Custom Search Engines

The Custom Search Engine (more commonly referred to as CSE) is a great little tool from Google, designed to "harness the power of Google to create a customised search experience for your own website." In effect, a little Google to use within a website.

How is this at all relevant or helpful to recruiters or sourcers? Well, a CSE doesn't just have to search your own website; it can be used to search through any public domain, in the same way that a standard Google search can. Where a CSE comes into its own is that it allows us to save a number of complex search strings and filters into one search engine, allowing us to cover more ground than would be feasibly possible with a standard Google or LinkedIn search.

The possibilities really are exponential, but I would suggest that every recruiter should have at two CSEs in their arsenal; one for campus recruiting, and one for diversity recruiting. In my experience these are two areas of sourcing which require a high volume of key words and as a result a lot of treading over the same ground. These problems are compounded by the recent limit on search terms imposed by Google, meaning you can no longer save a list of 50 top universities anywhere and past that in to your search.

I'd imagine at this point I've failed to demonstrate why a CSE could be so powerful; let me present a (relatively!) short guide to how to set up your own Custom Search Engine which will hopefully show a little better why they can be such a powerful tool...

Setting up a CSE

1) Define your Sites

First of all visit the Google CSE page, and click on 'Create a Custom Search Engine' on the right hand side. You will need a Google account of any sort to do this.

The Name and Description are self explanatory; you then get to the first good part: defining the sites you would like to search. Here is one of the most important (and powerful) parts of this tool; have a good think about the types of sites you would like to search. Interestingly you can use wild cards in your search. You could search across blogs (*.blogger.com or *.wordpress.com) for example.

In this case, we are going to leverage the CSE to search LinkedIn, so let's define our sites as follows:

2) Find the Control Panel

Click next until you reach the 'Get the Code' page, and click on the link at the bottom labelled 'Change the Basics'. This will take you through to the CSE Control Panel, which is where you are given a myriad of powerful customisation options.

3) Set your Refinements

Of particular use to us at this stage is the 'Refinements' page. This will allow us to define our custom search terms for our search engine. Click on the 'Add Refinement button at the top to get started, and for the purpose of this example we will add a simple elite college search for a couple of the elite colleges in the UK; Imperial College and Cambridge. Note you can add as many as you want; use your imagination!

The search terms use full Boolean operators (AND/OR, parenthesis and quotation marks). Again, be creative; use that 'minus' sign to eliminate certain terms (eg. -recruiter -recruitment to eliminate recruiters from your searches).

4) Run your Search!

Save your Refinement, go to your search page and go wild; let's try a basic Ruby on Rails search, and you will see that the search results are all from, of course, LinkedIn. This in itself has saved you typing in your usual Boolean X-Ray search string, but you will also notice your refinement saved just below the search bar ('Elite College UK'):

Clicking on this link will apply the additional search terms you have saved in your refinements; et voila, all results will now contain 'Imperial College' or 'University of Cambridge'.

Something else to note is that you can apply a refinement to a specific site; so you could, for example, have a 'UK Colleges' tied to uk.linkedin.com and 'French Colleges' tied to fr.linkedin.com.

The possibilities are endless; obvious examples would be to save refinements of every elite university in each country, saving lists covering diversity search terms (maternity, miss, mrs and the other usual suspects) or lists covering female names. You could set up a search which picks up pages containing e-mail addresses; or you could set up a search for CVs/Resumes using a refinement along the lines of 'format:.pdf OR format:.doc'.

See what you can come up with, and good luck! :)

Google Images as a Sourcing Tool

We've all used Google Images before; be it searching for screen caps from a movie or searching for last minute clip art to use in a presentation. As a professional tool to help us in the sourcing of UX Designers or Software Engineers however? Probably not very many of us. I read an article by Adam Wiedmer recently which discusses candidate sourcing via Google Images and it got me thinking; is this something we do on a day to day basis anyway? Those in the UK will know instantly what I am talking about when I mention a front-page with a censored image of an 'Unnamed Premiership Footballer' - I can't be the only one who used Google Images to find the offending front-page when the story first broke?

Sourcing of a sort, although more to cure my curiosity rather than serving any professional purpose; but how could this be relavent to us trying to source candidates? CVs are (for the most part) blocks of text after all. What Google Images can do is help us to eliminate some of what Glen Cathy describes as 'Dark Matter' - that is, those candidates who remain hidden to the usual search strings and methods.  A typical boolean search on Google will begin with some combination of (cv OR "curriculum vitae" OR resume OR portfolio) and so on, which immediately eliminates a number of pages which may contain information which is relavent to us. If we try plugging the skill sets and a location in to Google Images, we can find some results which would otherwise be hidden to us. We are going to achieve this by using the (unnervingly accurate) 'Face' image recognition which was introduced to Google Images following their acquisition of Neven Vision in 2006, and is found on the bottom left hand side of an image search:

As an example, a (very basic) search for a 'User Experience Engineer' in London:

(intitle:cv | intitle:"curriculum vitae" | inurl:cv | inurl:"curriculum vitae") -submit -jobs -template "user experience" engineer (london | 020)

Brings the following 1,650 results:

A successful first search with a set of CVs which look relavent; but lets try running the same search in Google Images but eliminating the 'must have' variations on resume, which will demonstrate a number of pages which we will not see running a standard text search:

"user experience" engineer (london | 020)  (-intitle:cv | -intitle:"curriculum vitae" | -inurl:cv | -inurl:"curriculum vitae")

This has given us 70,900 hits which can not have appeared in our simple text search:

While the pictures of the faces may not be much use, click through and read the page behind the images and see what you can find. One result which stands out is the big blue chap in the middle; a UX dev who has a very strong looking portfolio indeed, is based in London and is looking for work right now:

Other results include lists of speakers at UX events, several blogs, and even a little cheekily bio pages on company websites; all of which would have become more 'dark matter' using simple text based boolean searches.

-TW

X-Ray Searching - The Basics

The idea of X-Ray searching is one which has been covered countless times across various recruitment blogs (most notably and thoroughly in Glen Cathy's outstanding Boolean Black Belt) - yet it is a technique which I believe is criminally underused, particularly here in the UK. Anyone familiar with the concept, please switch off now; this is very much a bare-bones guide to X-Ray searching.

Background

X-Ray searching is a term denoting a technique for searching LinkedIn (or other similar publicly available databases) via search engines such as Google or Bing. The overwhelming majority of LinkedIn profiles are public, which leaves them open to being indexed by any search engine for finding using familiar Boolean search techniques.

How?

Step 1: Load up Google.com. As simple as that. Nearly, anyway; if like me you are based outside of the United States, you will be redirected to Google's localised site (so in my case Google UK) - this can actually skew your search results slightly, so click on the link at the bottom of the page which says 'Go to Google.com' (or, navigate to google.com/ncr).

Step 2: Build up your search string. The basic building blocks will be:

site:linkedin.com (inurl:pub | inurl:in) -intitle:directory -inurl:groups -recruiter

You then need to enter your keywords; in this example, let us try a very simple search for an iPhone developer with some kind of Agile experience.

site:linkedin.com (inurl:pub | inurl:in) -intitle:directory -inurl:groups -recruiter iphone (agile | tdd)

Next; location. A little different to a standard Boolean search, but only in ways which make it easier for us. Not too long ago, LinkedIn added the country code to the beginning of their URLs which makes for easier filtering; as a quick example;

site:uk.linkedin.com (inurl:pub | inurl:in) -intitle:directory -inurl:groups -recruiter iphone (agile | tdd) "London United Kingdom"

3. Search! :)

Why?

There are several reasons why one would X-Ray search rather than search directly in to LinkedIn; the big ones, for me, are that first of all it is simply quicker to run a search and skim through names via Google than LinkedIn, and that you are not limited to connections within your own network, but to anybody on LinkedIn with a public profile. You are also limited to 1000 search results when using Google; in a (free) LinkedIn account, your search is capped to 100 results.

The big loss when using Google in comparison with LinkedIn itself are the advanced search operators and the control that comes with them; it is difficult to control the distinction between current or past employers, for example.

Conclusion

Add this to your arsenal of search tools. If it is something you don't do at the moment, do it! :)

I have only touched the tip of the iceberg of what is possible with this introduction; I will add more advanced tips from time to time, but as with any kind of sourcing, it is up to the individual to take these ideas and really make them your own. Take the examples shown here, tweak them, change them, and use them.

-TW