Attribution models are really just cause and effect. You want to answer which business activities or touchpoints are driving which results or conversions.

Traditional marketing attribution tries to answer which sources of traffic—paid campaigns, organic Google searches, etc.—are driving which conversions.

Traditional tools, like Google Analytics, assume that both the touchpoint and the conversion happen on your website. That works fine for short, simple customer journeys like a very simple Shopify store. The customer clicks an Instagram ad and makes a purchase within the same session on the same site.

The Offline Conversion Problem

But most businesses have more complicated customer journeys, like these:

• leads that convert, sometimes months later, with sales reps tracked in Salesforce
• subscriptions where every subsequent payment happens automatically via Stripe
• direct mail or emails that convert over the phone without any web interaction

These journeys have multiple touchpoints, spread across different tools, where the revenue-generating conversion happens offline.

The Data Warehouse Solution

The only way to get the kind of robust, accurate, and business-specific attribution models you need for these kinds of journeys is to use your data warehouse.

With dbt-nexus, we do it in three steps.

1. In the client's data warehouse, we collect touchpoints from any data source and conversions from the true systems of record.
2. We unite the touchpoints and conversions into a single customer timeline with identity resolution.
3. We create multiple, business-specific attribution models using the same customer timeline.

Once we have the foundation of the customer timeline, we can create a new attribution model in hours instead of months. No new tool for every model. No waiting months for a new tool to populate data. No discrepancies across tools and datasets.

We've used this process to build traditional first, last, and multi-touch marketing attribution models for any customer journey (first vs last touch demo).

But once we realized that attribution is really just cause and effect, we used the same process to cover more bespoke use cases. Here are some examples that we've built for our clients using the same technique:

• A/B tests — which variant causes more conversions (live demo)
• Sales Reps — which sales reps close the most valuable customers (live demo)
• Referrals — which existing customers refer the most new customers, and which traffic sources bring in the referrers (live demo)
• Subscriptions — which channels drive the highest-LTV subscribers, with every Stripe renewal attributed back to the original touchpoint (live demo)
• Omnichannel commerce — true ROAS by marketing channel for a brand selling on its website, on Amazon, and through physical retail — every order joined back to its original touchpoint, regardless of where it was rung up (live demo)
• Click IDs for CAPIs — get the latest Facebook click ID for a purchase to send to the Conversions API (live demo)

All of these are built using conversions from whatever systems of record the business uses—Stripe, app databases, Salesforce, etc.—so the conversion and revenue numbers are 100% accurate and have the full history of all conversions.

They also use a variety of tools for touchpoints — Google Analytics, Segment, phone logs, form submissions with UTMs. Using the data warehouse, you can mix and match touchpoint sources to suit your model's specific needs.

That gives the data warehouse an unmatched combination of speed, flexibility, and accuracy that no off-the-shelf tool can provide.