4 Pillars of Social Media Algorithms … Trust x4

Ever wonder why some Digg submissions go hot at 25, while others don't at 270+? Same with Stumbleupon ... why do some posts do exceptionally well while others with similar numbers of thumbs ups get substantially less traffic?

One word ... TRUST!

Trust plays a huge role in organic search algorithms, so why wouldn't it in social media algorithms. After all, it makes perfect sense intuitively. Social media sites rightfully want the best content to rise to the top, so at the end of the day there are really only 2 considerations for a piece of social news/bookmarking content:

• is it being supported enough to justify high visibility?
• if so, is there any reason to think that its support is artificial? In other words, is there any reason to distrust its apparent popularity?

Ultimately this means, the support for a content piece only needs to be assessed when it is about to be promoted to increased levels of visibility. So, how can trust/distrust be assessed? Below, I've outlined the 4 Pillars of Trust social media sites are likely to consider, and some ideas how they may assess each. Not all sites will evaluate all 4 trust pillars (eg. Wikipedia cannot really consider 'Trust in Voting patterns'), and the relative weighting will be different too, but its a good conceptual starting point.

1. Trust in the Submitter:

a. where voting is a factor, is the submitter typically a blind voter (speaks to his/her motivations) or does the submitter typically look at the content first?b. has this user been flagged as an overly aggressive pusher (eg. Digg) or spammer (eg. Wikipedia) in the past?

c. have some of the user's previous submissions been buried or removed?

d. where voting is a factor, how long are the voters spending on the actual article page before voting (Digg/Reddit/Stumble etc. have some techniques of assessing this)?

e. are voters finding the piece from 'Upcoming' pages, direct from the article, or do they arrive at the article's Digg/Reddit page directly (an indication of a push)?

f. feedback from others on the submitter's comments

Trust or distrust is further solidified over time. Stumbleupon, Reddit and Wikipedia submitters typically earn the trust of the sites as they spend more time on the respective sites. Digg on the other hand, appears to assess distrust only ... new users start with a neutral score, and lose points over time as they become more active.

2. Trust in the Site Hosting the Article or Source of the Content:

a. how many times has the site been 'hot' or referenced previously?b. how many times have pieces from the site been 'buried' or removed previously?

c. is the site a .edu or a .gov?

d. has the host site redirected 'hot' submissions to other pages previously?

e. how frequently do pages from the host site get buried?

Trust in Site is another important variable to some social media sites, and less so to others.

3. Trust in Voting Patterns:

a. # and % of votes made from people arriving at the voting page via actual "upcoming" pages or via the piece itselfb. # and % of votes from friends

c. # and % of votes from 'out of network'

d. # of votes secured through shouts

e. vote velocity ... is it consistent with the overall voting velocity of the site

f. voter IP locations

g. comments

h. when the piece is given increased visibility on the social media site, does it garner more votes naturally?

4. Trust in the Voters:

a. are the Voters viewing the page the article resides on before voting, or merely blind voting?b. how long are the voters spending on the actual article page before voting (social media algorithms may have techniques of assessing this)?

c. are voters finding the piece from Upcoming pages or direct from the article, or do they arrive at the article's page on Digg directly (an indication of a push)?

Implications of Distrust:
Ultimately, each social media site must engage in this trust assessment process. Sometimes its purely a human assessment (eg. Yahoo Answers or Wikipedia), and sometimes its a combination of both (eg. Digg, StumbleUpon).

The implications of Distrust maybe as simple as either a Yes/No assessment in which case Distrusted content is buried or disgarded. In most cases though, I suspect that overall trust follows a continuum, where the scores of each of the relevant Trust Pillars are weighted to arrive at a final score.

Wikipedia for instance, will weight the Submitter Trust factors very heavily, and place less or no weight on the Site Voting Patterns. Digg on the other hand, appears to weight 'Trust in Site' most heavily. Each social media site will have its own unique blend of weighting amongst the trust factors.

Next week I'll delve more into Digg, and the weighting of the various trust pillars in its algorithm. Stay tuned ...

If you like this post, subscribe to my twitter feed