On the web, an ethical issue with A/B testing.

Talking about ethics on the web, there are more than just a few cases to be aware of.

Glen Greenwald’s talk on privacy mentioned a theory on mass surveillance, where we are conscious of our every action which could be watched at any moment, restricts our behaviour due to our tendency to respond based on expectations. Privacy is something that we need. It is not simply because we have something to hide, we need the personal space for our freedom.

However, too much perceived freedom may cause drastic actions.

There was the Twitter abuse, after Caroline Criado-Perez successful campaign regarding having a female figure on a banknote, where freedom of speech went too far.

Then there was Justine Sacco’s tweet which caused her downfall.

Moving away from Twitter, there was also Leah Palmer’s identity theft.

Even with all these happening, internet access is still not readily available for everyone due to affordability in some countries.

Now, let us shift our focus from a user’s perspective to a service provider’s perspective.

In January 2012, Facebook conducted an A/B testing without informing its users. The question of whether it was ethical or not, came from the fact that the point of the research was related to manipulating the user’s emotions.

A/B testing is used for testing a user’s reaction for two versions of a webpage. It is something that can improve the webpage based on the desired results of user responses.

It seems to be something normal and understandable for Facebook to do. Though, the user’s response may not be how they truly feel, as they are conscious that their reactions can be seen by others.

Whether or not the research was fruitful, is not of much concern in this discussion. The question still remains to be, was the attempt to manipulate user emotion without their informed consent of this particular social contagion experiment, ethical?

Without a doubt, our emotions are important to us, playing a part in decision-making for most of us. On the web, there are many things at work that we do not take notice of since they are behind-the-scenes. A/B testing for things like “What is the best phrasing of the product description to induce impulse buying?” would be quite common.

Although Facebook’s experiment involved only slight changes, not informing the involved parties explicitly when they are directly and emotionally affected, did not seem to be reasonable to me.

(400 words)



Greenwald, G. (2014) Why privacy matters. Available at: http://www.ted.com/talks/glenn_greenwald_why_privacy_matters#t-834564 (Accessed: 11 November 2016).

Editorial (2016) Twitter abuse: Easy on the messenger. Available at: https://www.theguardian.com/commentisfree/2014/jan/24/twitter-abuse-abusive-tweets-editorial?CMP=twt_gu (Accessed: 11 November 2016).

Ronson, J. (2015) How One stupid Tweet blew up Justine Sacco’s life. Available at: http://www.nytimes.com/2015/02/15/magazine/how-one-stupid-tweet-ruined-justine-saccos-life.html?module=ArrowsNav&contentCollection=Magazine&action=keypress&region=FixedLeft&pgtype=article&_r=2 (Accessed: 11 November 2016).

Kleinman, Z. (2015) Who’s that girl? The curious case of Leah Palmer. Available at: http://www.bbc.com/news/technology-31710738 (Accessed: 11 November 2016).

Kelion, L. (2013) UK jumps up internet scoreboard as digital divide grows. Available at: http://www.bbc.com/news/technology-24426739 (Accessed: 11 November 2016).

Hattenstone, S. (2016) Caroline Criado-Perez: ‘Twitter has enabled people to behave in a way they wouldn’t face to face’. Available at: https://www.theguardian.com/lifeandstyle/2013/aug/04/caroline-criado-perez-twitter-rape-threats (Accessed: 11 November 2016).

Arthur, C. and Swaine, J. (2016) Facebook faces criticism amid claims it breached ethical guidelines with study. Available at: https://www.theguardian.com/technology/2014/jun/30/facebook-internet (Accessed: 11 November 2016).

Hern, A. (2014) Facebook deliberately made people sad. This ought to be the final straw. Available at: https://www.theguardian.com/commentisfree/2014/jun/30/facebook-sad-manipulating-emotions-socially-responsible-company (Accessed: 11 November 2016).

A/B testing – the complete guide (2016) Available at: https://vwo.com/ab-testing/ (Accessed: 11 November 2016).

Instapage (2016) A/B split testing tips to improve landing page conversions. Available at: https://www.youtube.com/watch?v=oyoGU6wRsn4 (Accessed: 11 November 2016).


10 thoughts on “On the web, an ethical issue with A/B testing.

  1. Hi Gin!

    Great post on A/B testing!

    You mentioned about Facebook’s 2012 experimental testing on their users and like you, I agree that it does not seem reasonable for the platform to use and manipulate users, like us by our emotions to be guinea pigs in order to gain results. Though some people may argue that they’re merely using the test to improve their system, changing the algorithm without the consent of users seem rather unethical to me.

    However, I believe that by entering platforms like Facebook, we’re exposed to such practices, no matter ethical or unethical. The Internet is not 100% safe, after all. Though we can’t control how Facebook determines how to improve their site, I feel that we can limit the information that we put online, change privacy settings or for some people, decide to quit such portals altogether. Do you think there are other solutions?


    (150 words)

    Liked by 1 person

    1. Hi Dawn!
      Thank you for taking the time to read my post and comment!
      Yes, I think that there are other solutions. As you have said, we are unable to control how Facebook tries to improve their site. Other than what you have mentioned, I think that it is important to gain awareness of what Facebook is doing. This is so that we can understand the possible impact of Facebook’s actions and decide from there if we want to continue using the service as we have been.


  2. Hi Lee Gin!
    Thank you for introducing me to the concept of A/B testing!
    I think the issue of ethics is dependent on the effect the experiment will have on the subjects. Similarly to many field and social experiments done on the streets (Eg. how would we react to a lost child), the subjects are often unaware that they are being studied. I do not see the Facebook experiment as very unethical as it had little to no impact on the users. (https://techcrunch.com/2014/06/29/ethics-in-a-data-driven-world/)
    Furthermore, Facebook did not falsify anything and only changed the use of their words. I would think many marketing efforts out there are more manipulative than this! However, it is definitely an issue if the experiments involve lying, like the experiment done by OkCupid. (https://hbr.org/2014/07/were-okcupids-and-facebooks-experiments-unethical)
    Perhaps, people feel this is unethical because they do not like to be associated with the idea of being ‘experimented’. What do you think?

    (150 words)

    Liked by 1 person

    1. Hi Teresa!
      Thank you for taking the time to read and comment on my post!
      Your comparison with OkCupid helped make the distinction clearer.
      I agree that people may feel that it is unethical due to the idea of being part of a “human subjects research” as Max Masnick calls it (https://www.theguardian.com/technology/2014/jun/30/facebook-internet), especially without being asked for permission (which can also be thought of as a lack of respect).


  3. Hi Gin,

    Thank you for an informative post! It did also help educate me on one of the many ethical issues out there.

    I would agree that Facebook was not being open on the test that they were carrying out on its users and it’s unfair. Based on the articles linked, it shows that Facebook
    “is a for-profit company with its own needs, and its own agenda” and one of its needs was “towards maximising eyeballs on adverts”.

    Hence, my question to you is, would it be better for Facebook to let its users know about the test before carrying it out? If so, does this then not alter the way people feel, resulting in inaccurate results?

    Also despite “more evidence surfaces of the network’s fundamental amorality”, users are still active on it. What could be the reasons for them to stay faithful despite being “tested” without their knowledge?


    (150 words)

    Liked by 1 person

    1. Hi Juls!

      Thank you for taking time to read and comment on my post!
      With regards to Facebook’s experiments, I believe that they should let the users know about the test first. However, Facebook should also let them know that only a minority would be affected. With the idea that “I am not likely to be that lucky one”, it is possible for people to become less bothered by the experiment. A longer time frame would allow more time for people to forget that the experiment is still ongoing.
      Despite the increase in evidences of the network’s amorality, people may continue to utilize the network if it takes too much effort to withdraw or if there is no better alternative. The choosing of alternatives is an obstacle as different networks have different functions and connections.



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s