Job Search Engine Guide

Hidden Prey and the Job Hunt

   

Conventional wisdom has it that most jobs are never advertised – these ‘hidden’’ positions are offered and filled primarily via word-of-mouth through social networks.  So if job boards are your primary tools in the job hunt, you’re reducing your chances of finding the perfect job by over 50%.

   

It gets much worse!  Unless you are willing to invest substantial time, regularly use multiple job boards, and are adept at using Boolean operators, advanced search features, and filters, you’re reducing your chances of finding the perfect job by as much as 90%.  Here’s why:

   

Let’s assume that conventional wisdom is correct and 50% of available jobs are ‘hidden’, and 50% are advertised.  No single job board has every ad – Indeed and SimplyHired come the closest – so at best, you have access to roughly 80% of the advertised jobs on the largest job boards.  If you’re using only one job board, you’re accessing at best 40% of the available jobs (80% of 50%).  If you’re solely using CareerBuilder, Monster or HotJobs, then you’re accessing less than 20% of the available jobs.

   

‘Accessing’ the jobs isn’t the same thing as ‘Seeing’ the jobs though.  The simplistic queries we use, a lack of structured data, and the high rate of irrelevant results returned by job search engines conspire to hide many of the advertised jobs matching your interests. 

   

The chart below illustrates the point (click on the image for a clearer view).  The data is based on queries run today on Indeed – keep in mind that I’ve tabbed Indeed as The Best Site for Active Job Search. 

   

 

Notice that there are 17 jobs selling autos in Boston, MA.  Yet, if I search for ‘auto sales’ (equivalent to Auto AND Sales), Indeed returns 382 total results, but only 10 of the 17 relevant positions are mixed in the 382.  If I search for ‘car sales’, Indeed returns 488 jobs, but only 8 of 17 relevant jobs are mixed in the result set – and only one job shows up in the results of both queries.

   

If I try searching only the job titles for ‘auto sales (equivalent to Title:(Auto Sales)), Indeed returns a much more focused set of 13 jobs (which appears nice on the surface); but, I only see 8 of the 17 relevant jobs. 

   

Other options include using the Boolean operator ‘OR’ to broaden my search e.g. ((Auto OR Car) AND Sales) … this returns all 17 relevant jobs, but they’re mixed in a set of 869 jobs.  Only 8 of the 17 are on the first three pages of results, and a relevant result can be found as late as page 46.  A Penn State study found that “Fewer than 10 percent (of search engine users) bothered with the third page of results.”  How many (besides me) would make it to page 46?  Not many – as predicted by Optimal Foraging Theory.

   

Let’s recap:

   

§         50% of available jobs are never advertised

§         40% or less of available jobs can be found on any one job board

§         20% or less of available jobs are returned by simple queries

§         10% or more of available jobs can be lost in the clutter of irrelevant results

   

So next time you use a simple two-word query (without Boolean operators) on your favorite job board, keep in mind that roughly 90% of the relevant available jobs that you seek will remain hidden.   

   

Why Vertical Job Search Isn’t Better

   

In response to Irrelevance of Job Boards, Colin Kingsbury responded:

   

“Vertical search engines should be able to deliver much better results without requiring syntax simply because the context is pre-defined. When a user types "auto sales" into SimplyHired or Indeed, you know they are not looking for car dealers. Knowing that we are operating within a limited information domain creates all kinds of opportunities to make intelligent assumptions that deliver better results 95% of the time.”

   

I agree with Colin. 

   

   

Three possibilities explain the disparity between our expectations and reality:

   

1. Customers are satisfied with the status quo;
2. Job verticals are really advertising publishers, and they like the status quo; or
3. Verticals still lack adequate context.

The first possibility is unlikely, the second is probable in some cases, but the third is most certainly true – job verticals still lack adequate context. 

   

When I search for ‘auto sales’, the following key information remains unknown:

   

• Is this a job title, or two keywords?  If keywords, then the following titles could be equally valid matches:  auto sales rep, auto insurance sales, auto parts sales, auto sales receptionist, etc.

• What level position am I looking for?  General manager, sales manager, shift supervisor, sales associate, or sales clerk?

Certainly the job verticals could ask clarifying questions of each job seeker; however, the value is marginal unless similar information is collected for each job – and this is where job verticals run into trouble.  The jobs they index are from sites all over the Internet, posted without the benefit of a standardized format, and lacking in consistent information.

   

The basic problem is this:  Job verticals have tremendous volumes of unstructured data lacking sufficient context for highly relevant search.

   

Joel’s Podcast with HotJobs’ Dan Finnegan

   

Joel Cheesman continues to impress with a great podcast interview of Dan Finnegan, General Manager of Yahoo! HotJobs.  As always, Joel has perfect timing!

   

Yesterday, while I was posting that job boards aren’t in the labor exchange business, Dan was telling Joel all about HotJobs’ new advertising programs.  Listen to the podcast and count the times Dan uses the words: ‘advertising’, ‘advertiser’, ‘ads’, ‘classifieds’, ‘buying keywords’, etc. – the number is somewhere north of 20. Clearly HotJobs is primarily an advertising publisher. 

   

Advertising isn’t a four-letter word of course; but advertising isn’t the ‘outcome’ that labor exchange participants are seeking.  Ads are simply a tool in the process of finding the perfect employee or the perfect job. 

   

Searching, matching, screening, and qualifying – Dan didn’t use these words as often, but he did reference having the best algorithm for sifting through content.  And one of Dan’s statement really stood out:  “We’ve learned something really important … consumers gravitate to the content that they see as the highest quality and most relevant.”

   

Sometimes what we know, say, and do are different – Dan says that relevance is important, yet the following posts suggest HotJobs is near the bottom of the pack when it comes to delivering relevant results.

   

HotJobs – A Whole New Search?

   

Is Anyone Home at HotJobs?

   

HotJobs Rated Unacceptable for Semi-Active Job Search

   

8th Best Site for Active Job Search – HotJobs

   

Is the disconnect between the ‘spin’ and ‘reality’ intentional?  It’s hard to believe that with all of the resources of Yahoo at their disposal, HotJobs is unable to deliver more relevant results.

Job Boards Worshiping False Gods

   

Traffic, Size and Advertising – these are the false gods that lead some job boards to lose focus on what matters most – producing market efficiency in the labor exchange. 

   

Employers need to quickly and easily find the best candidates (active and passive) for specific positions, while jobseekers want to quickly and easily find the positions that satisfy their needs and leverage their skills.  Job boards create value when they act as matchmakers, speeding, simplifying and optimizing the connection between these market participants.

   

Why then did reactions to my Irrelevance of Job Boards post include:

   

• Improving relevance will cause traffic numbers to decline
• There is money to be made from the inefficiency of search
• Size matters more than quality

The answer is simple – many (perhaps most) job boards aren’t in the labor exchange business at all!  No, these chameleons are really advertising publishers, intent on maximizing revenue from preferential exposure.  Rather than selling a ‘labor exchange solution’ these publishers encourage an advertising ‘arms race’ among competitors for talent.  The more inefficient the market, the more profitable it is for the publishers. 

   

Google, and AdWords, epitomize this business model. Per Google:  “Google AdWords ads connect you with new customers at the precise moment when they're looking for your products or services.”  This is all well and good, but if Google’s search engine provided fully relevant organic results, then sellers wouldn’t need to pay Google $6 billion per year for preferential exposure.  Clearly, there is a great deal of money to be made off the inefficiency of search – and very little motivation for advertising publishers to improve relevance.

   

Thankfully, organizations like TheLadders, Jobster, and JobCentral are building successful businesses by improving the labor exchange.  If you’re an employer or a jobseeker looking for a true ‘matchmaker’ in the labor exchange process, then these three organizations should be at the top of your list.

   

The Job Hunt and Optimal Foraging Theory

   

Mostly domesticated, job hunters retain a subconscious appreciation for Optimal Foraging Theory (OFT).  At its core, OFT posits that organisms focus on consuming the most energy while expending the least amount of energy.  We survive by finding the most profitable ‘prey’.

   

Today, our favored hunting grounds are the electronic job boards, and our weapon of choice is the ‘search engine’.  OFT would seem to predict a migration to the job board with the most jobs (SimplyHired here we come), but availability of ‘prey’ is only half of the equation.  Rather than bringing our own weapons to the hunt, we must use the weapons provided by each job board.  So success rests on the interaction between ‘herd’ size and ‘weapon’ efficiency.

   

Enter Automaticity and job search design.  According to the Encyclopedia of Educational Technology, “Automaticity is the ability to effortlessly complete everyday tasks with low interference of other simultaneous activities and without conscious thought to step-by-step process.” 

   

The illustration at right “…shows how a scan of the brain might look before automaticity of a skill is acquired. The working memory is actively operating to process and categorize information.”

   

When the brain recognizes familiar tasks it processes the information and applies the correct rules to the procedure in order to reduce the demand on the working memory and allow for higher order processing of information.”

   

The illustration at right “…shows what the scan of a brain might look like after automaticity of a skill is acquired. There is a drastic reduction of activity in brain.

   

Applying ‘automaticity’ and ‘optimal foraging theory’ to job hunting – if search isn’t effortless, then we’re consuming more energy to capture our ‘prey’’, and we’ll continue looking for a more profitable hunting experience.

   

So let me ask you:

   

• Do Boolean operators confuse job hunters?
• Do job hunters infrequently seek a new job?
• Do you ask job hunters to choose between the simplicity of basic search and the power of advanced search?
• Have you optimized your basic search for short search strings?

   

The usual answers: yes, yes, yes and yes – which of course means that job hunters operate without the benefit of automaticity, expending substantial energy in an effort to overcome their limited understanding of search.   Furthermore, since the job hunt is infrequent, and Boolean operators have little use in daily life, ‘training users in the tricks of search’ is an impractical solution.

   

Frustration results, with the outcome predicted by OFT – job hunters seek a more profitable hunting experience.

   

Job Search Results and the Eye of the Beholder

   

“Our results look good to me.”   This is the most common response I hear when I talk to folks at job boards about poor search results; and this perspective was repeated in response to my Irrelevance of Job Boards post. 

   

Sometimes they’re right.  TheLadders, a service that I love, posted twice (here and here) to say they like their results – I do too.  As a niche service, TheLadders has a specific focus (100k+ jobs), a unique business model (monthly subscription), and structured data that produces shared success – they win when jobseekers win!  And they’ve earned the right to be proud.

   

But unlike the children of Lake Wobegon, every job board isn’t above average.  Many of them just think so; and this frustrates me to no end.  Because when a job board is unwilling to question their performance, then jobseekers are forced to either accept irrelevant search results, or to go somewhere else.

   

   

All of this begs the question: Why are my views so often different from the views held by the job boards?  Well, each of our opinions is clouded by the following differences in loyalty, focus, limitations and context:

   

Loyalty – While customers look for great results, regardless of provider, the staff at Job Boards know who signs their checks, and they are invested (often literally) in their products.  So right from the start, we view site strengths and weaknesses from a vastly different point of view.

   

Context – When I search, I know exactly what I want.  So I judge ‘accuracy’ from a self-centered perspective … ‘auto sales’ means selling cars to me.  Perhaps it doesn’t to you – and this is a conundrum for the designers of search engines. The search engine must find, on average, the best results for all users – not just me.  The sad part is that search can be all things to all people if the boards are willing to invest in understanding the context of each search, and jobseekers are willing to leverage the search options provided by each site.

   

Limitations – I’m reminded of the proverb “It’s easier to see a splinter in another’s eye, than to see a log in your own.  When I search, I don’t acknowledge my limitations in understanding the intricacies or tricks of each search engine.  At the same time, most job search engines are built for speed – trading off accuracy in the process – yet job boards seem to accept this trade-off as an unavoidable technology limitation rather than a design choice. 

   

Focus – It’s hard to see flaws from 10,000 feet – yet it’s hard to see past them at ground level.  Like TheLadders, Jobster is an excellent site, and they’re proud of the quality of their search results – Jason Goldberg sent me a sample query for ‘auto sales’ jobs within 50 miles of Boston posted within the past month, and the results were excellent.   But Jason was looking for jobs at a higher ‘altitude’ – nothing wrong with that, but it can lead to different perspectives.  What do I mean?  I just searched on Jobster for ‘auto sales’ jobs posted within 25 miles within the past 7 days, I see only one job – an Administrative Assistant. 

   

Narrowing the geography, and the time window, reduces the number of available jobs and begins to expose relevance flaws in every search engine.  Since some jobseekers look for fresh jobs each day, that are close to home, it’s easy to see why perceptions of relevance can vary between users and job board staff.

   

Looking at the search results from SimplyHired for ‘auto sales’ jobs at different ‘altitudes’ illustrates how the job pool shrinks dramatically the closer we get to ground level:

 

§         10,000ft view (auto sales, National, past 30 days) = 20,229 matching jobs

§         1,000ft view (auto sales, Massachusetts, past 30 days) = 541 matching jobs

§         100ft view (auto sales, Boston, MA, past 30 days) = 233 matching jobs

§         Ground-level view (auto sales, Boston, MA, last 24 hrs) = 40 matching jobs

 

Finding 10 good jobs from a pool of 20,000 is easier (for a computer) than finding 10 good jobs from a pool of 40 possible matches.  Ideally, the search engine would calculate relevance scores for each match and return only the jobs that exceed a certain level – a level that the user could change to meet their preference. 

   

In the end, what separates the best job boards from the rest is perspective – the winners are always asking the question: “Do our search results look good to you?”

   

HotJobs – A Whole New Search?

   

If you’ve followed my blog, then you know I’ve been pretty hard on HotJobs (e.g. Is Anyone Home at HotJobs?).

   

So I was excited to see that HotJobs has a “Whole New Search”.   

   

According to HotJobs…

   

“Job Search Results: we know you want the right ones... and fast! At Yahoo! HotJobs, we have listened to you.”

   

“We've been hard at work, to continue to provide you with an unparalleled job search experience. With Yahoo's trusted search as your guide, Finding The Right One is easier, and takes a lot less time.”

   

“So try the new, enhanced search experience at Yahoo! HotJobs.

It's a whole New Search!”

   

So try them I did – with ‘auto sales’ in Boston, MA (results at right … click for a larger image).

   

Wow … I continue to be stunned by how far out of touch HotJobs remains.  Can the jobs at right, including sports marketing, advertising, insurance research, food sales, etc. really be considered “…the right ones” for my search for ‘auto sales’?  

   

HotJobs may have been listening to someone … but it sure wasn’t me.

   

Revenge of the Job Boards

   

Queue the Darth Vader theme song … created a disturbance in the ‘Force’ did I.   J

   

Surprisingly, there was less than total agreement with my premise that many job boards provide irrelevant results.  The comments posted, plus emails received, combined with sentiments expressed, can be summarized as follows:

   

• Our results look good to me
• Our results are better than the other guys
• We’ll train our users in the tricks of search
• If we improve relevance, our traffic numbers will decline
• We make money off the inefficiency of search
• Size matters more than quality
• Vertical search should deliver better results

   

Interesting perspectives all; with each worthy of its own missive.  Next up: our results look good to me.

   

Irrelevance of Job Boards

   

Webster’s Online Dictionary defines ‘rel·e·vance’ as the ability (as of an information retrieval system) to retrieve material that satisfies the needs of the user.

   

Let me be clear – I’m not satisfied with the results I’m getting from many job boards.  In fact, ‘disgusted’ seems a more appropriate word.  So, I took a time-out from the Job Search Engine Guide. 

   

Are my expectations unrealistic?  Am I below average in my ability to use a job board?  You be the judge.

   

Below, you will find the top ten jobs emailed to me today by three scouts for ‘auto sales’ jobs in Boston, MA.  Not one of the 30 is selling automobiles!  Somehow, the job boards interpreted ‘auto sales’ to mean I’m looking for a job as an engineer, a mechanic, a pharmacy manager, a business analyst, a trainer, and a customer service rep.  Oh, and six of the 30 jobs were already closed.  Am I satisfied?  Not even close!

   

   

SimplyHired:

   

1) Direct Sales - Company Unknown

This role is a breakdown between research and advisory consulting services.

   

2) Sales Executive - Spectrus Real Estate Group

Commercial Real Estate knowledge A MUST!

   

3) SBS Sales Representative - Company Unknown

Prospecting by phone and foot canvassing your territory offering payroll solutions

   

4) Branch Sales Manager - National City

The job/position you have selected is no longer available.

   

5) Sales Assistant HFC - HSBC

The job/position you have selected is no longer available.

   

6) Sales Account Executive Beneficial - HSBC

The job/position you have selected is no longer available.

   

7) Sales Consultant, Benefits – Boston - ADP Claims Services Group

Identifying, developing and presenting business solutions through sales presentations and securing the recommendation of the user community.

   

8) **HIGH-ENERGY, EXPERIENCED SALES ADMIN NEEDED!** - Company Unknown

Perform moderately complex clerical, administrative and general office duties.

   

9) Engineer II – Gastroscope - Field Sales - Endoscopy

Design and development of electro-mechanical medical devices

   

10) Management Trainee - Enterprise Rent-A-Car

Management Trainees focusing on sales, marketing, customer service, business administration and general administrative tasks.

   

   

Indeed:

   

1) Sales / Sales Associate - Advance Auto

Front counter/retail

Sorry, this job posting does not exist or may have expired.

   

2) Sales / Sales Associate - Advance Auto

Sales clerk, stock work, clean shop, assist customers

Sorry, this job posting does not exist or may have expired.

   

3) Regional Account Representative

With Dealix Regional Sales Managers to maintain... account management and sales experience Consultative account management and sales experience Proven track...

   

4) Parts Associate - Advance Auto Parts

They are looking for Counter and sales help. Must be prompt for work.

Sorry, this job posting does not exist or may have expired.

   

5) *High-Energy Microsoft Office Admin Whiz Needed for Sales Office

Solutions for auto and truck dealers and vehicle manufacturers; and computer-assisted auto damage repair... for Human Resources, Sales or other departments...

   

6) Sales Assistant HFC - HSBC

Leading provider of consumer loans, credit cards, auto finance and credit insurance products in the United... support for the Branch Sales Manager and up to five...

   

7) Branch Sales Manager Beneficial - HSBC

Of our branch sales team, the Branch Sales Manager is... and operations of the sales offices. The position monitors and evaluates sales results, productivity...

   

8) Sales Account Executive Beneficial - HSBC

With at least 1 year of sales experience to join our... an excellent benefits package and a leading-edge sales/management training program. Where can your career...

   

9) Store Manager Trainee - Goodyear

Team. We are one of the nation s leading tire and auto service networks. Goodyear, the worlds largest... monthly and annual sales goals. Qualified candidates...

   

10) Retail Store Management - CVS

#1 retail pharmacy with more than $37 billion in sales and 5400+ retail stores in 36 states and Washington... service, completing sales floor merchandising...

   

   

CareerBuilder:

   

1) Field Specialist - MA, Framingham Mosaic   

Train wireless retail store associates on all aspects of selling our client's products

   

2) Field Specialist - MA, Brockton Mosaic   

Train wireless retail store associates on all aspects of selling our client's products

   

3) Entry Level Marketing & Promotions Nova Marketing & Promotions   

ENTRY LEVEL. Training in Event Coordination & Execution, Public Relations, Promotions

   

4) Customer Service Rep, NS Liberty Mutual Group   

Answers questions and concerns regarding account status

   

5) Fleet Mechanic Coca-Cola Enterprises Bottling Companies   

Performing mechanical repairs and preventative maintenance on company fleet vehicles

   

6) Merchandising Specialist - NH, Seabrook Mosaic   

Maximize sales of video game products in one of the largest retailers.

   

7) Merchandising Specialist - NH, Salem Mosaic   

Maximize sales of video game products in one of the largest retailers.

   

8) Merchandising Specialist - MA, Danvers Mosaic   

Maximize sales of video game products in one of the largest retailers.

   

9) Senior Business Analyst - Solicitation ADP - Brokerage   

Analysis of client data files for the purpose of developing and managing solicitation strategies

   

10) Area Sales Manager Capital One

Educate dealer personnel on COAF loan programs and direct marketing products.

    

Relevance has Multiple Perspectives ... One for Each Customer

   

What’s the primary purpose of a job board?  At the core is the process of connecting employers and jobseekers. Not just any employer with any jobseeker – no, each employer seeks the candidate that best fits their needs, while each jobseeker looks for the job that best satisfies their criteria.  One size does not fit all.  Or does it?

   

Periodically I’m asked to check out new job boards and provide feedback – as I did last week for a job board in beta that will remain nameless – and I can’t count the number of times the conversation has gone something like this:

   

Bob:   Your site has a nice clean look. My primary area of expertise/focus is search, and I tried a couple strings on the site ... the results seem weak at this point ... but as you say, you're still in beta.  Try a search for auto sales – most of the results are for auto parts sales.

   

New Job Board:  I think our organic search results are relevant and good.

   

Bob:  You're in good company when you say the organic results for auto sales look relevant. While true that all of the results contain the words 'auto' and 'sales', wouldn't I have likely used the word 'parts' if I wanted to sell auto parts?

   

New Job Board:  Point taken, however do consider the issue of international terminologies, I would probably look for "car sales" if my intent was to sell cars or automobiles.

   

   

Translation:

Bob:  I know what I want, and the search engine isn’t giving it to me.

New Job Board:  Looks good to me, you need to learn how to search. 

 

   

Bottom line – both perspectives are correct.  Our goals and skills are unique, so when we ‘search’, our perspectives on relevance are also unique.  However, and this is key, as a user, my perspective counts; the perspective of the job board does not.  The challenge for every job board is learning how to provide search results that meet the ‘relevance’ criteria for each and every unique user that falls within the job boards target audience.