Privacy-by-design analytics · No consent required · GDPR · ePrivacy
Between 25% and 45% of your sales never reach your analytics tool. They happened, you were paid for them, and the report you make decisions on does not count them. You are switching off campaigns that work and you do not know it.
Nothing to install · No commitment
1,000 real sessions. 1,000 measured: channel, campaign, keyword, landing page and sale.
“We have far too much direct. It is direct traffic that should not be direct.”
There is no mystery channel bringing in half your business. There is a cookie banner standing in front of your measurement.
“We know Meta contributes more than we can see, but we do not have the data.”
Discovery channels look worst, because they sit furthest from the final click that does get measured.
“We want to test a new channel, but we would not know whether it worked.”
A test you cannot measure is a test you will cancel for the wrong reasons.
Dreamplace Hotels recovered +30% more traffic vs GA4 and closed a 15–20% gap in sales attribution against their CRM. Palladium Hotel Group recovered 35% of unattributed bookings and improved Display CPS by +165%.
Same campaign · same spend · same month
The budget you were about to cut was your best-converting channel.
Every conversion counted means the channels that work stop hiding inside “direct” and inside traffic that never consented.
The requirement
Reconstructing a customer journey requires identifying a user over time. Cookies, fingerprinting, server logs: the method does not matter. Even anonymised, it requires consent.
The gap
Between consent rejection (20–35%), domains blocked by filters (10–15%) and visitors leaving before a heavy pixel ever fires, the hole compounds.
The bias
People who accept cookies do not behave like people who reject them. You are not measuring half your business: you are measuring one specific half and deciding as if it were the whole.
SealMetrics does not reconstruct journeys. It records aggregate hits that are unconnected to each other, with no user identifier, no cookies and no profiling. Since nobody is identifiable now or later, no consent is needed to measure.
The pixel fires before the banner and measures 100% of your sessions, sales and events. Real data: we do not model and we do not sample, because there is no gap left to fill with statistics.
The shift in mindset is this: you stop analysing people and start analysing products and campaigns. Which product gets viewed, added to cart, entered into checkout and bought — by channel, campaign, keyword, landing page and any property you choose to tag.
Palladium Hotel Group
The data SealMetrics delivers is agnostic, impartial and neutral. There is no black box.
Toni Andújar · Digital & Direct Sales Director
Dreamplace Hotels
The case that convinces the finance director: the number in the marketing tool and the number in the CRM stopped contradicting each other.
Almost every tool is privacy by policy: it collects the personal data and then promises to behave well with it. We never collect it. The data you would need in order to breach someone’s privacy does not exist in our database, so there is nothing to promise, nothing to anonymise afterwards and nothing to leak.
Neither first nor third party. No localStorage, no sessionStorage, no storage of any kind on the visitor's device.
No user ID, no client ID, no anonymous pseudonym. Two visits from the same browser are not recognisable as such, now or in two years.
No canvas, no installed fonts, no header combinations. Fingerprinting is explicitly ruled out, not merely unused.
Nothing we collect on your site is joined with any other. Your data does not feed an advertising graph because we do not have one.
Hosting, database and backups in Dublin. No international transfers depending on an adequacy framework that is currently under political and judicial dispute.
CNIL self-assessment, UK PECR self-assessment, subprocessor list and data subject rights procedure. Auditable without asking us.
Our architecture has passed the legal audits of Acciona, 3Cat, UNICEF and Desigual.
Measuring without consent has a price and we do not hide it in the small print. If what you need is in the left-hand column, we are not your tool and we will tell you so on the first call.
SealMetrics against GA4’s gtag.js and against the Adobe Launch + AppMeasurement chain. Same reference sites, same methodology, July 2026.
Weight transferred over the network · gzip
155× lighter than both. The tracker fits in a single TCP packet and arrives on the first round trip. The other two need 4 or 5 round trips just to download.
JavaScript the device has to parse · uncompressed
Parsing has no CDN. Google and Adobe serve from excellent edges, but half a megabyte of JavaScript is parsed on the visitor’s phone. On a low-end handset that is 0.8 to 1.5 s of CPU before any hit can fire.
Time until the hit leaves the device · fibre, installed in head
And these are their best figures: they exclude waiting for the banner. That window between load and hit is exactly where a visitor who leaves becomes invisible traffic.
| SealMetrics | GA4 | Adobe Analytics | |
|---|---|---|---|
| Hit transport | sendBeacon · survives page close | Beacon type | Image GET · cancelled on exit |
| Consent in the EU | Not required, by design | Yes · Consent Mode models the gap | Yes in most installations |
| Traffic measured in parallel | Baseline | 25–45% less (our client sample) | 25% less (30 days dual-tagged, field) |
The Adobe figure is the strongest evidence: same site, same visits, pageviews against pageviews, a full month, and a constant 25% difference — the signature of a structural cause, not an incident. Full methodology and limitations →
1.1 KB
A single TCP packet. No dependencies, no tag manager required.
Real time
Not in one report: across the whole platform. You know to the second when the last hit was processed.
Black Friday too
Real time guaranteed by SLA at peak. You optimise at 10:01 in the morning, not the next day.
100% of traffic
No consent, no cookies, no banner in the way.
Zero modelling
No modelling. No sampling. No estimates. If it is in the report, it happened.
+25% vs Adobe or GA4
Measured in the field, 30 days dual-tagged on the same site.
The whole funnel
View product, add to cart, begin checkout, purchase. And whatever event you can think of.
Any property
Size, model, brand, category, colour, weight, price range. Whatever exists in your business.
From data to why
A product that gets viewed, added and not bought: that is not a metric, it is a diagnosis.
Robust API
Everything you see in the interface is available over the API. No exceptions.
BigQuery · Looker Studio
Native connector. Pick the account and it syncs.
MCP + LENS
Connect your data to Claude, ChatGPT or Cursor. Or use LENS, our private AI, and keep the data here.
From data to why.
In the head, or with the Shopify, Magento or PrestaShop plugin. Your current UTMs work as they are: you do not have to retag anything.
The first block measures pageviews and entry channel. The second records the sale with everything you want to know about it: amount, currency, payment method and the full product line. That second block is what turns analytics into campaign decisions.
You can add and remove events mid-season: since there is no journey trail, there is nothing to break.
1 · Measurement pixel
<!-- SealMetrics Analytics --> <script src="https://t.sealmetrics.com/t.js?id=YOUR_ACCOUNT_ID" defer></script>
2 · Conversion pixel
sealmetrics.conv('purchase', 149.99, { currency: 'EUR', payment_method: 'credit_card', items: [ { product_name: 'Product A', price: '19.99', quantity: '2', category: 'Sneakers', brand: 'Brand' }, { product_name: 'Product B', price: '29.99', quantity: '1', category: 'Accessories', brand: 'Brand' } ] })
Every key you add to the object becomes a dimension you can cross against channel and campaign. Brand, colour, size, price range: whatever exists in your catalogue. No order IDs and no customer IDs — the pixel does not accept personal data by design.
3 · Or not even that: a prompt
Connect Claude to the SealMetrics MCP. Server:
https://mcp.sealmetrics.com/mcp
Then create my account, generate the pixels I
need and build the reports in real time.Paste it into Claude, ChatGPT, Cursor, Codex, Copilot, Windsurf or any MCP client. It creates the account, generates pixels tailored to the events and properties you need, and leaves the reports built. No signup form and no sales call.
1.1 KB · DEFER · SENDBEACON · NO COOKIES · NO CONSENT
One client saw TikTok Ads as almost flat in GA4. With 100% of the data, the real sales appeared along with a discovery weighting far above what was reported. Spend moved accordingly.
This Shopping campaign sells headphones and does not sell jackets. Exclude the second, raise bids on the first. The data sits at the level of whichever property you tagged, not just the channel.
Products viewed a lot, added to cart and never bought: that is a price problem or a product-page problem. It only surfaces when the funnel is complete.
“Where am I burning budget?” — “DemandGen: 9,230 clicks, 2 sales. Pausing it and moving spend to PMax_Catalog recovers it.” An answer, not the report you have to go find it in.
Annual contract. We only count the events you tag, so your GA4 figure is not comparable: you will start well below it. Seasonal peaks carry no penalty.
For context: GA360 starts at $150,000 a year. Adobe Analytics at $100,000.
We tell you how much traffic and how many sales you are losing today, using your own data. We review your measurement plan and our technical team talks to yours if needed.
Request your free auditNo credit card and no form. One prompt in your assistant creates the account, generates the pixels and builds the reports. Once you pass the million, the Growth plan activates.
Start freeDo you trust that data?
For how much longer?