Every click, session and order — deduped, attributed, and cross‑referenced against what the platforms tell you. The second opinion your P&L has been missing.
Most ecom brands stall between $5M and $20M because they push spend up on attribution data that breaks down at volume. Three numbers most operators only learn after the damage is done.
claiming credit for organic intent
vs causal lift, under last‑click
touch 3+ channels before checkout
Channel‑by‑channel forensics — branded Google, email & owned, paid social — and the customer journey behind a single order.
| Channel | Spend | Revenue | True ROAS | vs Platform |
|---|---|---|---|---|
| Google Ads | $89,240 | $313,019 | 3.51× | −0.7× |
| Meta | $74,120 | $212,803 | 2.87× | −1.1× |
| TikTok | $31,800 | $130,998 | 4.12× | +0.3× |
| — | $57,210 | 8.45× | +2.1× | |
| $18,330 | $33,727 | 1.84× | −1.1× |
Google, Meta, TikTok, Pinterest — each tracks in isolation, each counts the same order, each takes credit. The aggregated number in your spreadsheet is louder than reality.
Omnitrail runs your data through five attribution models in parallel. Compare, cross‑reference, and let the AI agent pick what actually drives incremental revenue.
A single tag captures every event, session, and order across your storefront — server‑side where it counts.
Five models run in parallel on deduplicated data. We compare what the platforms claim against what actually drove the sale.
The corrected signal feeds your ad agents so spend moves to what works. Forecasts refresh continuously.
A single tag captures sessions, orders and LTV across platforms — then cross‑references them so nothing is counted twice.
Server‑side Conversions API for Meta, Google, TikTok, Pinterest, Snapchat. Identity stitched across devices. 87% of sessions resolved to known users; the rest fall back to probabilistic matching.
Linear, First‑Click, Last‑Click, Data‑Driven, AI‑Powered. Compare side by side, never lock in one lens.
Every credit split is calculated in parallel. Export a model comparison for board decks. The Data‑Driven and AI‑Powered models run on your rolling 90‑day window, updated every four hours.
Campaign‑level true ROAS and CAC feed directly back to your ad accounts — so the next dollar goes where the last one worked.
Integrates with Meta Ads Manager, Google Ads, and TikTok Ads APIs. One‑click budget reallocation or full automation handoff to the AI agent. Rules respect your guardrails.
An AI agent that reads the real numbers — not platform‑reported ones — and rebalances daily within your guardrails.
Works from the corrected attribution signal, not the pixel. Suggests bid adjustments, creative swaps, and audience expansions. You approve the first seven actions; after that it earns trust.
Continuous forecast of revenue, CAC, and ROAS by channel — with confidence bands, not marketing copy.
Monte Carlo simulation on your corrected attribution. P10 / P50 / P90 bands. Reforecasts nightly, surfaces when actuals drift outside the 80% interval so you can step in before the damage shows up in a monthly review.
Every step reads the same corrected signal — no silos, no re-exports, no reconciliation. One number, one ledger.
The stack is closed: what the tag captures, the models measure, the agents act on, and the forecast anticipates — all from the same record. No second pipeline, no attribution team stitching numbers at month-end.
Your ad spend on the left, the revenue it actually drove on the right, and the attribution models doing the accounting in between. Hover a channel to isolate its path.
GA4 and platform pixels each track in isolation. They can't dedupe across channels — so four platforms end up claiming credit for the same order. Omnitrail sits above them, ingests all of it server‑side, and produces a single corrected signal that reconciles every claim against what actually happened in your store.
About eight minutes for Shopify or WooCommerce: install the app, map your ad accounts, and we backfill 90 days. Custom storefronts are a single JavaScript tag plus server‑side webhook. No engineering sprint, no data‑team handoff.
Yes. Our primary tracking runs server‑side through Conversions APIs. First‑party identity stitching keeps roughly 87% of sessions resolved to known users; the remainder fall back to probabilistic matching. We don't rely on third‑party cookies or pixel fingerprints.
You can, though we'd recommend not. Every model has a blind spot. Running five in parallel lets you cross‑reference — and the AI‑Powered model itself is built from the agreement between the others plus incrementality signal from your holdouts.
Omnitrail exposes a clean API and pushes corrected events to your warehouse (BigQuery, Snowflake, ClickHouse). Most teams leave existing dashboards in place and use Omnitrail as the trusted second opinion for decisions.
Tiered by monthly tracked orders. Starter plan starts at $499/mo for up to 2,000 orders. Talk to us if you're above 50k/mo — we do annual pricing with a dedicated implementation engineer.
A 25‑minute call. We'll plug a read‑only connection into your ad accounts and one storefront, and show you exactly how much of your reported ROAS is noise.