Job Search Engine Guide

You Can’t Buy My Love

   

But you can buy a detailed report describing the 2007 Top Sites for Job Search competition. 

   

Interested in the testing methodology and detailed results for each round?  Curious why your site scored below your competitors?  Looking for specific usability and performance analysis that can help guide future development?  It’s all available to you in a format that uniquely compares your site to each of the tested sites in the 2007 Top Sites for Job Search competition.

   

If you follow the Job Search Engine Guide, you know that I refuse to accept advertising of any kind.  If I ever do, then I think you should question my objectivity.  But objective tests like the 2007 Top Sites for Job Search competition require considerable time; and one way to make this a sustainable evaluation process, providing guidance year after year, is to sell the detail-level findings to job sites as competitive intelligence.

   

Price?  $4,995 for a customized report for your site; or $9,995 for the report plus an onsite visit to discuss the findings and recommendations.

   

Let’s be clear though … at the end of the day, what I really want is your commitment to site improvement … the fee is simply a crude way to make sure that we’re both focused on the same goal.

   

Do you want to improve?

   

2007 Top Sites for Job Search – Indeed Wins Again

   

Two rounds, two wins for Indeed.

   

In this second round, fourteen of the top general job sites were tested to find jobs for ten hypothetical jobseekers in Boston, MA.  Roughly five times the size of Tulsa, OK (round one test location), Boston provides plentiful job opportunities. Thus, this test puts more emphasis on the ability of each search engine to prioritize their search results.

   

Indeed continues to impress by providing the most relevant opportunities with very few duplicates.

   

   

Surprises in round two? 

   

GoogleBase dropped from third in round one to last in round two.   While GoogleBase did the best job of weeding out duplicates in this round, perhaps their methodology is too aggressive in eliminating marginal content and partial matches because in the end they returned the fewest relevant results.  As a result, GoogleBase won’t be back for found three.

   

Jobs.com jumped to third in round two, from ninth in round one – impressive, but I can’t explain it.  In round three we’ll see which of the previous rounds was a fluke.

   

Jobster tried my patience – their site was painfully slow yesterday – but no points were awarded or deducted for speed in this round.  Alexa data indicates that Jobster is “Slow (72% of sites are faster)”, but performance yesterday seemed atypical.

   

   

Best Job Sites – Round 2 Results

   

1. Indeed
2. SimplyHired
3. Jobs.com
4. CareerBuilder
5. America’s Job Exchange
6. Jobster
7. Monster
8. LiveExpo
9. Oodle
10. GetTheJob
11. Just-Posted
12. HotJobs
13. JobCentral
14. GoogleBase

   

   

Round three challenge?  Who has the most jobs listed for a small western town?  May the best site win.

   

2007 Top Sites for Job Search – Indeed Wins Round 1

   

Indeed takes the top spot in round one of the 2007 Top Sites for Job Search competition – not a surprise really since Indeed took overall honors in 2006.

   

In this first round of competition, fifteen of the top general job sites were tested to find jobs for ten hypothetical jobseekers in Tulsa, OK.  To score well, sites needed sufficient ad content for Tulsa, plus they needed search technology that was powerful enough to find matching jobs without returning mismatches. 

   

Surprises in the first round? 

   

Craigslist is the first site to be eliminated from further competition as a result of a last place finish.  I’ve never been a fan of Craigslist (see here, here and here for background), but going in I expected them to finish mid-pack … after all, Craigslist does have a strong following.  In the end, simplistic search technology and limited ad volume in Tulsa were Craig’s undoing.  Yes, the results would have been different if the test location had been San Francisco; but, doesn’t that make Craig’s more like a collection of regional sites as opposed to a national site?  I think so.

   

HotJobs and Monster in the bottom five … ironic that two of the big-three job boards garner so much traffic and yet performed so poorly in this round.   

   

Just-Posted, a site that almost no one has heard of (Alexa traffic rank = 1,559,714) was fifth, leveraging their next-generation search technology to muscle past many of the big guns.

   

Best Job Sites – Round 1 Results

   

1. Indeed
2. SimplyHired
3. GoogleBase
4. Jobster
5. Just-Posted
6. CareerBuilder
7. LiveExpo
8. Oodle
9. Jobs.com
10. JobCentral
11. Monster
12. HotJobs
13. GetTheJob
14. America’s Job Exchange
15. Craigslist   

   

   

So let’s head to the round 2 challenge – finding jobs for ten hypothetical jobseekers in Boston, MA.  May the best job site win.

   

2007 Top Sites for Job Search

   

Who will be first?  Which of the following fifteen sites provides the best combination of content, features, and ease-of-use for active job search?

   

To make my list of 2007 competitors, a site must provide national coverage of the job market and contain ads for most types of jobs in the economy – for simplicity, let’s refer to these competitors as general job sites.

   

As you can imagine, with 40,000+ job sites to choose from, there are many excellent niche sites that can be better choices than any of these fifteen general sites in certain cases – e.g. theLadders in the $100k+ market, and Dice in the tech market, to name only two.  However, as good as some of these niche boards are, their ad volume is often quite low, so even dedicated niche board users will often watch one or more general job sites when actively searching for a new job.

   

So which of the following general job sites is best in 2007?  That’s what we’ll find out by challenging the competitors to a series of tests, with the poorest performer in each test eliminated from further competition.  In a competition like this, it’s just as important for each competitor to have few weaknesses as it is to have strengths.  May the best site win.

   

   

First challenge – Find jobs for ten hypothetical jobseekers in Tulsa, OK.  Who wins?  Who goes home?  Stay tuned.

   

   

The Big Three

CareerBuilder

HotJobs

Monster

   

Vertical Integrators

Indeed

Oodle

SimplyHired

   

General Classifieds

Craigslist

GoogleBase

LiveExpo

   

Federally Endorsed*

America’s Job Exchange

JobCentral

Jobs.com

   

Next-Gen Search

GetTheJob

Jobster

Just-Posted

   

   

* In preparation for the shutdown of America’s Job Bank, the Department of Labor published Training and Employment Notice 27-06, which included a copy of a Federal Register Notice outlining the criteria for inclusion in a list of links of private sites providing functionality similar to AJB.   

   

Something for Nothing – a Gift or a Gaffe

    

Jobster joined the free classified club today – and Jason says “…the free posting plan makes Craigslist a near-term competitive target” (according to GigaOM). 

   

I’m probably in the minority when I say that free is the wrong price for job ads.  Frankly, I don’t believe that the best things in life are free … perhaps it’s my economics training and the concept of opportunity cost that gets in the way; or maybe it’s the adages that ‘if it sounds too good to be true it probably is’, or ‘there’s no free lunch’.  In any case, call me a free-skeptic.

   

What if I’m right?  What if …

   

… employers care more about access to quality candidates than they do about ad prices?  GoogleBase has been offering free job postings for over a year … with no significant buzz or impact to my knowledge.  Monster does over $1 billion per year in revenue (the opposite of free) by selling access to candidates … in January Monster had roughly 20 times the traffic of Jobster.  It would appear that employers are less concerned about the price than they are about broad applicant exposure.

   

… jobseekers care more about finding real job opportunities (open jobs advertised without spam) than they do about the absolute number of ads?  For all of the success of Craigslist, they’ve still found the need to institute fees for some ads in some markets to help reduce abuse.  And, at least at the outset, Jobster says that ads on their site will never expire … a sure recipe for a site that grows full of stale ads for closed positions.  So who will pay, if not employers, for the policing and pruning of inappropriate and old content?

   

Bottom line, I just don’t see how free ads improve the match experience for either employers or jobseekers.

   

Do you?  If so, let me know.

   

   

   

Update:  For a better post on Jobster's new business model, check out Andrew Marritt's thoughts on Resourcing Strategies.

   

Netflix Offers $1 Million for Improved Predictions

   

Netflix is curious … to the tune of one million dollars.  If there’s a better way to help customers find the movies they’ll love, then it could make a big difference to their customers and their business. 

   

So Netflix is offering $1 million to anyone who can improve their prediction accuracy by 10%.

   

I’m stunned.  In an industry where the average transaction is under $20 per month, Netflix is willing to pay seven figures for a 10% improvement in match accuracy.  Incredible!

   

What’s the value of the average job placement?  Even if an employer uses Craigslist or Google Base, where job ads are free, the time investment to place an ad and screen applicants is far greater than any transaction in the video rental industry.

   

Yet when was the last time you read about a job board’s major investment in search/match accuracy?  With the exception of Career Builder, who licenses search technology from FAST, other job sites seem content to roll their own relatively simplistic search … offloading the job of sorting through mediocre results to employers and job seekers.

   

With millions at stake, a major player in the electronic job market will eventually seize on search/match as a key differentiator.  Imagine the press coverage, customer interest, and the speed at which a firm could vault into the lead if they employed the concept demonstrated by Netflix?

   

Bravo Netflix … a brilliant idea!

   

Job Central Beta – My New Favorite Interface

   

An interface is a window through which we can see part of our world.  Some windows are small, providing a very limited view; others are large, providing a broader view. 

   

Job Central’s new interface provides a panoramic view of the job market – the best window yet. 

   

With a single search, I can now see how many jobs are in inventory at Job Central, Simply Hired, Indeed, Google Base and Jobster. 

   

Job Central handles the ‘translation’ work needed to ensure that the results across the sites are comparable.

   

So if I’m searching for a ‘sales clerk’ position in Seattle (click the image at right for a larger view), I can quickly see that:

   

-          Job Central has 4 jobs;

-          Simply Hired has 68 jobs;

-          Indeed has 54 jobs;

-          Google Base has 24 jobs; and,

-          Jobster has 356 jobs.

   

One of the really cool features is that any of the five sites can be the primary site – if Jobster is generally giving me the best search results, I can search Jobster from within Job Central, and continue to keep an eye on the other sites for unusual differences in coverage.  Brilliant!

   

A caveat – each site uses a different search algorithm.  So, even though Jobster shows more than five times as many jobs as Simply Hired on this search, it could mean that Simply Hired does a better job of finding what we’re looking for, with fewer errors of inclusion.

   

Honestly, none of the sites do a great job on search accuracy – notice that the top job in the results from Job Central is for a data entry clerk – I wouldn’t consider that a great match for a ‘sales clerk’.  For comparison, the top job on Simply Hired is for an Accounts Receivable Clerk – another poor match.  Jobster, Google and Indeed return sales clerk jobs as their top results; but they also include jobs for ‘warehouse clerk’, ‘inventory clerk’, and ‘accounting clerk’ in their top ten.  Again, not great.

   

Over time, Job Central intends to integrate O*NET tags into their search algorithm to improve accuracy.  When they do, Job Central will be hard to beat.

   

   

Disclosure – Job Central licenses O*NET-SOC AutoCoder software from my firm; however, the JC Beta interface does not yet leverage our software.

   

Job Boards – Free is the Wrong Price

   

Sujith Nair posed the following question in response to Automated Job Tagging in Action:  “How do you ensure that advertisers posting same job advt with similar title can be filtered out?”  Great question Sujith! 

   

In CareerBuilder Suffering from Tag Spam I stated that CareerBuilder, “…primarily due to ‘tag spam’ …dropped to last place in a new job search shootout against Jobster, Indeed, and SimplyHired.”  Interestingly, even though CareerBuilder suffered from tag spam, they returned no duplicate entries.  Compare this to a rate of roughly 10% duplicate ads returned by SimplyHired.

   

Two lessons here:

   

CareerBuilder, by charging for ads, effectively discourages duplicates and near-duplicates.  The ROI for advertisers on duplicate ads approaches zero.  Even Craigslist, the granddaddy of free classifieds, has moved to impose low fees on some categories and locations to reduce spam.  Quality is tied to price – and businesses built on free services are going to struggle with spam.  In contrast, TheLadders leads the pack in data quality by having two staff members review each and every ad before posting.  Charging for services increases quality, and reduces duplicates.

   

Vertical job search sites aggregate free content – some from high-quality sources (e.g. CareerBuilder and Monster) and some from lower-quality sources (e.g. Craigslist and GoogleBase).  No disrespect intended towards Craigslist and GoogleBase; but their pricing model (free or low-cost) yields a higher percentage of marginal content.  Thus, the verticals become infected with this lower-quality content.

   

So, to reduce duplicates and near-duplicates, the three best choices are: 1) charge for each job ad; 2) aggregate only from trusted employers or sources that charge for each job ad; or 3) charge end-users for adequate data cleanup.

   

If you care about quality, free is the wrong price.

   

   

PS:  Free test-drives are a completely different animal.  Potential customers should have the chance to determine the personal value of a service before plunking down their money.

   

If Google Ran the Postal Service…

   

…I’d need a larger mailbox to hold the 1.9 billion results delivered when I ask for ‘my mail’;

…I’d need a dumpster to hold the junk mail from advertisers;

…I’d need a ‘Return to Sender’ stamp to process all of the misrouted mail; and, 

…I’d always miss some of my mail because sender and receiver used different keywords.

   

The U.S. Postal Service isn’t perfect, but they’re darn close when compared to Google search.  Each day, when my mail arrives, I find 90+% addressed to me, with the remainder addressed to ‘occupant’.  Very rarely do I receive a piece of misrouted mail – the percentage is astonishingly low given the handwriting I see.

   

What makes ‘search’ and ‘mail delivery’ so different?  Simply this – Tagging and Taxonomy!

   

Each piece of postal mail is ‘tagged’ with a delivery address, using a structured taxonomy of zip codes and street addresses, to precisely describe its destination.  There is no guesswork in selecting the mail for my mailbox – relevance approaches 100%.  (Note:  I don’t like junk mail either, and rarely find the content relevant, but this isn’t the fault of the postal service – their job is to make delivery to the specified address regardless of content.) 

   

So, if Google could precisely tag all of the content they index, using a precise and structured taxonomy, then the quality of search results delivered by Google might approach that of the Postal Service mail delivery.   

   

Google search results could approach the quality of Postal Service mail delivery if:

   

1. A precise taxonomy was broadly adopted for all content;
2. All content publishers assigned appropriate tags;
3. Google ‘returned to sender’ all content without appropriate tags;
4. Google’s search algorithms we’re optimized to leverage assigned tags; and,
5. The process was resistant to tag spamming.

I’m not so Pollyanish that I believe this scenario is possible.  But there are options that can get us close … especially in the realm of job search.

   

Stay tuned …

   

Jobster – 7th Choice for Semi-Active Job Search

Continuing the Top Ten Sites For Semi-Active Job Search series, Jobster is the 7th choice.

Delivering an effective semi-active job search function is harder than I expected.  Success requires strength in four areas: 1) tools for saving and managing saved searches; 2) effective communication strategies (email or RSS); 3) broad coverage of available jobs; and 4) an accurate search/match engine.

Google Base didn’t make the initial cut, HotJobs was rated unacceptable, and JobCentral returned no results for the saved search ‘auto sales’ in Boston, MA.

Things aren’t much better at Jobster.  During a ten-day test run, I received emails from Jobster on five days, containing a total of six jobs – none of which were for auto sales.  But two factors place Jobster ahead of JobCentral for semi-active search: 1) broader coverage of the job market; and 2) receipt of occasional emails (5 out of 10 days) so I could tell that the search agent was functioning.

Effective?  Not in this test.  Perhaps in a different market, or using a different search string, the results would be different – your mileage may vary.