What's broken today
Commissioning, marketing and research each work from their own picture of the viewer, and most titles never get a clear answer to who the programme is for.
What it costs
Calls rest on instinct, and you find out who a title reaches after it goes out.
Start by discovering your segmentsPanels and focus groups take one to two weeks and cost money, and a sample big enough to split by segment costs more. Unreleased work often can't be shown to outside panels at all.
What it costs
By the time the research is back, the budget and the slot are already committed.
Chatbots and simulated surveys give confident answers, with no way to tell whether real viewers would agree.
What it costs
So the answer stays out of the decision, and the team goes back to waiting for a panel.
We put the prediction next to real viewers' resultsThe audience map
See where your target audiences and everyone else landed, and read any viewer's reaction in their own words.
Key
You can compare results with real viewers on our eye-tracking panel.
The real-viewer panel
For anything you can show outside the building, we recruit real viewers screened to match your segments. They watch your content on their own device with eye-tracking, and you see their results next to the prediction.
Read our methodology300+ real people in our research dataset
Brain and eye-tracking data
Our audience models are trained on neuroscience data. The real-viewer panel measures where real people look as they watch.
A published method
Our methodology pages explain how we capture data and how we handle ethics and privacy.
A sample answer
This is the kind of answer a research team gets on a new series before launch: how interested each segment is in watching, and what they say about it.
You
“Our new six-part crime drama launches in the spring. How will our four core segments react to the first-look trailer?”
North AI
Interest in watching, score out of 100
Crime drama regulars
45 to 64, watch the genre every week
“Slow burn, proper detective work. The opening scene had me.”
Younger drama viewers
18 to 34, finish a series in a few days
“Looks good, but I need the hook before I give it six hours.”
Family co-viewers
Watch with a partner or the kids at home
“Too dark for the one hour we watch together.”
Light viewers
Watch drama now and then
“Nothing here tells me why this one is different.”
Lead the next trailer cut with the premise. The first scene already wins the regulars, but younger viewers need the hook sooner.
Illustrative example. We made up the segments, figures and quotes to show what an answer looks like. They are not results from a real run.
How it works
North AI infers which audiences your programmes and your competitors' appear to reach, from published content and ads, and grades each segment by how much evidence supports it. You can also test a segment you have in mind.
Upload a programme idea, a trailer, a promo or a short archive clip, and see how each segment responds.
Write a survey and every simulated viewer in your audience answers it. You read the results by segment.
Start with a quick read, then go deeper with a full report on every segment.
Private to your team, if you prefer
We can keep your audience map private to your team. We agree it with you when we start working together. The map on this page comes from a report on our own commercial, not a client's.