Product · Creatives

Every ad, watched and tagged.

Mediatape watches every video and reads every image: the hook, the angle, the offer, the setting. Versions of one idea are grouped, whatever the file is called, so your AI can tell you which ideas sell.

Inside one ad

Every beat, line, cut and claim.

Read automatically from the file itself. Press play and follow along.

FW-12 · META · 9:16
Beats
Speech
Six hours in the rain.Still dry.Taped stitches, two layers, under 400 grams.Built for the wet stuff.
Shots
4 cuts measured

0:00–0:02 · hook · close-up

“Six hours in the rain.”

Hook viewer with an intense weather scenario testing outerwear.

All labels hook, persuasion, presenter, offer, claims
Hook
bold claim
Persuasion
benefit · “education” left out: the two checks disagreed
Presenter
person to camera
Offer
$149 · Free returns
Claims
Six hours in the rain. Still dry. · Taped stitches · two layers · under 400 grams

One creative library

Every creative, labelled by what it is.

Hook, format, message, offer, and CTA for every ad, with the visuals, on-screen text, and spoken words behind each label. Creatives are read from Meta, Google Ads, TikTok and X.

One creative library38 creatives · 112 ads · 3 shown
MetaGoogle and TikTok creatives PlannedConnected by content, not campaign names.

FW-01 / Decomposition

Show the weather. Then the shell.

Hook
Forecast: go anyway. Benefit first.
Format
Product demonstration · vertical video
Message
Keep moving when the weather changes.
Offer
No discount; product benefit first.
CTA
Find your shell

Content metadata

Visual
Person, yellow rain shell, wet mountain trail, rain
On-screen text
Forecast: go anyway.
Spoken
None (music only)
Setting
Outdoors
Language
English
Runs as
15 ads · 6 campaigns

One creative, many ads

Fifteen ad rows. One creative.

The same video runs in many ad sets, gets copied, cut down, mirrored, re-scored, and renamed. Mediatape recognises it by its footage and audio, whatever it’s called, so results add up to the creative that actually earned them.

In your ad accounts · 15 rows

  • Shell_rain_v3Meta · Prospecting
  • Shell_rain_v3 - CopyMeta · Prospecting
  • Shell rain v3 (2)Meta · Retargeting
  • rainshell_video_finalMeta · Brand awareness
  • rainshell_video_final_15sMeta · Stories
  • Reels – go anywayMeta · Advantage+
  • GoAnyway_flip_testMeta · Reels
  • + 8 more with similar names

Ranked one by one, no row reaches the top five.

In Mediatape · 1 creative

FW-01 · Product in use

15 ads · 6 campaigns

5 versions of the same video

  • The original videosame file8 ads
  • A 15-second editsame footage4 ads
  • A captioned versionsame footage1 ad
  • A mirrored copysame footage, flipped1 ad
  • A new soundtracksame footage, new audio1 ad
Cost per purchase, all 15 ads$23.54

Lowest in the account. Spend and purchases are added up, then the cost is recomputed. Example data · Meta, Oct 8–14, USD, 7-day click.

Tests you already ran

Your account already ran the test. Mediatape finds it.

Every time a script is filmed somewhere new, or a video gets a new first three seconds, you run an experiment without calling it one. Mediatape finds ads that differ in exactly one thing and lines up their results.

you ›

Does where we film change how our rain shell video sells?

Script family · “Rain doesn’t cancel my run.”3 versions · 11 ads
  • FW-07 · METAIn a forest, in the rain
    Cost / purchase
    $21.33
    Hook rate
    34%
    Spend
    $4,180
  • FW-08 · METAAt a city bus stop
    Cost / purchase
    $24.16
    Hook rate
    29%
    Spend
    $3,020
  • FW-09 · METAIndoors, at a window
    Cost / purchase
    $27.83
    Hook rate
    22%
    Spend
    $2,560
Labelled the same in all three
Presenter, script, music, offer, CTA, 15-second length
The one labelled difference
Where it was filmed

The forest version cost 23% less per purchase than the indoor one, and kept 12 points more viewers past the first three seconds. Where it was filmed is the clearest difference between them.

Found in your account · 9 tests

  • New location1 testSame script and presenter, filmed somewhere else
  • New presenter2 testsSame script, re-recorded by someone else
  • New opening4 testsSame footage, a different first three seconds
  • New voice or language2 testsSame video, dubbed or re-voiced
  • Not enough data yet14 pairsShown once both sides have 7 ad-days and 20,000 impressions

Matched by footage, audio and script, not by ad names. Labels other than the one that changed must match on both sides.

Semantic creative analytics

Analyze what your ads say. See what performs.

Group creatives by message, hook, format, offer, or CTA, even when the wording, execution, and ad names differ.

you ›

Do ads about weather resilience outperform ads about saving money?

Grouped by
Performance by message2 groups / 3 creatives
  • Weather resilience$26.93

    2 creatives · $6,060 spend · 225 purchases

    FW-01FW-02
  • First-purchase savings$32.14

    1 creative · $2,250 spend · 70 purchases

    FW-03

Weather resilience has the lowest cost per purchase at $26.93. The two groups also differ in format, so treat this as a lead to test, not a verdict.

Example data · Meta, Oct 8–14, USD · 7-day click. Group cost per purchase = total spend ÷ total purchases.

Video drop-off by moment

See where viewers leave, and what was on screen.

Each video’s retention curve, lined up with what was shown and said at every second.

you ›

Where do people stop watching our customer-story videos, and why?

Customer-story videos lose 30 points of viewers between seconds 2 and 4, while the logo card plays and before the product appears. Videos that open on the product in use keep 15 points more viewers at second 5.

0%25%50%75%100%0s5s10s15s0%25%50%75%100%0s5s10s15s
  • Opens on the product in use · 5 ads
  • Opens on a logo card · 9 ads
  1. 0–2sHook line: “The forecast changed.”
  2. 2–4sLogo card
  3. 4–9sProduct in use, in the rain
  4. 9–13sCustomer talking to camera
  5. 13–15sCTA: “Meet your everyday shell”

Creative and angle fatigue

See when the pitch itself is tired.

Each ad wears out. So does the angle behind it, even across fresh ads.

Also in the library

The details a creative review needs.

Everything here is available to your AI as a question, with spend and results attached.

  • Search by meaning

    Find creatives by what they show and say: “someone using the product outdoors”, “a price on screen in the first second”.

  • Angles you can correct

    Mediatape groups ads by the argument they make. Rename, merge, or split an angle through Claude, and your correction is used from then on.

  • Ad and creative age

    Days since launch for every ad and creative, so you can see which ones have run longest without a refresh.

Example questions

Ask about creatives the way you’d ask a strategist.

  • Which hooks keep people watching past three seconds?
  • Find creatives showing someone using the product outdoors.
  • Which of our re-recorded videos beat the original?
  • Which creatives have been running longest without a refresh?
  • Do ads that lead with the price cost less per purchase?
  • Where does our top video lose viewers, by placement?

Find out which ideas are working.

Get started