New Using Ad Spend on Videos A/B Testing?

Discussion in 'Support' started by IMI, Oct 25, 2018.

  1. IMI

    IMI New User

    Joined:
    Oct 25, 2018
    Messages:
    0
    Likes Received:
    0
    Hello @Andrew
    On your YouTube video about A/B testing, you mention it is best to only test videos that receive about 100 views per day. The channel I plan to A/B test does not get that many views (some of the videos I'll test only have about 50 views total). Would you recommend using ad spend while A/B testing these videos so that they get more views? And is 14 days still the recommended amount of time, or do you think 7 days would be enough to get conclusive feedback?

    Thank you.
     


  2. Phil

    Phil Administrator Administrator

    Joined:
    May 7, 2016
    Messages:
    434
    Likes Received:
    553
    Good news - that now we're able to gather CTR data, the "Total Views" doesn't matter as much. Because for every 50 views, you are likely getting 2,000+ "Impressions". TubeBuddy's A/B Test Status screen has a built-in mathematical calculation called "Statistical Significance" that will let you know when we're sure that a test is over.

    Do NOT use ad spend to get views up. YouTube does not collect data for videos shown as ads and if you're paying for views, it will skew the results a lot.

    So - run a normal test, and monitor CTR data.
    Here is what it looks like: https://pbs.twimg.com/media/DqdA5aFWsAMmQ5f.jpg:large
     


    Dimi likes this.
Loading...