← back to product
Pillar 04

Attribution
OS.

iOS killed last-click. AI search broke Google Analytics. Attribution OS uses first-party ID stitching and probabilistic modeling to tell you what actually drives revenue.

Attribution OS illustration — multi-touch customer journey with AI search, dark social, and true attribution* the real buyer journey — 27 touchpoints avgAI SearchChatGPT recDay 1LinkedInsaw sponsoredDay 3Dark SocialSlack shareDay 7Googlebrand searchDay 12Conversion ✓$2,400 contract signedLast-click (WRONG)Google: 100% creditAI Search: 0%LinkedIn: 0%Attribution OS ✓AI Search38%LinkedIn28%Dark Social18%Google16%attribution model⚠ platform overreportingMeta + Google + LinkedInall claim 100% of this saletrue ROAS per channel4.2x blended* incrementality testing runs automatically

34%

higher conversion rate from AI referrals vs organic

3x

platform overreporting ratio (Meta + Google + LinkedIn)

27

avg touchpoints before a B2B purchase

92%

attribution accuracy with first-party stitching

features

Know what actually
drives revenue.

01

First-party ID stitching

Attribution OS connects anonymous sessions to known users across devices using first-party signals — email hash, phone hash, logged-in state. Cookieless by design, not by compromise.

02

AI search attribution

When a visitor arrives from ChatGPT, Claude, Gemini, or Perplexity, Attribution OS captures the AI referral source and attributes it properly — not as direct traffic, not as organic.

03

Probabilistic multi-touch

Attribution OS builds a probabilistic model of each buyer's journey using first-party events. Every touchpoint gets credit proportional to its actual influence — not just the last click.

04

Platform overreporting correction

Meta claims 7-day click attribution. Google claims 30 days. Both over-count. Attribution OS runs incrementality tests to measure true lift per channel — and shows you what is real.

05

Dark social tracking

Podcast mentions, Slack shares, newsletter clicks — Attribution OS captures conversion events from channels that traditional analytics marks as direct. The dark funnel gets light.

06

Revenue-level reporting

Attribution OS connects marketing spend to actual revenue — not MQLs, not sessions. CAC, LTV, payback period, and true ROAS per channel, per campaign, per creative.

the overreporting problem

Every platform claims more credit
than it deserves.

Meta defaults to 7-day click + 1-day view attribution. Google defaults to 30-day. LinkedIn defaults to 28-day. If a customer sees a Meta ad on Monday, a Google ad on Wednesday, and converts on Friday — all three platforms claim the full conversion. Your blended ROAS across all channels is likely 2-3x your real number.

Meta Ads

7-day click 1-day view window

Reported ROAS

4.8x

True ROAS

3.1x

overreporting by 55%

Google Ads

30-day click 1-day view window

Reported ROAS

6.2x

True ROAS

4.2x

overreporting by 48%

LinkedIn Ads

28-day click 7-day view window

Reported ROAS

3.4x

True ROAS

2.4x

overreporting by 42%

faq

Common questions.

How does Attribution OS handle iOS 14+ tracking changes?

Attribution OS uses server-side tracking via our conversion API — bypassing browser-level restrictions entirely. We match events server-to-server using first-party identifiers, not third-party cookies.

What is incrementality testing and do I need it?

Incrementality testing measures the true lift from a channel by holding out a control group. If you turned off Meta completely, how many conversions would you still get organically? That gap is your true Meta lift. Attribution OS runs these tests automatically on your highest-spend channels.

Does Attribution OS require a developer to install?

No. Installation is a single JavaScript tag plus optional server-side event forwarding. Setup typically takes 30-45 minutes. For e-commerce platforms (Shopify, WooCommerce), we have one-click integrations.

How does AI search attribution work technically?

We maintain a regularly updated list of referral domains for all major AI assistants. When a session arrives from a known AI domain, we tag it as AI_REFERRAL in our event schema. This data is available in all Attribution OS reports.

Can Attribution OS integrate with our existing data warehouse?

Yes. Attribution OS exports to BigQuery, Snowflake, and Redshift. Raw event data and attribution models are available as tables you can query directly. Enterprise plans include custom schema support.

works best with

* your current ROAS data is probably wrong

See what actually
drives your revenue.

14-day free trial. No credit card. Cancel anytime.

start free trial →