Most 'data-driven' marketing teams only have one or two of these layers wired up. They make decisions confident they're using data, when really they're using a slice of it.
There are four. Most teams have ad-level performance and some kind of attribution. That's it. The behavioral layer is missing or bolted on. The enriched CRM layer is in a different tool that nobody opens. And the four don't talk to each other.
Here's what the four layers are, why each one alone is incomplete, and how to tell which one your stack is missing.
Layer 1: Ad-level performance
- Ad-level performance data
- Campaign, ad set, ad, and creative-level performance pulled from ad platforms like Meta, Google, TikTok, and YouTube.
This is the layer everyone has. Spend, impressions, clicks, CPM, CPC, CTR, in-platform ROAS. You see it inside every ad manager.
What it tells you: which creative platforms reward, which budgets are pacing, which audiences are firing.
What it can't tell you alone: whether any of those clicks turned into revenue. In-platform attribution is a sales pitch the platforms make to themselves. Meta will tell you it drove the conversion. Google will tell you the same. TikTok will too. They can't all be right.
Layer 2: Behavioral data
- Behavioral data
- A record of what a person actually does on your site or funnel: every scroll, click, video view, quiz answer, form interaction, and page visit.
Most marketing teams have this layer in pieces. They have a heatmap tool. They have form submissions. They have a video player that maybe reports plays. Stitched together, sort of.
Native behavioral data is different. It's collected at the platform level, not bolted on. When the funnel lives in the same system that does attribution and lifecycle, every event is captured the moment it happens, attached to the same identity, available everywhere downstream.
FlowOS does this natively because the funnels live in FlowOS. There is no integration step between 'user watched 80% of the VSL' and 'send them to the high-intent sales sequence.' The behavior is the trigger.
Layer 3: Enriched CRM data
- Enriched CRM data
- Downstream revenue, lead status, deal stage, lifetime value, and lifecycle stage from your CRM and ESP: Salesforce, HubSpot, ActiveCampaign, Klaviyo.
Layer 3 is what makes attribution honest. It's where you find out which leads actually closed, what they paid, and whether they stuck around.
Most marketing teams have this data. It just lives in a CRM that nobody on the paid team ever opens. The result: paid optimizes for lead volume, sales optimizes for closed-won, and the two don't align until quarterly business review when someone notices CAC payback is broken.
Wired up properly, layer 3 closes the loop. You know which ad drove the lead, which behavior signaled intent, and which deal closed at what price. You can finally optimize for closed-won revenue instead of form fills.
Layer 4: Multi-touch attribution
- Multi-touch attribution
- A model that credits every touchpoint in a buyer journey, not just the first or last click, using first-party server-side data with up to a 180-day lookback.
Attribution is the layer that ties the other three together. Done right, it's first-party, server-side, with a lookback window that matches your actual sales cycle.
Done wrong, it's last-click in Google Analytics. Or worse, it's whichever platform shouts loudest. If you sell anything considered, anything that takes more than a week to close, last-click attribution is lying to you.
Multi-touch with a 180-day lookback is the standard we wire on every system. It survives cookie deprecation because it runs server-side. It matches actual sales cycle. And because it talks to layers 1, 2, and 3, the credit it assigns is grounded in what actually happened.
If this layer is broken or missing, read our piece on first-party attribution in 2026 for what to build.
Why nobody has all four
Look at the market. Each category gets some of the layers, none gets all of them.
| Tool category | Layer 1 | Layer 2 | Layer 3 | Layer 4 |
|---|---|---|---|---|
| Hyros, Triple Whale, Northbeam | Partial | No | Partial | Yes |
| ClickFunnels, GoHighLevel | No | Partial | No | No |
| Salesforce, HubSpot, ActiveCampaign | No | No | Yes | No |
| Meta and Google ad platforms | Yes | No | No | Partial (theirs only) |
| FlowOS | Yes | Yes | Yes | Yes |
Attribution tools collect ad performance and run a model on it. They don't build your funnel, so they don't get the behavioral layer natively. Funnel builders collect behavioral, but most of them barely touch attribution and don't integrate enriched CRM data well.
CRMs collect the enriched layer, but they have no view into ads or behavior. And the ad platforms only know about their own clicks.
FlowOS sits at the center because the funnel, the tracking, the attribution, and the integrations to the CRM and ESP all live in the same platform. The data shares one identity model. That's the moat.
What to do if you're missing a layer
Missing layer 1 (ad-level)
Unusual. If you're spending and you don't have this, you're using one platform and not pulling the API. Set up reporting on the spend you have.
Missing layer 2 (behavioral)
Most common gap, the one no one notices. Get every funnel onto a system that captures events natively. Bolted-on tools always lose data at the seams.
Missing layer 3 (enriched CRM)
Pipe closed-won revenue and lead status back to your marketing system. Even a daily sync is better than nothing. The goal: marketing can optimize toward closed-won, not form fills.
Missing layer 4 (attribution)
Get off last-click. Build server-side, multi-touch, first-party. The 2026 article covers the playbook.
How we audit data layers in the diagnostic
When we run a 10-pillar diagnostic, we score the data layer (pillar 5) against this four-layer model. Most $1M to $20M brands score 4 to 6 out of 10. They have layers 1 and 4 in different tools and barely talk to layers 2 and 3.
Want us to score yours? It's free and takes 24 hours.