We all know how Linkedin works – people have their own profiles, which they update with their work experience, and over time, build a set of recommendations from other Linkedin users with whom they have worked.
Since all the information people save in their profile is uploaded directly by them, the one important way to ratify the accuracy of that information is to look at recommendations. And this is where things start getting subjective. Today there is a way to game this system – it’s not an easy way, but there is a way.
Depending on how we play the game, a user can decide to go out and ask for recommendations from a lot of people – depending on how big the user’s network is, and how effective the user’s influence skills are, there are users out there who will manage to get a lot of recommendations.
What does this say about the user? Does the presence of all the accolades mean that they ring true? Perhaps, but not always.
How do the rest of the Linkedin users, who might not be either as social media savvy or great at following up with tons of other users for getting recommendations, level the playing field?
The answer to that lies in an aspect that Klout is currently trying to uncover.
Klout, as you all know, is looking at defining, measuring and comparing a person’s influence across (read against) the others in the community. However, Klout has decided to look at activity, and recent activity at that, to go about that endeavour. Now, maybe their approach is best for them, from the point of view of building a sustainable business model wherein their partners pay them for access to influencers who are, rightly so, influential today.
However, since the majority of people in the world (including those on social networks), aren’t necessarily the most active users, it is bound to give a disadvantage to those influencers who are not very active on a regular basis. Let’s not get this wrong – we’re talking about those people who are influential, but just not all the time (read every minute of every day).
So how should Linkedin look at strengthening their social media strategy?
Well, they need to look at the other aspect of influence that Klout is not currently looking at. The long-term aspect. The aspect that defines people for who they really are, not just who they have been behaving like this week.
It works like this –
Every user on Linkedin gets a score (let’s call it a Reputation Score) that is dependent on the following factors –
– Education (type)
– Alma Mater (universities)
– Work Experience (number of years)
– Workplace (brand, number of years)
– Designation (number of years)
– Salary (number of years)
– Recommendation Score
Now since the world has millions of organizations, chances are, defining the appropriate weightage of score for Workplace and Designation will take years to fine-tune.
Education, Alma Mater, Work Experience, Designation across number of years, Salary and Age are still weighted factors that will take lesser time to get developed.
For the purpose of this post, however, let’s focus on the last aspect, viz. the Recommendation Score, which I believe will play a key role in strengthening the reliability of Linkedin for its community.
Say, we start with user A, who has received 2 recommendations, one each from user B and user C.
Since recommendations in today’s format are subjective, let’s look at leaving them (the text write-up portion) as subjective for those who want to read user A’s recommendations. However, since users have also started adding skillsets to their profiles, when user A is being given a recommendation, user B and user C should have access to the skillsets listed by user A. Apart from the subjective recommendation text write-up that they are giving user A, users B and C are also encouraged to rate user A on any number of user A’s skillsets (with a rating scale between say 1-to-5 or 1-to-10). This objective rating of skillsets goes towards user A’s Recommendation Score and we will just get to that in a minute.
Now, I know what you’re thinking – this system, like any other, can be gamed by people who would ask their raters to give them high ratings. However, if the Linkedin community is assured that the actual ratings given by users B and C will always be hidden from user A, chances are, those who want to preserve the sanctity of Linkedin (and most professionals will want that) will do an honest job of rating others.
Now to tackle how these skillset ratings translate into user A’s Recommendation Score.
Let’s say, for the sake of discussion, that user B belongs to user A’s team, and user C is the super-boss of user A (two levels up). Here, we could look at seniority of reviewers and give a higher weightage to those that are higher up the corporate ladder. That could be one approach. However, that does not necessarily level the playing field quite enough, as it still leaves a chance for user A to game the system to quite an extent (via edicts and ass-kissing, respectively).
So, an additional aspect that we will want to incorporate here is assigning a Reputation score to users B and C themselves, that in turn, are derived from their profiles (that contain Education, Alma Mater, Work Experience, Workplace, Designation, Salary, Age and most importantly, Recommendation Score).
Hence, say –
User B has a Reputation Score of 5.7
(derived largely from user’s Recommendation Score apart from E, AM, WE, W, D, S and A)
User C has a Reputation Score of 7.8
(derived largely from user’s Recommendation Score apart from E, AM, WE, W, D, S and A)
If User B gives a Skillset Rating of 6.0 (assume only for 1 skillset) for User A, and
if User C gives a Skillset Rating of 8.0 (assume for the same skillset) for User A, then
User A’s rating for that Skillset would be
((6.0 * 5.7) + (8.0 * 7.8)) / (5.7 + 7.8)
which in this case would be
(34.2 + 62.4) / 13.5 = 7.16
This 7.16 against this particular skillset goes towards building user A’s Recommendation Score, which clubbed with the other aspects of user A’s profile, go towards building a Reputation Score. In the future, if user A gives a recommendation to user X, the Reputation Score of user A can play the weighted influence it needs to have on user X’s Recommendation Score.
As we proceed with this methodology across all Linkedin users, we will find that ratings given to Skillsets go towards calculating Recommendation Scores which go towards building Reputation Scores. Over time, the Linkedin community’s individual Reputation Scores start playing a big factor in the integrity (read authenticity) of ratings that users obtain from the community.
Now, coming back to the other aspects of a user’s profile. It will make sense to start looking at giving incremental weightage for a user’s Experience (number of years) and Age, and maybe even Salary.
By this, I mean that although averaging out a user’s ratings for skillsets might make sense (for now), it might not make sense to treat Experience, Age and Salary as averaged-out factors, and instead give more weightage to those with more Experience, Age and Salary.
This would allow users with more Experience to rise higher than others in terms of Reputation instead of viewing two individuals with identical skillset ratings, but say with 10 and 20 years of work experience, as the same. One can always argue that the higher Reputation Scores of those (assume peers) that have rated the one with higher experience will end up increasing the user’s Recommendation Score over the user with lower experience who has been rated by peers (as both sets of peers themselves have higher and lower experience, and hence, higher and lower Reputation Scores themselves). However, this is an assumption, and it might come out to be true – how much of a difference it does make, is only something we will know once this model has been implemented and live data starts getting incorporated.
Also, if we want to start defining weightage to Education and Alma Mater, there could be two approaches to the same –
One is the centrally-driven approach that Identified uses wherein it ranks Universities according to their brand value and assigns a certain score to them.
The other approach, which in my opinion could start off as a crowd-sourcing-of-reputation project based on this post’s approach, is to look at the Recommendation Scores of those Linkedin users that belong to different universities, and derive a score for the Alma Mater and maybe also for Education. This would reduce the chances of getting it wrong, which are high since most social networks are, like it or not, U.S.-centric. This would also help in measuring the true value of one’s Alma Mater and Education as it would measure influence and reputation of its users over time (and not necessarily based on the latest global rankings of Universities).
Since we are looking at defining a Reputation Score that allows for incremental gain (and not just averaged out scores), we would want to look at evaluating a user’s Reputation Score on a relative scale of 1 to 100 across the community, a la Klout, instead of leaving it as a value between 1 and 10.
That being said, to sum it up, I would say that –
Creating Recommendation Scores by averaging out Objective Ratings received for Skillsets by using weightage given to Raters’ own Reputation Scores
are a better approach than
Letting Linkedin users rack up Subjective Recommendations one after another without paying heed to the credibility of those Raters.
In the long run, the advantages of Linkedin getting into the Reputation game will include letting organizations find jobseekers directly in a much more reliable environment, which will help Linkedin earn more revenue not only through 1:1 role-to-candidate recruitment fees, but also through providing a higher quality of paid access. Not just that, Linkedin can start using targeted advertising based on a user’s Reputation score, which will have direct correlation to those users’ purchasing powers due to high salaries. This is just the tip of the iceberg of possibilities, one which we will unravel as time goes by, if Linkedin is to tread down this path.