What is A/B Testing? | Data Science in Minutes

Поділитися
Вставка
  • Опубліковано 21 гру 2024

КОМЕНТАРІ • 68

  • @kingofthedong3908
    @kingofthedong3908 3 роки тому +22

    im a graphic designer with no idea what i am doing other than design beautifully, and i was tasked on learning how to do a/b testing for my first job. this is very helpful for me to understand the basic, thank you !

  • @veej_77
    @veej_77 4 роки тому +53

    I'm a data science student and found this brief video informative, especially if you land a job in marketing research, etc. We didn't cover this in school and I wondered why that was because I see it everywhere on job searches. What specific tools are used for this?

  • @analisamelojete1966
    @analisamelojete1966 3 роки тому +30

    Basically, this is hypothesis testing but with a fancy name. Things like means difference, T tests, F tests, etc. Just pure statistical inference.

    • @lucasl1047
      @lucasl1047 3 роки тому +4

      Yep, data science just like BI relabeled a bunch of well known statistical techniques with more fancy, gimmicky, names lol

    • @analisamelojete1966
      @analisamelojete1966 3 роки тому

      @@lucasl1047 Agreed!

    • @lucasl1047
      @lucasl1047 3 роки тому

      @@analisamelojete1966 Gracias Ana :P

    • @Datasciencedojo
      @Datasciencedojo  2 роки тому +10

      @Ana Lisa You are correct. At its core, A/B testing is design of experiments (hypothesis testing, t-tests, confidence intervals, p-values etc.). Now add a lot of business understanding (seasonality, customers), UX design (colors, buttons, novelty), engineering (cookies, coding, servers/clients) and this becomes A/B testing or simply online experimentation :)

  • @zxvbhvxz
    @zxvbhvxz 5 років тому +29

    This was super clear, thank you so much

  • @relital3362
    @relital3362 3 роки тому +5

    Short and sweet. Thanks!

  • @willzo4026
    @willzo4026 3 роки тому +3

    I love this teacher.

  • @canaykoseoglu4766
    @canaykoseoglu4766 2 роки тому +1

    Hello, this is extremely useful information. I'm running AB tests with WASK. I can run A/B tests between my Facebook and Instagram ads and my Google ads. Furthermore, based on the results of my AB tests, I can disable inefficient ads and continue with my efficient ads.

  • @AnonyMous-vt3qy
    @AnonyMous-vt3qy 3 роки тому +28

    thank you unknown girl 🙏😉

  • @jchenmason
    @jchenmason 3 роки тому +3

    Thank you, Aubrey Plaza!

  • @vf6373
    @vf6373 4 роки тому +8

    i fell in love with A/B Testing within a few seconds watching this.

    • @sionmarak1916
      @sionmarak1916 4 роки тому +3

      There's an underlying A/B meaning on that sentence :3

    • @050138
      @050138 3 роки тому +1

      Me too! And today is Valentine's Day! 😍

  • @rushabhgada
    @rushabhgada 4 роки тому +4

    As simple as it can get. Thank you!

  • @jarombray6226
    @jarombray6226 2 роки тому +2

    Super helpful! I loved the visualization. It helped me internalize the information.

  • @Gibson-xn8xk
    @Gibson-xn8xk 5 років тому +9

    Thank you so much. Extremely helpful information!)

    • @SK-yc8lz
      @SK-yc8lz 5 років тому +1

      You are working with data?

  • @AL-pj7jg
    @AL-pj7jg 3 роки тому +1

    good job. simple and easy way of explaining.

    • @Datasciencedojo
      @Datasciencedojo  3 роки тому

      Glad you liked it. Keep following us for more content.

  • @awb19892
    @awb19892 9 місяців тому +1

    excellent video

  • @TheIronhideChronicles
    @TheIronhideChronicles 2 роки тому

    Thank you very much for explaining this in simple terms

  • @ZanetDesign
    @ZanetDesign 2 роки тому

    Hi
    Here is a question that I just can't find an answer to regarding thumbnail split testing.
    when is the best time to split test thumbnails?
    I use Tubebuddy for split testing my thumbnails right from day one. But im worried im doing something wrong and getting unreliable results. here is my process...
    1. I design 2 thumbnails that I think will work.
    2. I release the video to the public with my thumbnail A
    3. I now splittest to thumbnail B and rotate daily
    4. I let it run until a winner produces a statistical significance
    here is the 2 questions no one seems to consider or answer
    A)
    Isn't the first few days an unrealistic test as the first 24 hours the CTR is not relevent to the thumbnail but is more due to alerts and subscribers seeing you have a new video out. This is almost impossible fairness as a split test goes, so my gut feeling is don't split test in the first week to avoid this.
    B)
    the fact that a thumbnail has potentially 2 very different audiences also makes split testing irrelevant. if my video is for search, the audience AND CTR is totally different to the suggested video audience.
    Question
    how do you deal with A and B?
    so what and how is this meant to be handled, what do you do to allow for and address both these realistic scientific approaches to CTR?
    any response would be appreciated
    Your biggest fan
    Zane
    Zanet Design
    ua-cam.com/users/ZanetDesign

  • @hemantsharma5385
    @hemantsharma5385 3 роки тому +1

    How can i find web log dataset on healthcare website for a/b testing

    • @Datasciencedojo
      @Datasciencedojo  3 роки тому

      Hi Hemant, please forward your query to our site chatbot & email id: datasciencedojo.com/

  • @Tahoza
    @Tahoza 4 роки тому +2

    So... t-tests/ANOVA?

  • @michaeljordan49
    @michaeljordan49 3 роки тому +1

    This is excellent!!

  • @LILY-os1vo
    @LILY-os1vo Рік тому

    very helpful! thank you

  • @brajeshprajapat1590
    @brajeshprajapat1590 3 роки тому

    thanks to give the information on a/b testing it helped me alot.

  • @NMLeepsa
    @NMLeepsa 4 роки тому +2

    I am curious to know how you had that background with graphs. anybody who can help? I want to make my own videos for my channel with such visuals in the background. Is there any software for that? any youtube video to explain how such video is edited

  • @palashmyaccount
    @palashmyaccount 5 років тому +3

    Great video. Keep up the good work!

  • @rajgurrala6228
    @rajgurrala6228 3 роки тому +1

    The presenter is Amazing. But I felt she is a bit nervous, But your nervousness is pretty attractive. Come on. you are the best and you will look great being more confident. Very informative 3mins.

  • @muneebbolo
    @muneebbolo 9 місяців тому +1

    tysm

  • @azizmamoughli267
    @azizmamoughli267 4 роки тому

    Who Does A/B test ? ( Job/Career) ?

  • @antwanwimberly1729
    @antwanwimberly1729 Рік тому

    USA 🇺🇸 #lanetix taught me

  • @VVV-wx3ui
    @VVV-wx3ui 5 років тому +17

    Am still not clear on A/B Test in DataScience. While I understand that this video is made to educate folks, but the message and the education part seem missing as I did not get anything from this video. Probably, the host is reading it out from the Tele-prompter and did not explain the need, approach and the conclusion. @"Data Science Dojo" -- pls pay attention and post right videos.

  • @topspeed8832
    @topspeed8832 3 роки тому +1

    Power

  • @gouthamrajesh2745
    @gouthamrajesh2745 2 роки тому

  • @somanshusharma6525
    @somanshusharma6525 3 роки тому +1

    The fact that its required item on job posting is absurd! If you can't trust your candidates to learn this then you should not be hiring them in the first place

  • @whatisahandle221
    @whatisahandle221 2 роки тому

    Question to users: For HR/benefits (Eg health insurance) and work intranet tools, is it just me, or are websites getting more and more clunky, click & menu happy, cumbersome, and poor in their search capabilities?
    All the new web tools my (large) company or new benefits groups have put out in the last ~3 years have seemed more and more like a Lego video game: you’ve got to click all the text bc the hyperlinks aren’t obvious, look carefully and explore all the text because of fancy picture & explanations vs simple titles or obvious organization, scroll way down on ever page, and explore all the hidden menus, because nothing is simply presented, and everything requires multiple clicks, even common actions.

  • @davidfitcher2953
    @davidfitcher2953 2 роки тому

    why don't we just call it Version Testing?

    • @Datasciencedojo
      @Datasciencedojo  2 роки тому

      Hello David, version testing is mainly used when our main focus is to keep track of and see the histories of different versions of a test case. Whereas in A/B testing we compare two hypothetical versions of a webpage or app against each other to determine which one performs better.

  • @sinan_islam
    @sinan_islam 2 роки тому

    You did not explain what is A/B test.

  • @sachinbabup
    @sachinbabup 3 роки тому

    Lindsay Lohan, is that you?

  • @vivekjoshi3769
    @vivekjoshi3769 3 роки тому

    Sorry but really not clear. I watched this twice & still not clear. Can anybody explain me ?

  • @HotShots00
    @HotShots00 3 роки тому

    she looks baked

  • @divertechnology
    @divertechnology 2 роки тому

    nice woman, she makes this less boring

  • @howdontanalytics6158
    @howdontanalytics6158 4 роки тому +4

    I mean....this video is not that helpful....? why everybody here is praising? I don't know man..

  • @ayushkr.sharma2303
    @ayushkr.sharma2303 4 роки тому

    😍😍😘😘

  • @pbanikk
    @pbanikk 4 роки тому +7

    You're so beautiful!

  • @maamriabassem3818
    @maamriabassem3818 5 років тому +5

    Amazing explanation, though you ought to smile a bit more honey 😂😂😊😊

    • @WackyMe4eva
      @WackyMe4eva 4 роки тому +13

      Would you make the same comment if it's a dude??

    • @maamriabassem3818
      @maamriabassem3818 4 роки тому +3

      WackyMe4eva I’m a teacher too, it had nothing to do with sex, just a piece of advice

    • @WackyMe4eva
      @WackyMe4eva 4 роки тому +8

      @@maamriabassem3818 what has that got to do with you being a teacher?