Summary and results
After a few weeks of experimentation with Craigslist I found two ways to stop unqualified resumes from reaching my inbox.
These results are shocking (at least to me). Let me rephrase it:
How it was done
I was always wondering how can I post a job listing on Craigslist that attracts only qualified candidates. A single job listing for an office manager fetches 100+ job applications in Seattle, 90% of which are not even remotely qualified. You get overloaded with emails while trying to find 5 good candidates for an interview. I wanted a recipe for a “secret sauce” that stops dozens and dozens of resumes that look like they were generated by robots from pouring into my inbox. I knew that it could be done. You can use agencies that pre-screen applicants or put up a billboard over highway. But could it be done as a part of my “normal” hiring process?
I decided to create a fake job listing, post it in several cities (modifying only a few lines of text for each city) and track the quality of candidates that apply for each position. Each modification to the job listing was designed to pre-filter people who really want this job from people who mass send resumes to all listings they can find.
Chicago and Dallas – baseline
I posted original version of the job listing in 2 cities – Chicago and Dallas. Response to these listings established baseline on how many candidates such listing may attract in other cities.
Chicago. Metro area population 9.5M, unemployment rate 12%, total unemployed 1,200,000, received 144 resumes
Dallas. Metro area population 6.5M, unemployment rate 9%, total unemployed 600,000, received 81 resumes
In other words I received 1 resume for each 8,300 unemployed in Chicago and 1 resume for each 7,400 unemployed in Dallas. I decided to use 1 resume for 8,000 unemployed as expected ratio for application in other cities.
I went through all Dallas emails and found only 10 applications good enough for a follow up phone interview. That established a quality baseline, that I should expect 1 good application for 75,000 unemployed or 12% from total in a city.
Received: Total = 81
Received: Good = 10
Quality = 12%
Experiment #1. Job listing without contact information.
Location: Phoenix
Phoenix. Metro area population 4.3M, unemployment 10%, total unemployed 430,000.
Total expected applications = 430,000 / 8,000 = 53
Expected good applications = 430,000 / 75,000 = 6
In Phoenix I removed all contact information from the job listing. I kept only the company name. There was no phone, website or email. Craigslist listing itself had no “reply-to” email. I assumed that whoever wants the job will google the company name, go to website, click on “Contact Us” and send resume to info@company.com address.
Surprising results: I received only 22 applications (instead of expected 53), but 7 of them were good enough for a phone interview.
You can see that just by requiring applicants to use their brain and 30 seconds of their time I cut down the number of applications in half without loosing any good candidates! Note that ratio of total applications to good ones jumped from 12% to 31%.
Received: Total = 22
Received: Good = 7
Quality = 31%
Experiment #1. Job listing that asks for security deposit
Location: Boston
Boston. Metro area population 4.8M, unemployment 9.3%, total unemployed 500,000.
Total expected applications = 500,000 / 8,000 = 62
Expected good applications = 500,000 / 75,000 = 7
In Boston I was asking for a $1 refundable security deposit before accepting a job application. I was collecting security deposit via Karmabond.com and promising to return it in 3 days, if the application matches my job description. This was not a payment for a job application, but a “good faith” deposit to escrow account. I know it looked fishy, especially with Craigslist warnings on top of each listing. But I tried it anyway.
Surprising results: Out of estimated 62 resumes I received only 5. That is 8% of expected volume. But 3 of these 5 applications where good enough for an interview. Yes, I did not receive 7 good ones as I expected, but I got 3. In other words I received at least half of the good ones and almost no junk at all!
Received: Total = 5
Received: Good = 3
Quality = 60%
Other observations
1. I was surprised that changing “ideal candidate would have” to “candidate must have” made no difference.
2. Through this experiment I received hundreds of cover letters. Each job listing specifically said – “If you want to impress us then send a great cover letter”. Only 6 (six!) cover letters were truly tailored to my job listing. Is it difficult to write a few words that show you have taken the time to visit our website and understand the job description?
Tip to job seekers: I understand that the job market is tough, but you should not mass reply (AKA spam) to all offers you see. This does not improve your chances of scoring interview. If you want to get an interview you need to write a killer cover letter.
Written by
Alexander Uslontsev
alex@karmabond.com
