Google CSEs Go Wild with LinkedIn URLs

booleanstrings Boolean, Google, LinkedIn 2 Comments

This post will be of interest to those who use or create Custom Search Engines, in particular, to X-Ray LinkedIn for profiles (and avoid high LinkedIn subscription fees). The post is somewhat technical except for the first few paragraphs. We have a webinar on CSEs if you want to dig deeper into CSEs.

You might have noticed that the LinkedIn X-Ray Custom Search Engine you may have bookmarked, that used to work, will now return few or no results on seemingly open-ended queries.

Here is a new CSE to X-Ray LinkedIn that fixes that issue – I recommend using this one:

http://bit.ly/BetterLIXray

Now, for the technical part. For those who use advanced CSE search operators and create CSEs, here is what has happened over the past few months. We are seeing more than one unexplained behavior; it applies to LinkedIn URLs only. Nobody seems to have any clue why this is happening, so this information below is more like a weather report, not a “how” and “why” post. (We had to work very hard to “fix” the LinkedIn Agent in Social List).

  1. CSEs no longer “see” the Person object in public LinkedIn profiles, and neither do CSE APIs. Profiles indexed more than a few months ago still have “Person” in Google’s cache and can be queried, but newer profiles don’t have it (and do not have “Hcard” either). Because of that, searching with more: operators will find only older, indexed a while ago, profiles. Creating a Person-based CSE, where you specify the object in the Control Panel, will also limit the results to old.
    (There is info on “Person” – for which we would love to search – when you examine the source code for public profiles, but somehow Google fails to see it).
  2. If you include both “linkedin.com/in” and “linkedin.com/pub” as sites to X-Ray in your CSE, it will be producing very few results, with the keywords mostly in page titles.
  3. If you include “site:linkedin.com/in OR site:linkedin.com/pub -pub.dir” as a string to add to each query and make the CSE search everywhere, it will work for a CSE definition. That is how I built http://bit.ly/BetterLIXray. However, if you make an API call for that CSE, it will be finding pages on LinkedIn, but all sorts of pages, like jobs – not just profiles. Odd.
  4. You can find profiles in particular countries by including URL templates, with the use of the asterisk (which you shouldn’t have to do), like “uk.linkedin.com.in/*”. You can make refinements for a list of countries in one CSE, but make sure you append the asterisks to the templates. Here is the current list of all countries if you need it: LinkedIn Countries.
  5. When searching, avoid using ORs, search several times instead – otherwise, you will be getting fewer results.

Social List (which is based on CSE APIs) is up and running! We have found a way around all of the weirdness.

It’s hard to say what to expect in the future months, but let’s hope that mutual understanding between CSEs and LinkedIn will be restored – and improved. It would be nice to search not only for schools, companies, and headlines but also for locations.

Cheers!

Comments 2

    1. Post
      Author

Leave a Reply

Your email address will not be published. Required fields are marked *