“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?”
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. 
