Senior Media Data Analyst

This is the role that owns the MPF day to day.

You run the taxonomy and the build, and you are the media team's second line of support. When something in the plan data is wrong or unclear, you find the cause, fix it and tell the team what changed.

Then you make it better. The aim is a platform the media team trusts, and that asks for less of your attention each quarter.

What you’ll actually do

Run the MPF

• Run the taxonomy approval route, and elevate only what would change the standard.
• Change the build when it needs changing - schema, automations, scripts - and test properly before production.
• Keep the MPF in line with the contract that governs what reaches client reporting.

Support the media team

• Be the media team's second line for the MPF and media data: find the cause, fix it and tell them what changed.
• Fix repeat problems at the cause rather than the symptom.
• Help teams set up new clients and campaigns in the MPF right first time.

Improve it with AI

• Take the two prototypes into production, measured against a baseline before anything touches live data.
• Build what does not exist yet: anomaly detection on free-text fields, automatic categorisation of placements.
• Write and maintain the scripts and integrations behind it.

Why this one is worth a look

This is not a seat at the end of a pipeline. You own the system every media plan starts in, and you have room to improve it.

You will be putting AI into live client data - measured properly, with the output going into what clients see - rather than running experiments that never leave the sandbox.

What we’re looking for

• Hands‑on paid media experience. You understand how campaigns are planned, bought and trafficked, and you can talk to a media planner in their language.
• Experience in a data, analytics or media technology role where you owned a system other people depend on.
• SQL.
• You have built with AI, not only used it - scripted against a model API, designed and evaluated prompts, and put something into a workflow other people depend on. Be ready to talk us through one.
• You can tell whether an AI approach is working, with a number rather than an impression. Evaluation, not vibes.
• You support people well: patient with users, and clear about what changed and why.
• Self‑sufficient. You can take an ambiguous problem, come back with a proposal, and deliver it.
• You write things down - decisions and reasoning, not only instructions.

Nice to have

• Querying a cloud data warehouse. Ours is Snowflake - you would use it to check what reached reporting.
• CM360 or an equivalent ad server.
• Experience owning a data standard or taxonomy, including an approval route.
• Experience migrating or cleaning a large legacy dataset.
• Experience supporting or training the users of a data tool.