BACKGROUND (feel free to skip):

Back in the beginning of the internet site counters were popular for administrators to provide an indication of the number of people that visit the website (there were no 'likes' or 'hearts' back then), and provided some social proof that the site was popular/useful/cool/etc. This was phased out due to the increased amount of traffic on websites these days, and the phenomenon of 'viral' content that can cause a surge in page views out of group behaviour (e.g. youtube video 'sensations').

Then it was the era of social media with facebook being the most influential in creating the 'like' action and sharing behaviour, to the extent that it has since become a de facto standard for the popular/useful/cool/etc. status of a website rather than simply page views because it requires the visitor to actively action or endorse their feelings about the website. However, this has also gradually lost its value as online marketers and social media managers can 'buy likes' (i.e. use various means to influence large groups of people to give endorsement - notoriously by buying and selling likes and using fictitious accounts to do so).

There is no question that in the age of our online and connected society, social proof is an important factor in shaping our preferences and perception of websites. However, given that the current forms of social proof is quickly losing its ability to provide genuine value about a website, can the new algorithms and increased use of analytics help provide a new type of social proof indicator? An example might be the differentiation between views and reads on publishing websites like Medium, which appears to be the amount of time a person spends on the page plus the scrolling patterns and interactions that would indicate a person actually reading the content. And what kind of form will it need to be in to overcome the issues of the current types of social proof?

I believe that the new form of social proof might take the form of a metric that tracks not the immediate action but the subsequent or follow up actions of the user. For example, if the article is about healthy eating, and the user clicks on one of the links to an organic food shop and makes a purchase then it would be counted as a social proof. So maybe an 'influence' or 'action' that is generated passively as a result of subsequent user action is what I might call it.

QUESTION: does this type of social proof already exists or is being researched/developed at the moment?

Also, are there any thoughts on how it address the shortfalls of current social proofs? I think with more intelligent bots and conversational interfaces becoming more popular it will be an interesting test to the value of this type of social proof.

5 Answers 5


Have you considered using smarter algorithms with the same metrics? There are a lot of algorithms out there that are used to provide social proof while also being attack resistant. One of my favorite and easier to implement systems is this:

Computing and using reputations for internet ratings
Mao Chen & Jaswinder Pal Singh

Here's the Abstract:

Ratings for products and services are increasingly important on the Internet, as they allow users to harvest the wisdom of the community in making decisions. However, the difficulty with ratings is that little is known about the people providing them. Interpreting ratings well requires that the reputations of raters be factored into the scores computed for rated objects, even though these reputations are not explicitly available. Taking advantage of the insight that reputation can be computed implicitly from ratings, this paper addresses the reputation problem for raters and its application to evaluating rated objects. We develop a general method to automatically compute reputations for raters based on the ratings they and others give to objects, and incorporate these reputations to generate value-added information about rated objects. We evaluate our mechanisms by performing experiments on data from major rating sites, and show that they have the desired properties of a good reputation system. In the process, we analyze some key characteristics of different types of Internet ratings. To our knowledge, this is the first investigation into automatically computing raters'reputations and applying these reputations to better evaluate rated objects.

Essentially it compares how well a person "rates" an item by comparing their rating to how well everyone else rated. This affects their reputation and is factored into all ratings they make. The more times a person takes part in the system (i.e. rating something), the more trustworthy (or untrustworthy) their opinions are, which increases (or decreases) the weight of their ratings. So for example, if they are trying to manipulate the system by voting against the crowd, they are punished on all their ratings - practically reducing their trustworthiness on the system and reducing their effect on all outcomes. In order to have a successful attack, a significant portion of the community must decide to target that specific metric and vote along the same lines. This is very difficult, though not impossible, to do. Making the metric time dependent (making older ratings/votes not count anymore) reduces the risk of this by requiring repeated mobilization of all the attackers in order to maintain the attack.

The paper mentions ratings and reviews on products for online retail, but it can be applied to other metrics such as the ones you described in your question. A mentor of mine said innovation happens through the combination of orthogonal fields. I hope you'll be able to apply this ratings system to your application and create an innovative product!

Though this paper was published 2001, most websites have not implemented all its recommendations. For example, Slashdot implements a similar reputation system using explicit votes instead of implying it from their participation.

From Center for Communication & Civic Engagement:

This “News for Nerds” site uses a moderation system for both articles and authors that allows for a largely self-governed news outlet that obviates most hierarchical editorial functions. Regular users earn “karma” by submitting stories that get chosen for posting, and by posting comments which are rated by other users. As users’ karma goes up, they have a stronger voice on the site, and as their karma goes down they lose moderation power.

The system shares similarities to StackExchange's system, especially the reputation and trust aspects. The implementations are different, but the core principles are the same.

I am personally using the system developed by Chen and Singh for a website I will be launching in the future. It will include the implied reputation portion as well as some other improvements taken from papers published in the last few years.

  • +1 Can you also provide an example of a website that implemented this algorithm to make the answer more complete?
    – Michael Lai
    Jun 20, 2016 at 20:41
  • 1
    @MichaelLai I updated my answer with a few similar examples I can think of immediately. As for websites that implement the whole feature set, I will have to do more research. Advertising the system one implements can lead to the easier exploitation of vulnerabilities (look at Google's constant war with SEO professionals!), so I imagine I may have some difficulties finding examples. Obfuscation isn't security, though, so I'm sure I'll find at least one. Jun 20, 2016 at 21:16

most types of social proof on and off the internet can and will be faked by those who don't have it. e.g. the fake rolex watch.

a good example of social proof i think would be

1 product ratings with verified purchases and by reviewers with some kind of reputation (e.g. top reviewer) on ecommerce websites and apps.

2 reviews on glassdoor, where the reviewer has been verified with their linkedin profile

3 the mashable velocity graph where it shows how fast people are sharing a particular article.

when it comes to adding social proof to content. the lines are still blurry, the effort to building such a feature compared to the perceived value it adds is not much, which is probably why no one has come up with something yet.

one idea might be coming up with a system that limits the amount of shares or likes users can make per day. this would force users to make more valuable shares or likes. of course this will be countered by online marketers with fake accounts.

  • 2
    Not only do people who don't have social proof want to buy it, but even well known brands might want to do this to keep the competition at bay...
    – Michael Lai
    Jun 14, 2016 at 10:37

When someone has the "verified" icon on Twitter, he/she must be very popular.

enter image description here

It's safe to say that this type of social proof won't lose its value, because the selection is done by people. It can't be faked.

How does the Twitter verification work?

We concentrate on highly sought users in music, acting, fashion, government, politics, religion, journalism, media, sports, business and other key interest areas. We are constantly updating our requirements for verification. Note, verification does not factor in follower count or Tweet count. We do not accept requests for verification from the general public. If you fall under one of the above categories and your Twitter account meets our qualifications for verification, we may contact you in the future.


Social proof on your own website

enter image description here

A link to the (in this case) Awwwards page should be included for validation.

Social proof will be valued most when representative brands or companies acknowledge a person or website. Until algorithms become unfoolable.

Like you said, the amount of viewers and likes won't always guarantee quality.

  • Can you explain in your answer how the verified icon is awareded on Twitter in a bit more detail?
    – Michael Lai
    Jun 17, 2016 at 23:42
  • @MichaelLai Sure, I've added the explanation from Twitter. Hope it helps! Jun 18, 2016 at 9:25

No matter what social system you put in place there will always be some who seek to subvert it for their own personal gain. Whether we like it or not, this is what it means to be human rather than a robot.

Review-based social systems where someone describes their experience is a little more valuable because prospective buyers have a better chance to recognise the genuine reviews from the fake ones. Obviously you can combine multiple systems al la Amazon with their 5 star rating + description.

We have to consider the needs of the user as well as the business.

"However, given that the current forms of social proof is quickly losing its ability to provide genuine value about a website..."

Do you have any evidence to back this up?

Stack Overflow uses a combination of social mechanism combined with some clever use of gamification mechanics all designed to help you increase your social standing and intrinsic motivation, so that we all donate our time to help our peers for free.

So a more holistic approach is needed and I am sure some use of analytics may prove to be helpful. I am not sure what analytics will help the user though, so I need to ponder this. I guess it might be possible to use analytics to identify patterns which suggest programmatic solutions which are adding likes to your website.

  • Perhaps it is too early to say, but there are businesses out there buying and selling likes from 'suppliers', so at least that's one form of social proof that is losing its value. Facebook has added 'reactions' that will probably make it more difficult. There are also many social media platforms that automate some of the user actions (e.g. invites and posts) which reduces the values of these actions as proof of user interest and engagement (that add to the social proof).
    – Michael Lai
    Jun 14, 2016 at 9:56
  • Buying 'likes" is not necessarily a bad thing, but will depend on how they use what they bought.
    – SteveD
    Jun 14, 2016 at 10:01
  • For the users this is diluting the value of the social proof if it is bought and not earned.
    – Michael Lai
    Jun 14, 2016 at 10:23
  • Sadly I do not know what the supplier offers or how it relates to what the business buys. Do you have a link to something which can help me understand how this "like" transaction works?
    – SteveD
    Jun 14, 2016 at 10:28
  • Google "buy facebook likes" and you'll see what I mean. I did a search and this was the first thing that came up: buy-cheap-social.com. You'll also see plenty of articles that suggest why this is a bad idea, but it doesn't prevent people from doing it.
    – Michael Lai
    Jun 14, 2016 at 10:37

Facebook has added 'reactions' that will probably make it more difficult. There are also many social media platforms that automate some of the user actions (e.g. invites and posts) which reduces the values of these actions as proof of user interest and engagement.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.