My business partner Julia Tverskaya and I had the pleasure of chatting with the CTO of Blekko Greg Lindahl at Sourcecon last week. Greg’s presentation was the one that we both enjoyed the most!

Why should sourcers check out Blekko? Here are some reasons.

Blekko is a search engine that has made creating custom search engines its major feature. They are called “slashtags“. People can influence what is found; this is democracy applied to search engines.

Blekko doesn’t show junk search results, including – what I especially appreciate – dynamically generated pages that are pages of results of a search query on some sort of site aggregating and selling information. (Many of us are tired seeing “Email, address, phone numbers, everything!” in search results for people on Google.)

Among slashtags already created is “/people”. It points to pages that contain online profiles and bios. For those of us who source for software engineers they also have a wealth of slashtags poinitng to all sorts of software development technologies. I’ve asked Greg if it would be possible to have a slashtag for pages with contact info and he said not yet, but it’s possible to create one.

The drawback right now is the small number of pages indexed, compared to search giants. This may be temporary; the company is small, too, and has good chances to expand coverage in the future.

In the light of my participation in the upcoming SourceCon’s “Paid Resources” panel, I thought I’d share a few thoughts on the subject.

Roughly, paid resources fall into these categories:

1. Data collections, along with software to search it:  resumes; profiles; contact info; and competitive intelligence.

Data may be self-entered (as in job boards or LinkedIn) or collected elsewhere (for example, Zoominfo, theSocialCV crawl the web; Jigsaw collects data via its members).

2. Search tools; matching tools

I think that search tools need to be taken with a grain of salt. True, it feels nice to have your search strings automated or for software to search across different sources. But this can be compared to using an automatic transmission vs. manual in a car. Which one gives drivers better control?

As for matching tools (looking for resumes that match jobs) I personally don’t think any tool is achieving this at the time; the task is way difficult. Sorry!

3. Parsing and sorting tools

Parsing and sorting tools, on the other hand, make a lot of sense and could save volumes of time, money, and help dig out data that we’d never reach without them. Tools from Broadlook such as Diver help dive into hundreds of links letting you know what’s hidden there. eGrabber has tools that do similar tasks; the ResumeGrabber is an excellent resume parser that puls resumes from links, folders, etc.

To wrap this up: don’t expect parsing tools to search for you. (Diver comes with a list of sample strings;  but it is just the packaging, so to speak.) You must learn to search well to take advantage of tools that boost the productivity of processing your search results.

The key in successful search is correctly imagining what you are going to find. I recently wrote how one can construct  a Boolean search string by knowing its “anatomy“.

Here’s another creative approach that may help us dig out additional good results.

This time, let’s imagine a page with a link pointing to what we are looking to find, say, someone’s resume. As an example, a web developer may have a blog about developing sites with a link to her resume on that blog.

So here’s a sample search to try:

“resume * PDF” “web developer” ajax css +javascript -jobs

Here are some variations:

• “download my resume” “web developer” ajax css +javascript
• intitle:”about me” resume “web developer” ajax css +javascript

(you get the idea).

This string below will look for profiles on LinkedIn with links to online resumes:

site:www.linkedin.com “websites * * * resume” ajax css +javascript

You can also use this approach to look for pages with links to someone’s LinkedIn profile:

“linkedin profile” “web developer” ajax css +javascript -site:linkedin.com

Above all, this technique helps to cross-reference information.

If you search for people on LinkedIn you will see ten profiles per page. If you search within a group, it’s 20 profiles per page but it’s 10 again if you switch to advanced search. This seems pretty limiting.

Here are a couple of tricks that will let you see more.

Install Firefox if you don’t have it, and its add-on called Fastest Fox. (Even if you don’t care for this tip, it’s useful to have FireFox due to its rich collection of Add-Ons.)

1. Search within a group on LinkedIn. Scroll down, keep scrolling… you will be seeing new profiles as you scroll, 20 more at a time.

2. Search for people. Click on the first profile in the results and scroll down. You will be seeing profiles added to your page as you scroll, and you will end up seeing multiple people search results all on one page. (Then you can even disconnect from the web and go explore. Or, you could use a clever profile parser and get your profiles all sorted out in an excel file.)

P.S. By the way Fastest Fox also works with Bing search results.

A Boolean search string in a search engine has its structure. Its elements are: keywords, key phrases in quotation marks, operators (like site:), and special characters (like “*” in Google). Unfortunately, this structure has little to do with what we have in mind when we search; so the challenge is to translate our search for people with certain professional background into those cryptic-looking strings. To facilitate this translation, let’s think of the structure of our search this way:

1. What type of a document are we looking for? This could be: a resume, an online profile on a particular network, a list of people, someone’s contact info, or some other type of a page. To get to that type of results (perhaps risking some false positives but that’s OK) we could start the search string with:

• A resume: intitle:resume OR inurl:resume -jobs -job…
• An online profile (example): site:zoominfo.com/people…
• A list of people: filetype:xls name title company phone email…
• etc.

2. What kind of terminology are we looking for? These are your keywords, like Java, or C++, or SAP, etc.

3. What is the physical location? We add names of towns, area codes, zip codes, and other location pointers.

4. What are the target titles? Target companies?  These seem straightforward.

5. What do we want to exclude?  (Example: exclude “managers”) Use the minus to exclude words.

This doesn’t cover it all – but in many cases will help someone new to complex Boolean searching to get going. To construct your initial search just follow the structure: 1-2-3-4-5. Then (as always) alter the string depending on the results.

As an example, let’s search for resumes of Java developers in NYC from one of the large Wall Street companies. Let’s construct the search following the steps above; this is an example to illustrate the concept – and a fine search string to play with:

intitle:resume OR inurl:resume -jobs -job Java NYC OR 212 OR “new york” developer Bloomberg OR Goldman OR Morgan OR Merill -manager

That is my point for today. I am very interested to hear if it would help to create Boolean strings.

This site went over 100K+ hits this week…

…which has prompted me to take a look at some other statistics. These nice numbers below show the growing importance of Sourcing and Internet research in our work.

I have created the sites below in the last two years:

http://booleanstrings.com – 100K hits

LinkedIn Group (very active!) Boolean Strings: 11,000+ members. New discussions daily. Big thanks to the active sourcing community for keeping this up and running!

Top companies:

• Google (138)
• Kforce Inc. (84)
• Microsoft (65)
• Oracle (59)
• Korn/Ferry Futurestep (43)
• SourceRight Solutions (41)
• Amazon.com (38)
• KPMG (34)

Top locations:

• United States (8042)
• San Francisco Bay Area (1106)
• India (865)
• Greater New York City Area (628)
• United Kingdom (605)
• Washington D.C. Metro Area (440)
• Greater Chicago Area (407)
• Greater Atlanta Area (403)

Social Network “Boolean Strings”: 3,900+ members. 800+ discussions. About 4K hits per month. Bi-weekly chats on Thursdays (the next one is next week). Once again, big thanks to the members for excellent contributions. We run sourcing contests here too.

67 countries represented:

• Afghanistan
• Aland Islands
• Albania
• Algeria
• American Samoa
• Australia
• Bahrain
• Belarus
• Belgium
• Bermuda
• Bosnia and Herzegovina
• Brazil
• Canada
• China, mainland
• Czech Republic
• Denmark
• Finland
• France
• Georgia
• Germany
• Hong Kong
• Hungary
• India
• Indonesia
• Ireland
• Israel
• Italy
• Japan
• Jordan
• Kenya
• Korea
• Kuwait
• Lebanon
• Malaysia
• Malta
• Mexico
• Myanmar
• Netherlands
• New Caledonia
• New Zealand
• Nigeria
• Norway
• Oman
• Pakistan
• Peru
• Philippines
• Poland
• Romania
• Russian Federation
• Saudi Arabia
• Singapore
• South Africa
• Spain
• Sweden
• Switzerland
• Thailand
• Tunisia
• Turkey
• Ukraine
• United Arab Emirates
• United Kingdom
• United States
• United States Minor Outlying Islands
• Venezuela
• Vietnam
• Virgin Islands, U.S.
• Yemen

New experimental sites

• Hidden Resumes
• Hidden Leads:

– 1,000+ searches daily.

Other LinkedIn Groups:

• Twitter for Sourcing and Recruiting: 11K+ members
• Social Media Recruiting (co-managing with Bill Boorman): 8.7K+ members
• Software Development Jobs (very active; ~50 posts per day): 14K+ members
• Management Consulting Jobs: 21K+ members
• 100K+ Jobs Only: 4K+ members

Here are some personal and business numbers:

• 18,900+ connections on LinkedIn
• 14K+ followers/included in 670+ lists  on Twitter.

My agency Brain Gain Recruiting has been in business for 6+ years now. Both my partner Julia Tverskaya and I have won SourceCon challenges.

In the last two weeks we submitted 11 candidates (all in Software Engineering) of which 8 were intervied, 3 got and accepted offers.

What will the numbers be like in a year from now? 🙂

Given a company domain name company.com, and the first and the last name of a person, we can construct the address – if the company follows a certain pattern in email addresses (such as fist-dot-last-at-company.com). How can we find out what the pattern might be?

1. What NOT to do

Do NOT search on Google for

“*@company.com”

You may be lucky and find some email addresses. But you will be just as lucky (or unlucky) if you search, for example, for “*\company.com”. Both @ and \ are special characters and both are ignored. If you are in doubt, please compare the results of these searches:

•  “*@company.com“
• “*@@@@@company.com”
• “*\company.com” (or try using any other special symbol instead of \)
• “*company.com”

2. Check these sites that publish collections of patterns (thanks to Gary for pointing these out):

• Emails4Corporations

These sites cover several thousand companies; hopefully, their coverage will grow.

Have access to Jigsaw or Zoominfo? Check a few individual addresses there and you will see what the pattern is like.

3. Google for emails of company employees by using the search string

“email * company.com” (plus some optional keywords)

and/or “email * * company.com” (I have added one more asterisk). Sometimes you will have to exclude addresses starting with “support”, “info”, or “sales” to get to “human’s” emails; we all know how to do that. It’s not necessary, however, to remove false positives if we see enough results to come to a conclusion.

Here’s a simple example:

“email * companyname.com”

If nothing can be found, you may want to find out whether the company website domain and the email domain differ and correct the search appropriately.

4. Verify

Sites like Mailtester.com work all in the same fashion, so it’s enough to just try one of them. They do work if the company mail server allows verification of individual emails. They don’t work if the server doesn’t allow that, or if there’s a catch-all email address, or if the server is down at the moment.

5. Cross-reference for verification

This method still works fine! You could use Facebook’s and many other sites’ “find friends” functions as well. To be found, of course, the person needs to be a member with the address registered.

(Cross-referencing via social media also provides a glance at the person’s profile and lets you verify if you do want to send that email.)

I find it a bit annoying when Google search decides to omit similar results – which it does most of the time! To control this behavior, we need to go to the last page with the results and click the link “repeat the search with the omitted results included”. There’s a shortcut to doing this, which is adding &filter=0 to your search URL that is still a distraction. There’s however an even nicer way to get there.

In Chrome, go to Options/Basic/Search Engines. Select “add a new search engine”. Put anything for its name, a short word like “all” in the second column and something like

{google:baseURL}search?q=%s&filter=0&num=100

in the third column.

From now on, if you type the word “all” as the first word, just before you type in your Boolean search string into the Chrome address bar, your search will display all results right away.

Taking this further, this technique allows us to create shortcuts to control any Google and Bing search options that previously couldn’t be expressed within the search strings and had to be set up in the “options” or in the “preferences”. Some examples for Google include date ranges, Image, Blog, Patent, and Google-plus searches.

Here is a chart outlining the “core” syntax elements that most of us use daily.