Wednesday, December 5, 2012

What’s the most effective way to X-Ray search LinkedIn?


I’ve recently come across some blog posts and some Boolean Strings discussions on LinkedIn that inspired me to go back and tinker with searching LinkedIn via Google and Bing.
For example, I continue to see people talk about:
  1. Whether or not you should use “pub” and/or “in” (e.g. site:linkedin.com/in | site:linkedin.com/pub)
  2. Whether or not you should use -dir
  3. Using country codes in site: searches
  4. Using different phrases to target public LinkedIn profiles – e.g., “people you know”
My first reaction when people are curious about the most effective ways of retrieving public LinkedIn profiles is to encourage them to experiment on their own first instead of looking for answers to copy and paste. Quite literally 99% of everything I know about sourcing (and recruiting!) I learned through being curious and experimenting.
People learn by doing, and more specifically by failing/struggling, and not by copying and pasting somebody else’s work.
It’s also important to realize that you cannot and should not implicitly trust sourcing advice (or custom search engines) you find online or in training sessions/materials – there is never only one way of doing anything, and the CSE’s you use and the syntax you copy and paste may in fact artificially limit search results and prevent you from finding the best people.
So my advice is to take other people’s work (including mine!) and experiment.

LinkedIn X-Ray Search Syntax Experiment

I saw someone use “people you know” in their LinkedIn X-Ray search, presumably to target and isolate LinkedIn profiles and eliminate false positive non-profile results.
Can you guess what I did?
Yes, I tried it.
If you didn’t already know why someone would use such a phrase, it’s because it’s a fairly unique phrase found on public LinkedIn profiles, and it appears to be indexed by search engines like Google, so you can search for the phrase to find LinkedIn profiles without having to use things like (inurl:pub | inurl:in) to return profiles, or -dir, -jobs, etc., to prevent non-profile results from being returned.
You can see it at work when viewing your search results:



This is where it pulls the phrase from on the LinkedIn profile:



It does a decent job of isolating profile results, to the point where inurl:pub, inurl:in, -dir, -jobs (or any similar techniques) simply aren’t necessary.
But of course the next thing I did was wonder what other phrases would work.
All you need to do is look around a public LinkedIn profile and look at other words and phrases that seem unique to profile results and would not commonly be mentioned elsewhere on someone’s LinkedIn profile and experiment by including them in your searches and inspecting the results.



A quick scan yields a number of possibilities – here are a few:
  • “search for people”
  • “name search”
  • “join linkedin”
  • “full profile”
  • “also viewed”
  • “viewers”
  • “overview”
  • “million”
  • “contact”
  • “expertise”
  • “see who”
  • “introduced”
Now, I haven’t tested all of those and the others I didn’t even bother listing, but I did test a few.
Here they are below. Bear in mind the search itself is highly limiting, by design, as I wanted to have a reasonable number of results to make quick comparisons of any variances between result sets:
site:ca.linkedin.com “current * * engineer” java hadoop “people you know” “location * Toronto, Canada Area”
site:ca.linkedin.com “current * * engineer” java hadoop “viewers” “location * Toronto, Canada Area”
site:ca.linkedin.com “current * * engineer” java hadoop “overview” “location * Toronto, Canada Area”
site:ca.linkedin.com “current * * engineer” java hadoop “million” “location * Toronto, Canada Area”
site:ca.linkedin.com “current * * engineer” java hadoop “contact” “location * Toronto, Canada Area”
site:ca.linkedin.com “current * * engineer” java hadoop “expertise” “location * Toronto, Canada Area”
Don’t be fooled by the variance in Google’s estimate of results (763 to 1,220), which are different for each search.
Scroll to the bottom of the results for each search and you will see all of them only have 2 pages of results, and the actual number of results being returned ranging from 13 to 15 (13 being the most common). I will leave it up to you to compare the differences, if any, between the search results.
If you don’t have time to click and execute each search, I’ve performed a quick run through you can view here (best viewed full screen in 1080p):

View The Video and Read The Complete Boolean Black Belt-Sourcing/Recruiting Article

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.