for BusinessSign in

SeeMee for Business · early access

See where your ad loses people.

And where it lands, on your own customers' faces, before you spend on it. They watch your video with their camera on; you see what happened, moment by moment.

Create a workspaceHow it works

Every new workspace starts with a finished sample test, so you can explore the results before you send anything.

More people smiled at video B.

This difference is bigger than chance would likely give. The rows below show by how much, and where it's too soon to say.

Video A36 completedVideo B36 completedB compared with A
Facing the screen, on averageA86%B90%Can't tell apart yetB +4 points (−1 to +9)
Smiled at some pointA11 of 36 (31%)B20 of 36 (56%)Video B betterB +25 points (+2 to +44)
Stopped before the endA6 of 42 (14%)B3 of 39 (8%)Can't tell apart yetB −7 points (−21 to +8)
The answer comes first, with how sure it is. Made-up numbers, from the test in the chart below.

Facing the screen, moment by moment

Two cuts of a 20-second ad, 36 viewers each. Made-up numbers; point at or tap the chart to read them.

Cut A: product first Cut B: story first

At 0:07, in cut A's slow middle, only half its viewers are still facing the screen. Cut B keeps three in four. That's the edit.

Three steps, one afternoon.

  1. Upload your cut, or two.

    MP4, MOV or WebM, up to three minutes. Upload two versions of one ad to compare them.

  2. Send it to your customers.

    Share one link anywhere, print its QR code, or import your list and get a private link per customer for your own email tool. Add a discount code as a thank-you.

  3. Watch the results come in.

    Play your ad above the curves: where people faced the screen, where they looked away, where they smiled. Open any response to see that person's reaction beside your ad.

Numbers you can act on.

Small tests can mislead. The results say how sure they are, in plain words, so you don't recut an ad on noise.

  • Counts, not scores.

    Every number comes with who it's out of: “12 of 30 smiled at 0:08”. Percentages only from ten people up.

  • No emotion reading.

    SeeMee reports what faces visibly did, like smiling or looking away. It never claims to know how someone felt.

  • A winner only when there is one.

    Two cuts are called clearly different only when the data says so, with at least 30 people on each.

Fair to the people watching.

Your customers are doing you a favour. SeeMee treats them that way, and tells them so before they press Start.

Read the viewer privacy notice
  • The camera is used only while your video plays. Never the microphone.
  • Face analysis runs in the viewer's own browser. No app to install.
  • Each viewer sees, before they start, exactly what you'll get.
  • Anyone can delete their response, and its video, at any time.

Test your next ad on the people it's for.

Create a workspace

Early access is for brands that send their own email and make their own video ads. Workspaces are checked before tests go live.