- Lab
- Experiment 01
Live
Media planning thatshows its working.
A working media strategist: it takes a brief, allocates the budget across channels, estimates net reach and frequency against the real audience universe, then challenges its own plan before a human signs it off.
- 12 channels modelled
- 3 strategic scenarios per brief
- Swiss CPM benchmarks, editable
- Runs offline — no data leaves the page
Give it a brief.
Every number below is calculated from the inputs and the cost assumptions in the model section. Change the budget, the market or a CPM and the whole plan moves.
Campaign brief
Pre-filled with an example brief so you can see the output straight away.
Read for keywords: premium, brand safety, performance, frequency, local, B2B, and exclusions such as “no TikTok” or “no print”.
Budget allocation
Net reach against budget
Weekly spend
Evaluation layer
What the agent is missing
Scenario comparison
Nothing here is
a black box.
These are the planning assumptions the engine runs on. They ship as Swiss market benchmarks; overwrite any of them with your own rate card or post-campaign data and the plan recalculates. Edits stay in this browser.
Channel assumptions
How the numbers are produced
1. Audience universe
Market population, cut to the age range with Swiss-style single-year population shares, then narrowed by the selected segment. A target of 25–54 in Switzerland resolves to roughly 3.8m people.
2. Channel scoring
Each channel scores on objective fit, reach per franc ceiling / CPM, attention quality and brand safety. The strategic mode changes the exponents, not the channel list, so Efficiency and Impact genuinely disagree.
3. Allocation
Budget splits in proportion to score, then the engine drops any channel that cannot clear its minimum viable spend and re-splits. No channel takes more than 40%, and a test reserve is held back.
4. Reach and frequency
Per channel, reach = ceiling × (1 − e−build × GRP / ceiling). Channels combine inside a family with a duplication penalty, then across families with random duplication. Frequency is impressions divided by people reached; effective reach at 3+ contacts uses a zero-truncated Poisson contact distribution.
5. Challenge
The evaluation layer tests the result against thresholds: concentration (HHI), frequency floors and wear-out ceilings, spend at the efficiency floor, brand-safety index against the brief, and funnel fit against the objective.
Observe. Decide.
Challenge. Repeat.
The workflow is built as separable stages rather than one prompt. The highlighted stages are the ones running in this page today; the rest are where live data and models plug in.
The point is not to replace the strategist. It is to give the strategist more leverage.
Contact
Let’s talk.
If one of these sounds like you, write to me. A few lines are enough.
- A leadership roleIn media, marketing or AdTech.
- AdvisoryA media, investment or growth question that needs an independent, senior view.
- A partnershipMedia, AdTech and technology companies looking for strategic collaboration.
- An exchangeAbout where media is going — or about something in the Lab.