Job Search Engine Guide

New Monster Home Page – Same Old Monster

   

Effective today, Monster has a new look; and to my eye, it’s a winner … a much improved visual design.

   

But a new coat of paint doesn’t make a silk purse out of an old dog – hmmm, perhaps I’ve been reading too much Dilbert.

   

First, the search box real estate seems overwhelmed by the branding and sales real estate.

   

Second, Monster retains archaic drop-down boxes for ‘Categories’ and ‘Job Location’.  That’s really lame.  The industry standard these days is one or two search boxes for keywords and location; with the underlying application interpreting the user input and providing filters for subsequent refinement. 

   

Third, interstitial ads still separate users from the content they seek … imagine if Google forced us to view full-page ads before returning search results?  Would Google still be number one?  I’d bet against it.

   

And finally, the search results are terrible.  The screen capture at right shows the search results for Auto Sales in Boston, MA … supposedly in relevance order.   I say supposedly because not a single result is relevant.

   

The same search on Indeed.com returns four relevant jobs in the top five ads … that’s relevance that users can see.

   

Sorry Monster, I like your new duds, but your act is wearing thin.  With your resources, you could be the best in the business.  Why do you settle for less? 

   

Remember, today’s the day.

   

Job Search Agents – Is that a 6 or a 9?

   

You’ve probably seen it … the Subway commercial showing a woman buying a one-size-fits-all meal.  At first she orders a number ‘6’, but then changes her mind and asks for a ‘9’.  No problem says the smiling clerk, as he spins the box around. 

   

I swear the same clerk moonlights at the job boards.

   

Six months ago I posted about the irrelevance of job search agents … I said I was disgusted … little has changed.

   

Almost every day for the past six months I’ve received emails from CareerBuilder and Jobster with “new jobs matching auto sales in Boston, MA” (my test criteria). I’d stopped looking at the emails because the ‘matching’ jobs were usually anything but.

   

But it’s a new year, and thoughts turn to new opportunities.  Could it be that job search agents have improved?  Could it be that there are real opportunities that match my simple criteria?

   

Uh, no … it appears not.  The screen captures at right (click the images for larger versions) show the results are wholly irrelevant.

   

I’m guessing (and would love real numbers if anyone has them) that 10% of workers are active jobseekers, while another 20% are semi-active … occasionally checking a job board, but more likely relying on job search agents … with the remaining 70% of workers completely oblivious to job opportunities unless personally contacted by trusted individuals. 

   

If true, consider the implications for advertisers … the cost per qualified impression could be cut by two-thirds if job search agents delivered relevant opportunities to semi-active jobseekers.  And since advertising spending follows normal supply-and-demand principles, it seems clear that overall spending would increase if cost per qualified impression decreased.

   

Why job boards continue to ignore the opportunity to increase revenue (and customer satisfaction) is beyond me.  And unfortunately, employers and jobseekers are paying the price for their inattention.

   

The Ultimate Job Search Engine

   

Are you looking for (or trying to build) the ultimate job search engine? Jeff Tokarz, founder of Just-Posted, asked me to describe the key features that jobseekers look for in job boards and job search engines.

   

Check out the post on Jeff’s blog and let me know if I missed anything that is important to you.

   

And Jeff, thanks for asking!

   

Bob  :-)

   

The Polls are Open for the 2006 Recruiting.com Awards

   

Have you voted for your five favorite recruiting-related blogs?  If not, the Recruiting.com survey takes only a few minutes.  Not sure how to vote?  Here are five voting tips:

   

1)       Say thanks to Recruiting.com.  The rules may prohibit voting for them, but no matter – Jason, you have my vote!

   

2)       Vote for the blogs that make you think, laugh and learn … of the many great ones, five have meant the most to me during the past year: Cheezhead, Jason Goldberg’s Jobster Blog, Amitai Givertz's Recruitomatic Blog, EXCELER8ion and Talentism.  Thanks guys, and Shannon!

   

3)       Avoid the temptation to vote for blogs tugging at your heart strings with shameless, self-serving pleas, like …

   

- “If you … want to be nice to me...”

- “Thanks to your efforts this site has (sic) now a Finalist” … uh, there are no ‘finalists’ … every nominated site, even the ones nominated by the author, are still in the running.

- “If you like this blog, take a minute and please vote for it.”

   

4)       Ignore suggestions from some bloggers that you vote “five times for them”, or that you “vote early and often.”

   

5)       Use your vote to send a message to newbies and casual readers … a message that says “When you see the Recruiting.com best-blog logo, you are viewing a trusted, informed source that stands above the masses in the recruiting blogosphere.

   

So, who has meant the most to you during the past year?  Who would you recommend to a newbie?  Let your voice be heard.

   

I’m Liking Just-Posted and Jeff Tokarz

   

Relevance … specifically job search relevance … and the lack thereof on most commercial job boards … is one of the reasons I blog.  The technology exists to return more relevant job search results, yet the market does not seem to care.  Why?

   

Perhaps the market doesn’t care about relevance because:

   

-          better search reduces page-views per user, and sites are measured by traffic metrics rather than customer outcomes;

   

-          advertising dollars are inversely related to market efficiency – the less efficient the search process, the more advertisers are willing to spend to be heard above the noise;

   

-          the most profitable player in the market … Monster … does not deliver superior search results; or perhaps,

   

-          market participants … jobseekers, employers and recruiters … don’t care about search relevance.

   

I’m not willing to concede that market participants don’t care about search relevance.  I’m convinced that if a job board delivers broad coverage and superior search at a fair price, market participants will migrate to this site and away from less efficient players.  Not surprising then that I’m liking Just-Posted and Jeff Tokarz. 

   

The Just-Posted site combines broad market coverage (using spidering technology ala Simply Hired and Indeed) with substantially higher search relevance … my initial tests indicate a 70% improvement in accuracy as compared to the industry average.  And Jeff is one of the few other voices in the industry who consistently advocates for improved relevance.

   

If you too care about job search relevance, keep an eye on Just-Posted and Jeff Tokarz.

   

The Irrelevance of HotJobs

   

It’s been five months since I last looked at HotJobs.  The picture wasn’t pretty.

   

Back in June 2006 I reflected on the disconnect between Dan Finnegan’s statement regarding relevance and the reality of HotJobs.  Specifically, Dan told Joel Cheesman:

   

“We’ve learned something really important … consumers gravitate to the content that they see as the highest quality and most relevant.”

   

Yet, my analysis of HotJobs suggested that they were near the bottom of the pack when it came to delivering relevant results.   See the following references:

   

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

   

Surely, five months later, in light of Dan’s statement regarding the importance of relevance, HotJobs would have improved, right? Unfortunately, no; HotJobs still has the lowest relevance of the big job boards. 

   

The screen capture (click the image for a larger view) shows today’s search results for a ‘transportation supervisor’ position in ‘Portland, OR’ … ten jobs listed … no relevant results!

   

As I said five months ago, it’s hard to believe that with all of the resources of Yahoo at their disposal, HotJobs is unable to deliver more relevant results.  Makes me wonder if Dan really knows what is going on at HotJobs.

   

Jobster Commits to Improved Search

   

Late last week Jobster rolled out changes to their search engine – you can read about the changes here. 

   

By far the most important change is illustrated at right – Jobster wants your help to improve the search experience.

   

This is huge for a couple of reasons:

   

First, most job search engines take the position (publicly) that search is good enough.  So it’s refreshing to see a market leader ask for user feedback.

   

Second, despite rapid innovation in the recruitment space, search technology has remained essentially unchanged for the past ten years – we still enter a couple of keywords, we still hope that employers are using the same words, and we still sift through volumes of marginally related results – or worse, we stare at some variant of the message ‘no jobs match your search criteria.’ 

   

The problem is simply this – keyword-based search is a poor proxy for intention-based search.  If I search for a ‘computer programmer’ job, I’d also like to see ads for ‘application developers’ and ‘software developers’ – in fact, I want to see all jobs that focus on my intention – application development.  Today’s keyword search technology doesn’t come close. 

   

So even though the vertical search engines (e.g. Indeed, Jobster and Simply Hired) have dramatically simplified the job hunt by rounding up ads previously hidden on thousands of low-traffic sites, much of the gain in jobseeker efficiency is lost through ineffective search.  And of course, when jobseekers have trouble finding the best-fit ads, employers don’t see the best-fit candidates – all parties lose.

   

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!

   

Monster Beta Job Search – Improved but Flawed

   

Monster no longer dominates the U.S. online recruitment market – Career Builder has passed them in revenue, job postings and traffic – but Monster has deep pockets and a dedicated R&D team.   So don’t underestimate the power of a wounded Monster.

   

Innovation is a surefire way to disrupt the status quo; so if Monster can leverage their R&D spending to jump ahead of competitors in the areas of search relevance and user productivity, then perhaps they can reestablish their dominance.

   

Enter the much-improved Monster Beta Job Search.

   

The user interface is clean and flexible, plus it scores well on the use of standard usability principles. Definitely first class.

   

The addition of filters on the left mimics other sites, but it is still a helpful addition. And Monster becomes one of the few sites to offer a filter by occupational category – a key to improved relevance and productivity.

   

Weaknesses remain however:

   

1)       The occupational categories used by Monster are non-standard, poorly defined, and overlapping.  In the example above, there are separate categories for information technology and computers, software – aren’t computers and software part of information technology?  I have no idea what distinguishes one category from the other.  And if I’m confused during the search process, was the employer any less confused when assigning a category?  When will a job site turn to the combination of a proven occupational taxonomy and expert code assignment?

   

2)       Keyword search manipulation is not transparent.  Monster deserves props for going an extra step to help job seekers – when a search string finds no results using the default ‘AND’ operator, Monster reruns the search using the ‘OR’ operator.  They just don’t tell you when they’ve done this.  So in the example above, I’m seeing an ‘Accounts Receivable Clerk’ as the most relevant result for my ‘Ajax Software Guru’ search.  Definitely confusing.

   

3)       Coverage is limited to ads placed on Monster – somewhere under 1 million ads, and probably closer to 500,000. Monster doesn’t publish this metric because it would be hard to reconcile with their advertising claim of providing “access to the largest database of jobs”. So even if Monster does develop the best search engine, they’ll still only be offering access to 20% or less of the online job ads. 

   

An improved search experience is always welcome; and even with the current flaws, Monster’s new beta job search is improved. What I see so far won’t be enough to change their market fundamentals though – they’ll continue to see declining market share for as long as they charge premium prices for less than premium results.    

   

Do Standard Occupation Codes Work?

   

“I do not think ‘standard’ codes work.”  That’s Jon’s response to my Targeting Your Job Search post.

   

Jon goes on to say:

   

“There are way too may job titles to cover all the opportunities on the internet.  In addition jobs vary greatly based on each company.  For example:  Do you think a project manager for Coke needs the same skill set as a project manager for Amazon.com?  There are simply too many variations out there to try to "box" jobs in.”

   

Good feedback Jon. 

   

I have to disagree with you though, because many standard coding systems do work.  Zip codes work for mail delivery.  Clothing size codes work pretty well … at least better than if stores had racks and racks of clothes with no size codes.  And libraries are organized using standard codes … it’s much easier to find the book I’m looking for when the books are in order by code.

   

Still, most standard coding systems used for jobs fail for six reasons:

   

1)       Too few codes – in It’s Alive! … Windows Live Expo Jobs I complained about the limitation of their 34 standard codes.  Thirty-four categories do not provide enough precision to be useful.

   

2)       Too many codes – paradoxically, having too many codes is no better than having too few codes.  That’s why In Tagging Jobs for Search I disagreed with Charlene Li on the value of user-generated tags.

   

3)       Poorly defined codes – it’s not enough to have a list of code titles.  To be useful, each code should be defined and provide a list of commonly included job titles.  Better yet, the definitions should list excluded codes.  For example, look at the ONET definition for wholesale and retail buyers, except farm products.

   

4)       Overlapping codes – if a job site lists the standard codes ‘manager’ and ‘sales’, and you have to choose one, which one is correct for a ‘sales manager’ job?  To succeed, standard codes cannot overlap.  That’s why ONET provides many different codes for sales-related jobs, as illustrated in the image at right.

   

5)       Untrained users – standard coding systems are like foreign languages.  Not surprising then that infrequent users aren’t fluent.  So asking untrained users to assign standard codes results in unacceptable error rates.

   

6)       Apathy – adding accurate occupation codes to jobs costs money; so as long as users remain satisfied with the limitations of current keyword-focused job search engines, there is little motivation for adoption.

   

With all of the above problems, I can see why Jon doesn’t believe in standard job codes.  But by combining a well-researched occupational taxonomy (e.g. ONET) with an accurate automated coding system (e.g. O*NET-SOC AutoCoder) the problems are eliminated, and standard codes do work for grouping related jobs.

   

So, what does this mean for Jon’s project manager jobs at Coke and Amazon?  First, we’re left to guess about the specifics of these two example jobs – is the job at Amazon for a software engineering project or a product warehousing project?  Is the job at Coke an R&D project for a new flavor of Coke, or a cost reduction project? 

   

Since project managers must have the skills associated with the type of projects they manage, ONET does not group all project managers together.  Without more detail, O*NET-SOC AutoCoder offers several management occupations as possibilities.

   

In the end, no coding solution solves every problem (e.g. library books can still be hard to find); but accurately structured data provides a better starting point than a pile of unstructured data.