connection setup and main-thread execution matter more than the file size, and both show up in interaction responsiveness.
Is hosting the script locally safe?
It works and it is a workaround rather than an officially supported setup. Keep the local copy refreshed on a schedule and verify data is still arriving after changes.
Should I delay analytics until interaction?
It is the largest single improvement available, at the cost of missing visitors who leave without interacting. Decide deliberately and note when you changed it.
Do dashboard plugins slow my site down?
Slightly, through additional event tracking. Their admin weight only affects logged-in users. The bigger question is whether you will read the dashboard you paid for.
Is self-hosted analytics accurate?
For traffic and content performance, yes, and often more so since it is less affected by blockers. It lacks the depth and advertising integration of the major platform.
What is the lightest way to add tracking?
A minimal snippet plugin with no dashboard, ideally with the script hosted locally and delayed until interaction. Or a self-hosted tool, which avoids the tag entirely.
The verdict
Ask first whether you need the standard tag at all. A content site wanting to know which articles are read can answer that with a self-hosted tool at almost no cost in load time.
If you keep it, use a minimal snippet plugin and skip the dashboard. The tracking is identical; a licence buys charts inside wp-admin, and the free official plugin provides those along with search data.
Then host the script locally and delay it until interaction. Together those two changes remove the connection cost and take the tag out of the critical loading path, which is the largest improvement available.
And measure it yourself, throttled to mobile, three runs each way. Every figure in an article like this one is architecture rather than evidence about your site, and the test takes twenty minutes.
