The Meta paradox
The pixel reports 340 conversions. The CRM records 180. Both teams defend their own collection boundary.
Platform overview · one measurement pipeline
Your analytics stack
is only as good as
the inputs it sees.
SealMetrics connects aggregate traffic, recorded revenue, decision-ready reports and AI supervision in one EU-hosted platform. No stitched-together evidence. No modeled fill-ins presented as observation.
Cookieless capture · declared last-click attribution · EU-hosted in Dublin
Platform signal mapDefined inputs
01 / CaptureEligible aggregate eventsObserved
02 / AttributeRevenue under a declared modelReconciled
03 / UnderstandNine reports + LENS AIGrounded
04 / ActivateAPI · MCP · BigQueryPortable
One evidence chain · from event to decision
The problem before the product
Four systems can agree on the meeting.
They still disagree on the number.
GA4, ad platforms, your agency report and the CRM can each be internally consistent while measuring a different slice of reality. The cost appears when budget decisions depend on a baseline nobody can reconcile.
The Direct bucket
Consent-shaped visits lose acquisition context, so productive campaigns can reappear as Direct or none.
The reconciliation loop
Marketing, finance and the agency spend another 90 minutes aligning totals instead of moving budget.
What SealMetrics is
One platform for complete-data analytics.
SealMetrics is a cookieless, consentless web analytics platform for eCommerce that captures 100% of traffic and attributes revenue using last-click on the full dataset. GDPR-compliant by architecture, hosted exclusively in the EU.
The useful claim is not “more data” in isolation. It is a reported total you can compare with backend revenue, under an attribution model your team can name.
The architectural shift behind that claim — unmodeled data, no consent dependency, output an AI agent can read — is set out in modern web analytics.
Four layers · one evidence chain
From first signal to
a decision you can defend.
Each layer uses the same defined inputs. That removes the hand-off where collection, attribution and reporting usually start telling different stories.
Cookieless aggregate measurement
Observe eligible events without analytics cookies, persistent visitor IDs or fingerprinting.
Revenue under a declared model
Connect recorded outcomes to channel, campaign and creative using last-click on the complete dataset.
Reports plus LENS AI
Inspect nine reporting surfaces or ask defined metrics a plain-language question.
API, MCP and BigQuery
Move the same evidence into your warehouse, BI layer and supervised AI workflows.
The product, for real
Not another dashboard.
A tighter evidence loop.
The interface is built around the decisions an eCommerce team has to make now: which channel deserves budget, where the funnel leaks and whether the reported revenue matches the store.
01 / ATTRIBUTION
Revenue attribution that names its assumptions.
Channel, campaign, ad set and creative totals use last-click on observed aggregate events. No per-user profile is needed to rank recorded outcomes.
REVENUE / LAST 30 DAYSIllustrative data · not a live account
02 / LENS AIDEFINED METRICS
LENS answers against defined metrics.
Ask why revenue moved, which products leak value or what changed yesterday. Inspect the metric and source before the answer changes spend.
GROWTHWhich products are leaking revenue?
77% of premium e-bike carts do not convert. Review product-level financing.
COSTWhere is budget being burned?
Illustrative data · not a live accountDemandGen: 9,230 clicks, 2 sales. Compare the spend with PMax_Catalog.
03 / ACTIVATEAPI · MCP · BIGQUERY
Your data leaves through documented doors.
Use the REST API, native BigQuery export, webhooks or the MCP server. LENS can run with your own model key; managed Private AI is available by plan.
GET /v1/revenue?group=channel
200 OK
{
"model": "last_click",
"rows": 1284210,
"sampled": false
}Nine core reports
Every report uses
the same inputs.
Sources, pages, funnels and revenue do not live in separate products. LENS sits above these reporting surfaces as an assistant, not as an invented tenth source of truth.
Overview
Headline totals and the timestamp of the last observed hit.
Evolution
Metrics over time, with period-against-period context.
Sources
Channel, campaign and referrer totals on the complete dataset.
Pages
Pageviews, landing pages and content groups.
Conversions
Conversions, microconversions and last-click revenue by channel.
Funnel
Step-level drop-off without sampling thresholds.
Geography
Country-level totals derived without storing IP addresses.
Devices
Device, browser and operating-system totals.
Properties
Products, baskets, AOV and your own event properties.
Published customer evidence
The product has to survive
a real comparison.
Palladium Hotel Group and Dreamplace Hotels compared SealMetrics with their previous measurement and published the resulting figures.
40%
of inbound traffic previously had no attribution
35%
of GA4 bookings had no channel assigned
+165%
Display Cost-per-Search on DV360 after the change
“The data SealMetrics delivers is agnostic, unbiased and neutral. There is no black box.”Toni Andújar · Digital & Direct Sales DirectorRead the full case
+30%
more traffic measured than Google Analytics
15–20%
sales-attribution gap against the CRM, closed
“It is the tool that gives us the real data — and the one we make decisions with.”Eduardo Martin · Analytics & CampaignsRead the full case
Direct connectors and open interfaces
Keep the stack.
Replace the missing evidence.
Start with a native commerce connector where available, or implement the documented event contract. Keep GA4 during the comparison period and change the stack only after the backend confirms the result.
See all integrationsShopifyWooCommerceMagentoPrestaShopOpenCartWordPressNext.jsReactBigQueryMCP serverREST APIWebhooks
A fair product test
Do not trust the product page.
Trust the reconciliation.
Run SealMetrics beside the current stack, define the same events and compare both reported totals with the revenue your store recorded.
Keep your current stack · compare the same period · decide from backend evidence