The analytics stack every cross-border seller quietly hates — and what a small EU challenger just exposed about it
If you run a Shopify store, an Amazon FBA brand, or a DTC operation selling into the EU, you already know the feeling: your analytics bill went up, your dashboard got busier, and your actual understanding of where revenue comes from got worse. GA4 is the default nobody chose, the consent banner is a tax on conversion rate, and the AI tools you pay for can’t see any of your traffic data because there’s no clean pipe between them. So when a solo builder launches a cookieless, EU-hosted analytics tool with an MCP server baked in, that’s not a niche Product Hunt curiosity — it’s a signal about where the whole measurement layer is heading for operators who sell across borders. Statable Analytics, built by Yevhen Skrypets, is worth dissecting not because it will replace your stack tomorrow, but because the design decisions inside it map almost perfectly onto the problems cross-border sellers will be forced to solve in the next 18 months.
What problem Statable actually solves (and which ones it doesn’t)
Strip away the launch-day framing and the product is three things stapled together: a cookieless web analytics tool, an EU-hosted data store, and an MCP server that lets an AI agent query your numbers directly. The maker’s own origin story is the most honest part of the pitch — he was inside Google Analytics daily, took a GA4 course, and realized he wasn’t the audience for the product anymore. That’s a sentiment I hear from sellers constantly, just phrased differently: “I don’t need 40 reports, I need to know which ad set made money last week.”
The stated design constraints are unusually specific, and they’re the whole thesis: collect as little as possible, no fingerprinting, and shape the product around European regulation. Statable is cookieless, EU-hosted in the Netherlands, and — per the maker — needs no consent banner for analytics. For anyone selling into Germany, France, or the Nordics, that last clause is not a feature, it’s a conversion-rate lever. Every consent banner you can legally remove is a few basis points of checkout completion you get back.
Then there’s the MCP server, which is the part most operators will skim past and shouldn’t. The maker describes it plainly: agents had no way into analytics, and the MCP server fixed that. You connect with one OAuth sign-in and choose whether the assistant can configure anything or only read. In practice that means asking Claude what moved last week and getting an answer from your own data, or describing a signup funnel in a sentence instead of clicking through a builder.
The funnel caveat you need to read twice
Here’s where I’d slow any seller down. In the comments, someone asked the obvious question — does cookieless still capture enough detail for meaningful conversion analysis? The maker’s answer is refreshingly blunt: for conversions within a single visit, yes — goals, custom events with properties, funnels, source, UTM, referrer. But for a journey spread over several days, it’s weaker. The visitor ID is a hash that rotates every 24 hours, so someone who reads on Monday and signs up on Thursday counts as two visitors, and the signup gets credited to Thursday’s source.
That single sentence should determine whether you care. If your product is an impulse buy — a $30 accessory, a viral TikTok Shop item — you’re fine. If you sell $400 furniture with a two-week consideration window, cookieless analytics will systematically undercount your upper-funnel and over-credit your lower-funnel. That’s not a Statable flaw, it’s physics of the cookieless model, and every vendor in this category will hit it.
How it stacks up against the incumbents you’re probably already paying for
Let me be the annoying blogger who names names, because “cookieless analytics” is a category now and you deserve a map.
Against GA4: the comparison isn’t close on complexity, and that’s the point. GA4’s problem for cross-border sellers isn’t accuracy, it’s that the tool was built for a world where a marketing analyst sits between the data and the decision. Most DTC operators don’t have that person. Statable’s bet is that a smaller, opinionated dataset plus an AI query layer beats a bigger dataset nobody opens.
Against Plausible and Fathom: these are the real incumbents for cookieless, privacy-first analytics, and both have years of head start, mature dashboards, and self-hosting options. Statable’s differentiator against them isn’t the cookieless part — it’s the MCP server and the EU-hosting-plus-no-banner combination aimed explicitly at European regulation. If you’re already happy on Plausible, Statable is not going to move you on reporting alone.
Against Matomo: Matomo is the enterprise-grade, self-hostable, GDPR-friendly option that agencies deploy for clients with strict data-residency needs. It’s heavier, more configurable, and more expensive to run well. Statable is the “I want this working in an afternoon” version of the same instinct.
Against the AI-native layer: tools like PostHog have been adding AI-assisted querying and product analytics depth for a while. Statable is narrower but arguably cleaner for the specific job of “let my agent read my traffic numbers without giving it write access to anything.”
Why Amazon sellers should care less than Shopify ones — but not ignore it
If you’re pure Amazon FBA, most of this essay doesn’t apply to you, and I’d rather say that than pretend otherwise. Your traffic data lives inside Amazon Seller Central and Brand Analytics, and no third-party web analytics tool can see inside that black box. Your search-term reports, your Helium 10 or Jungle Scout data, your PPC console — none of it flows through a JavaScript snippet.
Where Amazon sellers should pay attention is off-Amazon. If you run a DTC site alongside your FBA business — and increasingly you should, for margin and for customer data — or if you drive external traffic to listings via TikTok, influencer, or Google Ads, then the measurement problem is identical. And the MCP angle is genuinely interesting for Amazon sellers because so much of the operational work is already agent-shaped: pulling reports, reconciling ad spend, summarizing weekly performance.
For Shopify and DTC operators, this is squarely your problem. Your storefront, your checkout, your post-purchase flow — all of it is measurable, and all of it is currently measured badly.
Where the math breaks
Two things I’d flag before anyone migrates.
First, the 24-hour rotating hash means your multi-touch attribution is effectively single-touch within a day. If you’re running Meta Ads prospecting into a Klaviyo email flow into a retargeting campaign, the model will credit whichever touch happened on the conversion day. Your Meta ROAS will look better or worse than reality depending on timing, and you’ll make budget decisions on a distorted picture.
Second, the “no consent banner” claim deserves your own legal review, not blind trust. The maker’s framing is that the product is shaped around European regulation and cookieless, which is the standard argument for banner-free analytics. It’s a defensible position and Plausible and Fathom make similar claims, but “cookieless” and “no consent required” are not automatically the same sentence in every EU jurisdiction. If you sell into Germany specifically, get your DPO or counsel to sign off. This is not a knock on Statable — it’s a knock on treating any vendor’s compliance marketing as legal advice.
The MCP permission model is the part worth stealing
Buried in the comment thread is the most operationally interesting detail of the whole launch, and it’s the thing I’d want every analytics and ad-tech vendor to copy.
Someone asked whether read-only access still exposes raw visitor-level data — IPs, hashed IDs, individual sessions — or whether it’s aggregated only. The answer from the maker’s side: read access is aggregates only, visitor IPs aren’t stored, they’re used on arrival for geo lookup and to build a visitor hash, then discarded. The hash is salted daily and never exposed. No tool lists individual visitors or sessions. With read access, the agent sees totals — visitors and pageviews by page, source, country, device — and page paths, UTM tags, and custom event properties come back as your site sends them. Query strings are stripped from page URLs.
The follow-up in that thread is the line that should be printed on every AI-tooling datasheet: most MCP analytics tools just hand the agent full write access and hope for the best. That’s the read-versus-configure split, and it’s the detail that gets a tool approved by a security-conscious team rather than blocked.
Why this matters more than the dashboard
Cross-border sellers are about to hand a lot of operational surface area to agents. Reorder points, ad budget shifts, listing edits, customer service replies. The question that will separate tools that survive procurement from tools that don’t is not “how good is the model” — it’s “what is the blast radius if the model is wrong or manipulated.”
Statable’s answer — read access means aggregates only, no raw visitor data, daily-salted hashes, query strings stripped — is a template. If you’re evaluating any AI tool that touches your store, your ads, or your customer data, ask the same three questions: what can it write, what raw data can it see, and what’s the retention window on anything identifiable. The maker also gave the practical warning that follows from this: keep emails and customer IDs out of your UTM tags and event properties, because those come back as your site sends them.
That’s a genuinely useful operational rule regardless of which analytics tool you use. Audit your UTM conventions this week. I’d bet money you have at least one campaign passing an email address or a customer ID through a query string.
What cross-border sellers can borrow from this launch
Even if you never sign up, there are four transferable lessons here.
1. Consent-free measurement is a conversion feature, not a compliance chore. If you sell into the EU, every banner you can remove is revenue. The cookieless route has real analytical costs, but frame the decision as a trade against checkout completion, not as a pure legal question.
2. Aggregates-only is the right default for agent access. When you connect any AI tool to any business system — analytics, ad platforms, your Shopify Admin API, your Klaviyo account — start with read-only and prove you need write. The maker’s split between “configure” and “read” permissions is the pattern to demand.
3. Multi-domain and subdomain handling is table stakes for cross-border. A seller typically runs a .com storefront, a localized subdomain or ccTLD per market, a blog, and an app. The maker’s answer on this is worth noting: by default each domain or subdomain is its own site with its own dashboard under one account, but you can combine them by putting the same snippet on all of them and listing the hostnames in an allowlist. Anything else sending that snippet — staging, a copied site — gets dropped, and a Hosts report splits combined traffic by hostname. If your current analytics setup can’t separate DE traffic from FR traffic from US traffic cleanly, that’s a gap.
4. The free tier is aimed at a specific audience. Statable is free forever on .edu and GitHub or GitLab Pages. That’s a developer-and-student play, not a merchant play, and it tells you something about who the maker expects to adopt first. Cross-border sellers are a downstream beneficiary, not the primary target — which is fine, but calibrate your expectations on merchant-specific features accordingly.
Where my judgment says it falls short
Three honest reservations.
The reporting depth is probably thin for serious operators. The maker’s own framing is that he built the tool he wanted, which means it reflects one person’s priorities. If you need cohort analysis, LTV curves, or multi-touch attribution modeling, you’re looking at the wrong category entirely — you need something closer to Triple Whale or a warehouse-plus-BI setup.
The 24-hour hash is a hard ceiling. It’s an honest trade, disclosed clearly, but it means Statable cannot be your source of truth for anything with a consideration window longer than a day. Use it for traffic and on-site behavior; keep your revenue attribution elsewhere.
Compliance claims need your own verification. As noted, “cookieless” and “no consent banner required” are related but not identical claims, and jurisdiction matters. The maker’s positioning is explicitly European-regulation-shaped, which is a good sign, but it’s still your legal exposure.
What I’d watch / test next
This week, three concrete moves.
First, audit your UTM and event-property hygiene. Pull your last 30 days of campaign URLs and custom event payloads and check for anything personally identifiable — emails, customer IDs, order numbers. If you find them, fix the convention now, before you connect any agent to any analytics tool. This is free and it’s the highest-leverage thing on this list.
Second, run a read-only agent test against whatever analytics you already have. If you’re on GA4, see whether an MCP-capable bridge exists and whether it can be scoped to read-only. If you’re on Plausible or Fathom, check their API and agent story. The point isn’t to switch tools — it’s to find out this week whether your current stack can answer “what moved last week” in a sentence, or whether you’re still clicking through a builder.
Third, if you sell into the EU and run a consent banner, quantify what removing it would be worth. Look at your banner’s accept rate and your checkout completion rate for users who accepted versus those who didn’t, or run a geo test. If Statable’s no-banner claim holds up under your counsel’s review, that number is your business case — and it’s a much better reason to evaluate the tool than the MCP hype.
I’d also watch whether the maker documents the daily-salted-hash and stripped-query-string details on the site itself rather than only in a launch comment thread. One commenter made exactly that point: that config belongs in the docs, not buried in a thread. How a solo founder responds to that kind of feedback in the first 30 days tells you more about the product’s trajectory than day-one ranking ever will. Statable hit #2 on launch day, which the maker called surreal — but rankings fade and documentation doesn’t. That’s the thing to check back on in a month.
