Donate to Absurd Job Vacancies! Donate to Absurd Job Vacancies! Donate to Absurd Job Vacancies! Donate to Absurd Job Vacancies! Donate to Absurd Job Vacancies!


Examine the expectations and inferences underlying selected job positions. Consider timely topics in career preparation and the struggle for fulfilling employment. Analyze what could be improved in either situation. If this blog reminds you too much of work, then peruse my namesake blog for lighter fare.

Fuck UWM and all universities! UW-Milwaukee and their brethren are mediocre. Click banner ads on ClixSense instead; it's a better use of time than a college education in the UW System.

Sunday, January 20, 2013

Suggestion for Tweaking the LinkedIn Skills Endorsement Algorithm

Many of you maintain a LinkedIn profile. Towards the end of my Taleo Universal review, I mentioned how very well LinkedIn exports your profile data to those ATS which permit LinkedIn to do so. Also, LinkedIn Jobs enables the user to apply to third-party vacancies in a matter of clicks by allowing upload of another resume to compliment the resume-style information on your profile. (It should be noted that many organizations posting to LinkedIn Jobs are now requiring applicants to apply on their own company website anyway, or else the applicant will not be a candidate -- check the vacancy details to make sure!)

However, some LinkedIn functions perform sub-optimally: the “Skills & Expertise” feature began as a cluster of word bubbles comprised of up to 50 skills to improve the number of keywords for which a user search would return his or her profile. LinkedIn then allowed “endorsements” or binary affirmations by first-degree connections for each skill and tweaked the search algorithm to factor these “yes/no” bonus points in its ordering of users in search results. That is an adaptation of the search engine convention of raising the order of pages having more inbound links than others, although it is unclear whether LinkedIn weighs endorsements from particular users more heavily than those from others. Such complexity in ranking would be similar to the Google innovation of ranking and averaging inbound links into a factor to be multiplied against the rest of the Google Page Rank algorithm.

The issue arises when LinkedIn encourages users to abandon current skills, many of which are already endorsed many times, with new skills having only one endorsement (the event which triggered the suggestion). For example, LinkedIn keeps on telling visitors to my profile to endorse me for "Budgets," but that is already on my profile under "Budgeting," which upon mouse-over parses to "Budgets" within the standardized LinkedIn vocabulary. (I chose "Budgeting" as one of my 50 skills before LinkedIn standardized its skill vocabulary.)

However, LinkedIn does not scan which definitions a person's skills parse to, only the unparsed text as displayed on the profile. So while someone searching for profiles having the skill "Budgets" will find my profile in the search results the same as he or she would while searching for "Budgeting" (due to LinkedIn translating them into the same term "Budgets"), the LinkedIn endorsement suggestion algorithm is poorly designed and consequently does not detect synonyms for those standardized skills in its database.

So why don't I retitle the skill "Budgeting" as "Budgets?" Because LinkedIn allows for deletion and addition, but not for editing. Therefore, I would need to delete the skill "Budgeting," losing three endorsements in the process, and then add "Budgets" to accept an endorsement for a “new” skill (to which the displaced skill already parses anyway).

Conversely, I’ve asked connections to endorse me for currently listed skills, such "Budgeting," and to ignore whatever “new” skills LinkedIn suggests, such as “Budgets.” I have all the 50 skills which I want to display and cannot sacrifice prior skill endorsements to include a replacement.

Be as vigilant when reviewing the blue box of newly endorsed skills which LinkedIn says you haven’t yet listed; chances are you already have a much-endorsed synonym for the skill in your skills list and that it parses into that “new” synonym-skill appearing in the blue box above your profile headline. Don’t accept the new endorsement unless you have a bunch of space remaining on your skills list for redundant skills.

It would be like listing the same job on your LinkedIn profile twice but with slightly different titles; your recommendations for the same employment position would be divided among your electronic representations of that same period of experience unless you tediously asked each recommender to make certain to click on the other position as well.

No comments:

Post a Comment