GA4 alternatives and comparisons
Peeko vs Cloudflare Web Analytics when free is enough and when it is not
By Emily Parker · Nov 11, 2026 · 10 min read
Both tools answer the same first question: how many people came to your site, and from where? They differ on what happens after that. This post reads Cloudflare's documentation and product pages as of November 2026, states Peeko's side from what we know about our own product, and says plainly where Cloudflare is the better choice. We did not run the two tools side by side on one site, so nothing here is a measurement comparing their counts.
What each tool is
Cloudflare Web Analytics describes itself as privacy-first analytics that helps you understand the performance of your pages as your visitors experience them. It is available on all Cloudflare plans, and the product page leads with "Free website analytics." It works with or without the Cloudflare proxy. If your site is proxied, Cloudflare can turn it on for you. If it is not, you paste a JavaScript snippet before the closing body tag.
Peeko is one script tag with a site key. It sets no cookie and needs no configuration once it is on the page. Beyond page views it records five kinds of click with no code (outbound, phone, email, whatsapp and form), takes named events, and holds goals, funnels and revenue. A separate server side path counts AI crawlers, which do not run JavaScript.
Cost and limits, in numbers
Cloudflare's blog post introducing the product says, "Best of all, our analytics is free." The same post says there are no caps based on traffic volume, but the limits page does set other caps:
- Sites not proxied through Cloudflare: 10 per account. Proxied sites have no limit.
- The dashboard shows aggregate data for up to 1,000 websites at once.
- Rules, which let you track specific sites or paths, apply only to proxied sites and depend on your plan.
Source: Cloudflare, Limits · Cloudflare Web Analytics docs, 2026-11. Retrieved Nov 8, 2026.
A Free plan allows 0 rules, which means that on a Free plan Cloudflare injects the snippet on all subdomains of a proxied site. If you want to track one path and ignore another, that needs a paid Cloudflare plan.
Peeko has one plan on a slider from 100,000 to 10,000,000 events a month. As of November 2026, the pricing page lists 9 dollars a month for 100,000 events, and the range runs to 199 dollars. Yearly billing is two months free. Every feature is on every size, with unlimited sites and unlimited team seats. The trial is 14 days with no card and covers one site. An event is a page view, an outbound click or a custom event, counted over the UTC calendar month, and nothing stops recording if you pass your limit. AI crawler hits are never billed.
So the price gap is real: zero against 9 dollars at the smallest size. The question is what the 9 dollars buys, and the rest of this post is about that.
Cookies and what identifies a visitor
Both tools set no cookie. Cloudflare states that it does not use "any client-side state (like cookies or localStorage) for analytics purposes," and that it does not track users over time "via their IP address, User Agent string, or any other immutable attributes."
Peeko's approach is different in one respect, and you should know it before choosing. A visitor is the first part of a one way keyed hash of their address and browser string. The address itself is never written to a database or a log. The hash behind the People number mixes in the date and is re-salted daily, so nobody can be followed from one day to the next. A second hash without the date tells a returning visitor from a new one and strings page views into a journey. That second hash is the identity behind goals, funnels and revenue.
If your priority is a tool that keeps no per-visitor identity of any kind, Cloudflare's stance is the stricter one. If you need to know whether the same person came back and what they did next, that second hash is what makes it possible.
What gets counted
Cloudflare's high-level metrics page defines a visit as "a page view that originated from a different website or direct link" and a page view as "a successful HTTP response with a content-type of HTML." The blog post explains the design: Cloudflare counts page views that come from a different site instead of counting unique IP addresses. A visit here is a unit of arrival, not a person. The FAQ adds that the beacon script "will fire on every pageview," and that single page apps send a measurement on each route change.
Two things from the same docs are worth holding in mind:
- The beacon can be blocked. The FAQ says "Cloudflare is aware that the analytics beacon is blocked" by ad blockers, and that Cloudflare's edge analytics cannot be blocked because it measures every request received. The edge version is described as available to Pro, Business and Enterprise plans. Peeko's tracker is also a script tag, so we would not promise that it is exempt from blockers either.
- On proxied sites, the automatic setup does not inject the beacon if the
Cache-Controlheader ispublic, no-transform. The get started page says Web Analytics will not work in that case unless you add the snippet by hand.
On bots, the dimensions page lists an "Exclude Bots" filter that excludes bot traffic from the dataset. The pages we read do not say how a request is classed as a bot, so we cannot say how it compares to anything else.
Sampling and retention: the largest difference
This is the point that matters most if you plan to look backward. The Cloudflare FAQ states: "We retain unsampled beacon data for the past 7 days, after this point data is aggregated down to around 10%." It also says that the ingestion pipeline does not apply sampling, so every beacon received is recorded, and that you can currently access data for the previous six months. The GraphQL API exposes a sampleInterval field that tells you which level of sampling a query used.
Put those together. For the last week you see everything that arrived. Older than that, you see an aggregate at around a tenth of the original volume, and only for about six months. The pages we read do not say how the dashboard scales the aggregate back to a total on a low traffic site, so if exact counts of old weeks matter, check them against a number you know, such as signups.
Peeko keeps page views, events and crawler hits for three years by default. The site owner can set that window anywhere from 30 days to ten years. A goal is a definition and not a counter, so a goal you write today counts over months before it was written. That is what makes a question like "how many people read this post and then signed up" answerable after the fact, and our post on a two step funnel built from traffic you already recorded walks through one.
Events, goals and revenue
The Cloudflare pages we read list these dimensions: country, host, path, referer, device type, browser, operating system, site, bot exclusion and navigation type. Referer host is on the dashboard, and referer path is available only through the GraphQL API. None of the pages we read describe goals, funnels, custom events or revenue. Cloudflare's design, which does not track users over time, also means a funnel that follows a person is not what the tool is built for.
Peeko's side, for comparison:
- A funnel holds two to five steps, with a completion window of 1 hour, 1 day, 7 days or 30 days, covers at most 90 days, and follows a person rather than a visit. A site can hold ten funnels and twenty goals.
- Custom events are a
data-peeko-eventattribute or awindow.peeko.eventcall, with one name and one optional string. - Revenue arrives by a one line
window.peeko.revenuecall or a verified Stripe webhook, and the two totals are never added together. Peeko never holds a Stripe API key, only the webhook signing secret.
If your site has nothing to sell and no signup, you may never need any of that.
AI crawlers and AI referrals
A JavaScript beacon only fires in a client that runs JavaScript, and crawlers do not. That is why Peeko counts crawlers on a separate server side path, with the @usepeeko/bot-track middleware: ESM only, Node 20 or newer, zero runtime dependencies, under 3 KB gzipped. It matches user agents against a curated list of 55 agents from 21 companies and files each hit under AI answers, Search indexing, AI training or Other AI bots. No address and no visitor hash exist on that path.
Visits that arrive from ChatGPT, Claude, Perplexity, Copilot and Gemini are counted from ordinary page views by referrer, with no middleware. You could read the referer list in Cloudflare's dashboard yourself and pick those hosts out by eye. Our post on counting visits from AI assistants by referrer shows what that grouping looks like. The Cloudflare pages we read do not describe a per-site count of AI crawler fetches.
Where Cloudflare Web Analytics is the better choice
Pick Cloudflare Web Analytics if any of these describe you:
- You want to pay nothing. Free against 9 dollars a month is a real difference, and a hobby site or a side project with no revenue may not need more than free.
- Your site is already proxied through Cloudflare. The automatic setup is on by default for proxied sites that previously used Browser Insights, so there may be nothing to install.
- You care about Core Web Vitals. The docs say three metrics are measured: Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift, each rated against Google's thresholds. That is a performance product as much as a traffic one. We make no claim about Web Vitals for Peeko, so if they are the reason you are looking, start with Cloudflare.
- You want no per-visitor identity at all. As described above, Cloudflare's stance is stricter than a daily salted hash.
- You only need the last week or so in detail. Seven days of unsampled data is plenty if you check the dashboard after each launch and move on.
Cloudflare Web Analytics is also a sensible second tool next to Peeko. They do not conflict, and the free one adds Web Vitals.
Where Peeko fits
Peeko fits when you need to look back past a week, when you have a goal such as a signup or a purchase, or when you want to know which AI crawlers fetched your pages. It also fits when more than 10 sites are not proxied through Cloudflare, since Peeko has unlimited sites. An MCP server at usepeeko.com/api/mcp and a command line (npx @usepeeko/cli) return the same numbers, which suits people who ask their coding agent how traffic is doing. Our post on reading traffic from a terminal with no dashboard open shows the commands.
How to decide
Start from the question you will ask in three months. If it is "how many visits did the site get," Cloudflare answers it for free and the 10% aggregate is likely fine for a rough trend. If it is "which page led to a signup" or "did the crawler list change," you need records that last and a way to join steps, and that is what the nine dollars is for.
If you are not sure what is on your site now, the free analytics checker shows which of 18 analytics tools a page loads. It fetches one page and reads its HTML, so it cannot see tags that load later, such as those added by a tag manager or held back by a consent banner. For how this comparison fits next to the other tools, see our reads of Peeko vs Plausible and Peeko vs Google Analytics 4.
All Cloudflare figures above are from Cloudflare's documentation as of November 2026 and can change. Check the linked pages before you decide.