Implementing the New Hotness – Matthew Ferrree – SourceCon Spring 2016 @getreqt
Hi, I’m Matt and today we’re going to talk about Sourcing. I have 30 minutes to walk you through some of the more advanced ways people are finding candidates. To set the stage, this presentation came about after September’s SourceCon – was anyone in Dallas last year? Basically we had a tech sourcing round table – which there will be one tomorrow I’ll be leading – and one of the major themes was
“ THERE IS ALL THIS COOL STUFF OUT THERE BUT NOBODY REALLY KNOWS HOW TO USE IT IN A PRACTICAL MANNER” So here we are.
Honestly, there are tons and tons of amazing tools out there that will aggregate contact information, give you discussion points based on machine learning algorithms, interview assessments that require zero person-to-person interaction, and services that present offers to candidates prior to interviewing. We live in a strange new world.
So much of it though comes off like those old travelling snake oil salesmen of the late 1800s.
Lady Bewlyan’s Magical Candidate Generator
Or this. Dr. Bob’s Machine Learned String Writer
Sometimes you pay a fortune for these tools, get them set up, and a week later you’re:
Sometimes you end up thinking you’ve made a huge mistake
Other times you love a tool and the company gets acquired or shuts down.
So let’s look at a few ways to take advantage of the new sourcing hotness so we can minimize the other two scenarios.
Unfortunately, that’s not how it works in the real world. Before we dive into the new hotness let’s level set.
There is no silver bullet You will still have to do the work to produce results Nothing you implement will be perfect. There is a ton of trial and error involved.
So who am I and why should you be listening to me?
So this is the million-dollar question: WHAT IS THE NEW HOTNESS? Well, there is no correct answer. As alluded to earlier, there are so many new technologies, products and services coming up every day, it’s almost impossible to keep up. However, there are a few emerging areas that are showing some amazing results. Before we jump into this I want to give you the three criteria I use when choosing my tools:
Is it cheap or preferably Free? Does this thing actually deliver results, ie is it solving the problem I need it to or am I just spinning my wheels? Finally, is it reasonable? Is the juice worth the squeeze? Is what I’m getting out of this worth the hassle? When considering this, I have found these areas to be most useful:
So, what is the new hotness?
Deductive search – Formulating your search strategy around reasonable conclusions instead of throwing strings in a search engine.
APIs – There is a world of easily accessible useful data hidden in places only developers typically go.
Automation – Tools like data scrapers will not only help make you more efficient, but can also be employed to build targeted contact lists and search strings.
Here is the literal definition of Deductive Reasoning, but the gist is this: You’re making an assumption based upon things you understand to be true.
A very simple example
How does this translate to Tech Sourcing?
So how do we turn that into a search strategy?
Java is code. Code is written by engineers. Engineers are candidates.
If I can find code, I can find engineers / candidates
Instead of searching a general search engine – let’s go where there is code and get exactly what we want.
There are quite a few source code search engines. Some are specific to open source projects, some are like social networks, others like symbolhound allow you to search for non-alphabetic and non-numeric characters such as symbols, which is useful for STEM candidates.
Typically I would throw together a string like this and toss it into LinkedIn, Google, or my ATS
But that’s already old and busted – we’re essentially at the mercy of whatever that search engine has indexed and we’re going to have a relatively high signal to noise ratio.
1 term: Android
By doing a source code search I don’t need to look for titles like software engineers, because I am looking at the work of a software engineer.
I don’t need to add specific languages because I know anyone working with Android will be working with Java
I can search for 1 term and practically be guaranteed that my results will spot on.
Lets do it.
This is my result from Searchcode 640k results including the context of where the result came from (online profile, code repository, open source project, etc.)
Even though I’ve got a half million results, not all of these are going to be helpful. We still want to know who wrote this code and how to get in touch with them. Let’s add another term.
Try adding a domain name. Now I’ll pull results from people I know are software engineers, working at a target company, in the correct domain.
Close to 3k results but this time we’re searching for contact information hidden within the code itself. By shifting the way we search we come up with new results – here we’ve got a bunch of very targeted email addresses.
Let’s come back to this in a minute and talk about Automation.
There are many ways to automate your workflow, but I feel the best bang for the buck is with data scrapers. Data scrapers essentially allow you to take information from the web and turn it into something structured.
One free scraper is Import.io – you can sign up for free at Import.io
If you want to shell out a little more I’d recommend using Outwit. Outwit allows you to further automate by automatically scraping specific content from your results like contact info or documents, without any added effort on your part.
So let’s go back to our code search results
If I run this through Outwit I can automatically grab every email address returned in my results and format the results in Excel within seconds.
Here you can see the results and a preview of the data Outwit is collecting. There is a place for you to enter the URL you wish to scrape, a quick option for scraping contacts, and a preview of the Data outwit collects.
Here are the results – sweet, sweet emails.
Let’s try a different workflow. What if we went to Wikipedia, found a list of female tech organizations, and ripped that into Import.io.
Here is a list of Organizations for women in science from Wikipedia. If we could rip this list and turn it into a string we’d save a ton of time and energy. So let’s do that with Import.io and a Google Sheets string builder
Import.io click new extractor Navigate to the site you want to scrape Click ‘On’ Select the terms you want to scrape Choose ‘multiple rows’ Save
This is what your import.io Dashboard looks like. You can see the items scraped. Download your data as a csv
Give a shoutout to Jan Bernhart @janbernhart
Copy your phrases in Column G
Paste your terms in Column B E is a test to make sure everything looks correct Copy your string in Column L
This is what your string looks like
Now the newest Hotness… APIs
API Stands for Application Programming Interface. API’s allow someone to use the features of a product / service in their own application. For example “Log in with Facebook”, Apple Pay, Google Maps, etc. Many provide free access for our intents.
One of the most useful API’s for sourcing is GitHub. Github is a network for engineers to show off their coding projects or for companies / projects to host their code. It is x-ray searchable on Google and users often list their location and contact info. However, many do not, and by using github’s API we can often times pull up hidden contact info. Here’s how it works:
Github allows certain aspect of their service to be integrated elsewhere, for example given a username, return all that user’s projects.
This is your typical github profile. Name, username, a website, projects, but in this case, no contact information.
By making this call to Github’s API I am asking to display all public events taken by a user, and I will be looking specifically at projects they have contributed code to.
Bookmark this url
By replacing the xxxxx’es with the username
I get his email address.
Full Contact is another service providing a free API. With full contact you can look up other aspects of a person’s online presence. For example, given an email address show me everywhere this person hangs out online: I get their Twitter, LinkedIn, Google Plus, and some other information.
For doing quick research there are tons of Sentiment Analysis tools. Sentiment analysis allows you to provide text, a phrase, or URL and it will automatically tell you the content, weather the article is positive or negative, or return similar items.
For quickly researching Niche domains or quickly building strings with alternate terms, Sentiment Analysis tools, like Aylien, are a huge help.
In this example I asked for phrases similar to “machine learning”
Emailhunter is another great tool. Email hunter will return emails associated with a specific domain. Paid version allows you to download emails in csv format.
This works well for open source projects, industry groups, organizations, etc.
"Ada Lovelace Day" OR "American Association of University Women" OR "Association for Women in
Science" OR "Equality for HER" OR "European Platform of Women Scientists" OR "Institute for Women
in Trades, Technology and Science" OR "Organization for Women in Science for the Developing World"
OR "The Scientista Foundation" OR "Women In Bio" OR "Women in Science International Trust" OR
"WomenTechWorld" OR "Society of Women Engineers" OR "Ada Initiative" OR "Anita Borg Institute for
Women and Technology" OR "Association for Computing Machinery Committee on Women" OR
"Association for Women in Computing" OR "BCSWomen" OR "Black Girls Code" OR "BlogHer" OR
"Center for Women in Technology" OR "Computing Research Association Committee on the Status of
Women in Computing Research (CRA-W)" OR "DC Web Women" OR "Girl Geek Dinners" OR "IEEE
Women in Engineering" OR "LinuxChix" OR "National Center for Women & Information Technology
(NCWIT)" OR "Systers" OR "Women in Technology International" OR "Women Techmakers" OR
"WorldWIT" OR "WOUGNET" OR "Association for Women Geoscientists" OR "Trowelblazers" OR
"Association for Women in Mathematics" OR "American Association for Women Radiologists" OR
"American Medical Women's Association" OR "Association of Women Surgeons" OR "Kappa Epsilon"
OR "Medical Women's International Association" OR "Society of Woman Geographers" OR
"Sociologists for Women in Society" OR "Women'sNet" OR "WOUGNET" OR "WomensHub" OR
"WISET" OR "American Association of University Women" OR "Association for Women in Science" OR
"National Research Council" OR "DC Web Women" OR "Graduate Women in Science" OR "HWIC" OR
"HWIP" OR "National IWITTS"