Improving Pitch’s Marketing Attribution

Project Overview

I redesigned Pitch's marketing channel definitions and their attribution logic to better understand a users' touchpoints along the customer journey and to improve Pitch's marketing spend. I updated the attribution logic of 8 marketing channels and introduced 19 marketing sub-channels for more granular insights.* I build an attribution model in dbt which automatically attributes website sessions to a marketing channel and sub-channel. In addition, I established uniform guidelines for UTM links to ensure accurate and automatic channel attribution.


During exploratory research, I found errors in Pitch's attribution model. These errors were caused by overlapping channel definitions. So I developed a proposal for new channel definitions and attribution logic. I collaborated with all 12 relevant stakeholders to ensure the redesign satisfied everyone's needs. The stakeholders included, CMO, CFO, Head of Product Growth, and all marketing leads.


For the technical implementation, 2 data sources were added to Snowflake via Airbyte. I created 1 attribution model and updated 7 pre-existing data models using dbt. The new attribution model works retroactively, meaning historical data was also updated with the new marketing channel and sub-channels. In addition, I created a UTM generator in Google Sheet so that marketers can automatically generate UTMs for their campaigns and thereby increase the attribution accuracy for their campaigns. The whole project was documented in detail in Notion for transparency and future reference.


The updated attribution approach accurately attributes 70% of signups to a specific marketing channel, with the remaining 30% being users who didn't accept cookies**. The new model improved attribution accuracy by 24% and reduced attribution to the 'other' channel from 10% to 1%. In addition, data consumers are now able to filter dashboards in Metabase for marketing channel and sub-channel, making it easier to understand how a user's acquisition channel affect Pitch's business KPIs.

The improved channel definitions and attribution modeling allow Pitch to understand where users come from and which marketing efforts acquire users who engage and pay for the product, as well as those who sign up but never return. As a result, Pitch can now better allocate resources to marketing activities that drive real business impact by acquiring more users who become paying customers.

Curious about improving your attribution model or thinking about a similar overhaul in your organization? Let's dive into a conversation.

*Sub-channels provide a more granular view of how different marketing activities within an acquisition channel contribute to its overall performance (e.g., paid-social vs. paid-video within the paid marketing channel).

**This level of non-attributable signups is standard in the SaaS industry.