Vidalytics
Feature Requests
Changelog
←
Back to changelog
PoweredΒ byΒ Canny

new

πŸ“’ Feature Updates : Settings Experiments + AI Recommendations

Optimizing a video used to mean duplicating it, managing two entries in your library, and figuring out what to test on your own.
Not anymore.
Settings Experiments let you test any single video setting β€” CTA timing, autoplay, thumbnail, gating, and more β€” directly on your existing video, with no manual duplication and no library clutter.
And Vids AI Recommendations take it one step further: instead of guessing what to test, the system analyzes your video's performance and tells you exactly which setting change is worth running β€” so you can launch the experiment in two clicks.
Together, they make video optimization faster, smarter, and easier than ever.
1️⃣ Settings Experiments
πŸ†• What's New
A New Type of A/B Test β€” Test Settings, Not Just Videos
  • A/B test any individual video setting (CTA timing, autoplay, thumbnail, gating, and more) without duplicating your video
  • The original settings become the
    Control
    ; your changed settings become the
    Variant
  • Traffic is split 50/50 on the same embed β€” no code changes required
No Library Clutter, No Extra Billing
  • Everything is managed behind the scenes β€” your video library stays clean
  • The variant does not appear as a separate video and does not affect your storage billing
  • Traffic sent to the variant is billed as usual
Full Flexibility After Launch
  • Edit variant settings at any time after the experiment has started using the regular settings editor
  • Monitor performance in your existing Experiments report
One-Click Winner Application
  • When a variant wins, apply its settings directly to your original video via the winner pop-up β€” no manual copying
πŸ› οΈ How It Works
Settings Experiments work on the same embed as your original video. Your original settings run as the Control; the new settings run as the Variant. Traffic is split 50/50 automatically, and stats are tracked just like a regular split test β€” all without you needing to manage two separate videos.
πŸ› οΈ How to Use It
  1. Open the
    Experiments
    creation flow on your video
  2. In the variant selector, choose
    "Player Settings"
    (instead of "Another Video")
  3. Open the
    Variant settings
    and change the setting(s) you want to test (CTA timing, autoplay, thumbnail, gating, etc.)
  4. Confirm your traffic split and experiment settings, then save
  5. Monitor results from the
    Experiment Profile
  6. When a winner emerges, click
    Apply
    in the winner pop-up to push the winning settings to your original video
Variant settings experiment 2 RN
πŸ’‘ Benefits
  1. Test Any Setting in Seconds:
    No manual duplication, no managing two video entries β€” just pick a setting, change it, and launch
  2. Keep Your Library Clean:
    The variant is handled entirely behind the scenes β€” your library stays organized
  3. No Extra Costs:
    Variant videos don't affect your storage billing
  4. Faster Optimization Cycles:
    Edit variant settings even after the experiment starts, and apply winners instantly
  5. Familiar Reporting:
    Everything tracks inside your existing Experiments report β€” no new dashboards to learn
⚠️ Restrictions
  • Settings Experiments are available starting with the Pro Plan and up just like Experiments.
  • Only settings available on your plan can be tested.
2️⃣ Vids AI Recommendations
πŸ†• What's New
AI-Driven Experiment Recommendations on Vid Stats
  • Vids AI analyzes your video's performance data and surfaces a recommendation card directly on the
    Vid Stats
    page when a high-confidence testing opportunity is identified
  • Each recommendation includes a summary, the key metric it targets, and the reasoning behind it
Two-Click Experiment Launch
  • Click
    See Full Report
    to review the full context and proposed setting changes
  • Click
    Create Experiment
    to launch β€” a 2-step wizard walks you through confirming the variant settings and configuring your experiment
One-Click Winner Application
  • Same as manual Settings Experiments β€” apply the winning settings to your original video instantly from the winner pop-up
Decline & Dismiss
  • Not interested in a recommendation? Dismiss it with one click β€” it won't appear again for 7 days
πŸ› οΈ How It Works
Vids AI continuously analyzes eligible high-traffic videos in the background. When it identifies a high-confidence opportunity β€” based on your video's current settings, performance data, and benchmarks β€” it surfaces a recommendation card on your Vid Stats page. Recommendations are generated automatically; they appear when the system finds a meaningful opportunity, not on demand.
πŸ› οΈ How to Use It
  1. Go to a video's
    Vid Stats
    page
  2. If a recommendation is available, you'll see a card with a summary, the key metric, and why it matters
  3. Click
    See Full Report
    to review the full context and proposed setting changes
  4. Click
    Create Experiment
    to open the setup wizard:
  • Step 1:
    Review and confirm (or adjust) the variant settings
  • Step 2:
    Configure traffic allocation, goal metric, and guardrails β€” then confirm
  1. The experiment runs as a Settings Experiment
  2. Monitor progress from the
    Experiment Profile
  3. When a winner emerges, click
    Apply
    to push the winning settings to your original video
Captura de pantalla 2026-05-06 a la(s) 5
πŸ’‘ Benefits
  1. No More Guessing What to Test:
    AI tells you exactly which setting change is most likely to improve performance β€” backed by your real data
  2. Launch in Two Clicks:
    Go from recommendation to live experiment without any manual setup
  3. Data-Backed Decisions:
    Every recommendation is tied to a specific metric and rationale β€” you always know why a test is worth running
  4. Apply Winners Instantly:
    One click to push the winning settings live β€” no manual work
  5. Non-Intrusive:
    Recommendations only appear when there's a genuine, high-confidence opportunity β€” no noise, no false signals
⚠️ Restrictions
  • Vids AI Recommendations are available on
    Pro plans and above
    (Pro, Premium, Enterprise, Agency)
πŸ“
Notes
  • Recommendations appear
    automatically
    based on your video's performance β€” there is no way to manually request one.
  • Only one recommendation is shown per video at a time
  • A dismissed recommendation will not reappear for
    7 days
  • If you manually change your video settings after a recommendation appears, that recommendation will disappear
  • If a video enters a split test, any pending recommendations for it will be hidden automatically
If you have any questions about either of these features, feel free to reach out to our Customer Happiness Team at hi@vidalytics.com