Job Specifications
About us and our clients:
We’re Ballpoint, a creative growth agency based in London. We’re young and ambitious. Since launching 2 years ago, we’ve grown 2.5x YoY and plan to continue that trajectory for the next 5 years.
Ballpoint was born as it became clear creativity would be crucial for modern paid social advertising. The best agencies of the 2010s were either ‘performance’ agencies or ‘creative’ agencies, we’ve been both since day one.
Many of our clients are similar to us, early innovators in their fields. We work with growing start-ups through to household names, each with their own growth challenges for us to break through.
The role in a nutshell:
A freelance econometrician to own our Marketing Mix Models. You’ll iterate, and productionise MMMs in Meta Robyn (R) and Google Meridian (Python), then pressure‑test them with lift studies and hold‑outs.
The goal: bullet‑proof budget recommendations our clients can take straight to the CFO. You’ll plug into our existing data engineering, media buying and creative pods to make measurement a competitive edge.
Important info
This is a flexible role in regards to days/hours, for the right person
This is a 1 month contract full time or 3 month part time position, paying between £400 - £500 a day depending on experience and skillset
What you’ll do:
Day to Day
Pull granular spend / KPI data from BigQuery, Snowflake, GA4, Meta & Ads APIs.
Clean, transform and QA datasets; code lives in Git + CI.
Fit Robyn (ridge + evolutionary) and Meridian (Bayesian + JAX) models; compare accuracy, stability and ROI outputs.
Calibrate with geo‑tests, holdouts and causal‑impact checks.
Turn outputs into dashboards and optimisation scenarios; brief media buyers on re‑allocations.
Document everything so any analyst can reproduce in one command.
Skills you must have:
R (tidyverse, prophet, Robyn) & Python (Jupyter, numpy/pandas, JAX, Meridian).
Solid econometrics: Bayesian regression, adstock, saturation, priors, cross‑validation.
SQL (BigQuery), dbt or similar for pipelines.
Experiment design: lift, geo‑split, incrementality.
Clear storyteller – can defend a model to finance, founders and creatives.
Nice-to-have skills:
GCP or AWS infra, Docker, Airflow.
ML‑Ops (CI/CD for notebooks, Feature Store).
Causal inference libs (DoWhy, CausalImpact).
Visualisation in Looker, Tableau or Streamlit.
Experience pushing MMM results into bid‑modifiers or scripted budget allocations.
Why we think you’ll love it at Ballpoint:
A culture that celebrates big ideas and supports bold thinking
Work across brands like Tilly Sveaas, Mother Root, Origin Coffee & Thriva
London based office with monthly socials, plus summer and Christmas parties
You'll join a small but mighty team and be able to make your mark from day one
Application process
Please fill in the Typeform - we aim to respond to all applicants within 14 days to let you know if we’ll be progressing your application.
If selected, the remaining process will be:
30-minute intro call with our Head of Creative Studio
In-person technical interview with Ballpoint’s founder
Final decision and offer