Marketing attribution at Neo Financial
Built one attribution view across five paid channels, with CAC and LTV models that guided where the budget went.
Problem
Neo spent across Google, Meta, TikTok, Spotify and affiliate partners. Each platform reported its own results. There was no shared view of what a customer cost to acquire or was worth.
Role
As Director, Marketing Operations & Analytics, I led a team of analysts. I also owned martech partnerships and the martech budget.
Solution
- I directed a marketing analytics data mart for attribution and paid media across all five channels.
- I built CAC and LTV models to guide budget allocation and judge return on spend.
- We modelled in dbt on Databricks, with a GitHub review on every logic change.
- Hightouch synced results back into marketing tools, so insights showed up where decisions were made.
- I reviewed vendor contracts against actual usage before each renewal.
Five platform reports
- Search, social, video, audio and affiliates
- Each platform credits itself
One attribution model
- Databricks and dbt, run by Airflow
- One definition of a conversion
CAC and LTV
- By channel and by segment
- What a customer costs, and what they're worth
Back in the tools
- Synced to ad platforms and the CRM
- Answers where decisions get made
Invest in the right channels
Budget moved toward channels whose customers paid back, not the ones that reported best.
Tailor outreach
Knowing CAC by segment showed who was worth reaching, where, and how often.
Where each channel lands
As each platform reports it, every channel looks fine. Channel C looks cheapest.
Outcome
- Paid channels in one attribution data mart
- 5
- Paid channels in one attribution data mart
- Models used to set budget and judge return
- CAC + LTV
- Models used to set budget and judge return
Marketing planned its budget against one shared view of channel performance, instead of five platform reports. Knowing CAC by channel and segment let us tailor outreach and invest in the channels that paid back.