Website Review: JobsMiner.com

In the late nineties, when I was running the career programs for my previous employer, Moore & Associates, I made sure to pass out a bright yellow handout to every client — listing about 50 employment websites that were imperative for the average job hunter to utilize at the time.

Since then, the field has diminished considerably.  Not only have many previously popular sites winked out of existence, unable to sustain a following, but the rise of the new “aggregator” class of sites has led to even further consolidation throughout the web advertising world.  In fact, over the past few years, things have been extremely stable.  Almost boringly so.  As most of you know, I’ve been preaching for years that job hunters only need to pay attention to three key employment websites: SimplyHired.com, Craigslist, and the LinkedIn.com “Jobs” page.

Has there been a disturbance in the force, however?  Is there a now a fourth significant job site entering the scene, worthy of our attention?

The early returns suggest that yes, there might be.  This new site, called JobsMiner.com, appears to have potentially cracked the code and found a way to harvest and aggregate job listings from the “untamed country” of the Internet — the social media universe.

How have they accomplished this feat?  Well, according to their own self-description, “We have created the first social media job search engine based on a unique technology called Makam. This search engine crawls social networks, blogs and forums, filtering out the clutter and presenting only opportunities that are relevant to the job seeker.”

The big question: does it work?  Does this site actually tap into a pool of job listings that, until now, were virtually impossible for job hunters to access in a convenient, systematic way?

Based on my sample searches so far, the answer is yes — kind of.  Without question, this site seems able to dig some pretty obscure job postings out of the bowels of Twitter, Facebook, LinkedIn, and various corporate blogs scattered around the web.  The problem?  Pulling data from social media job postings is like herding cats.  Nothing is standardized.  So as good as this mysterious “makam” technology may be, it’s not perfect.  It can’t always determine whether a job is located in Seattle, St. Louis, or Sao Paolo, Brazil.  And you also can’t search jobs exclusively by job title, which is a major restriction.

In a test search I ran for CFO jobs in Seattle, for example, there were 10 jobs that came up on the first page.  Of these 10, only six were actually CFO jobs — two were accountant jobs that reported to a CFO, one was a duplicate, and the other one was a nurse practitioner job with no relevant connection to the search criteria whatsoever.  And in terms of geography, one listing came up from Tacoma (close enough, I guess) but several others were from (oops!) Washington, DC.

There’s also no alert feature on the JobsMiner site I’ve been able to find, either, despite the “FAQ” section of the site saying that such a feature exists.  So unless I’m blind as a bat, and somehow missing the link in question, it seems you have to check this site manually on a regular basis, which takes time and makes it difficult to keep track of which jobs you’ve already seen — and which you haven’t.  Keep in mind, too, that the site also picks up jobs from LinkedIn and Craigslist, so you might end up duplicating some effort on that front, as well, if you’re already tracking those sites.

So in conclusion, I’m not completely sold on JobsMiner quite yet.  It’s got tons of potential, though, and to be fair it’s also still in Beta testing mode — so perhaps further improvements will be forthcoming, down the road.  For now, I’d encourage any active job hunters to try it out for a while, compare it to the listings you’re already receiving from other sources, and see if it produces any fresh results you haven’t seen previously.

I’ll be dying to hear any feedback you have around it, once you’ve put it through its paces!

Advertisements

3 Responses to “Website Review: JobsMiner.com”

  1. Again, thank you for your valuable feedback. We will certainly look into filtering by (deduced) job titles — you are correct that these may never actually be written in the first place.
    I’ll be very glad to have you revisit our site!

    Thanks again,
    Shlomi Noach

  2. Shlomi:

    Much thanks for your detailed response — I’m quite excited that you folks noted my article and are willing to provide such immediate feedback around some of the issues I raised! Makes you really appreciate the power of social media…

    For starters, I can totally understand the technical challenge involved in taking a bunch of non-standardized data and trying to filter it down into a useful set of results. I’m already quite impressed, in fact, by how successfully you’ve been able to do this, since I haven’t seen any other sites out there manage to pull it off. And I’m encouraged to hear you’re already working on a number of other features to make your site even more valuable, such as alerts for non-Facebook users and the like. I’ll be definitely keeping an eye on things and watching how things evolve on JobsMiner, with enthusiasm.

    As for your question about MY comment about searching job titles, yes, many job websites (e.g. Indeed, SimplyHired, Craigslist, to name a few) allow the user to specifically search the leads by words/phrases in the job title — versus the body text. This is incredibly useful, since you can thereby only turn up jobs with the word “marketing” in the title of the job, for example, versus a 100 other jobs that simply mention marketing in the body text of the job listing in some tangential or unrelated capacity. With your site, however, this may simply not be possible — since the users posting the jobs on social media don’t designate the “job title” in a certain place. It could be contained anywhere in their tweet or Facebook listing. So no worries if you’re not able to conquer that aspect of things, since honestly, I don’t see how it would even be possible without an actual human being screening each ad for “context” and identifying the exact title within the text.

    Keep up the great work and please rest assured I’ll continue recommending JobsMiner as a terrific new resource to my client base!

  3. Hello,
    My name is Shlomi and I’m CTO for JobsMiner. Thank you for evaluating and writing on JobsMiner!

    I would like to relate to some technical issues you have discussed.
    You are absolutely right in speaking of “pretty obscure job postings” and noting that “Nothing is standardized”, and this is undoubtedly one of the greatest challenges we face.
    Determining at all that some twitter post makes for a job offer is in itself a complex task, and figuring out the exact nature of the job is even more complex.
    The results you got from searching for CFO jobs reflect those challenges. In particular, the two job offers you got for positions *reporting* to CFOs are amongst the more obscure issues we need to tackle.

    We are working towards being more precise in location diagnosis. You got results for Sao Paolo, Brazil, which is highly undesired (at this stage we wish to only present job offers within the US). In some cases, the job location is explicitly expressed within the text. Other times it is deduced from some meta-data. And sometimes there is simply no reference, and we use more vague heuristics. Things slip through, which shouldn’t.
    The filtering algorithm is constantly being improved, and such cases should become more and more rare. Our 7 year experience with our technology in our Israeli product shows that while “noise”, as we call it, may never be completely eliminated, but can be reduced to very low percentage.

    The “reporting to” CFO is a classic phrasing issue we must deal with. We are already covering many phrasing variations; there are many more to handle. We note that people express themselves in a very relaxed manner when posting in Twitter or Facebook; it takes time to catch up with all possibilities.

    You should definitely not get “nurse” results on CFO search (not in the couple first pages, at least!), and I expect that an improved ranking algorithm should be online within a week or two, which fixes that.

    I’m not sure what you mean by “you also can’t search jobs exclusively by job title”. Was your intention to filter only job offers with exact job title as given by the user? I’ll be happy to discuss; would you care to elaborate?

    Washington vs. Washington DC is yet again an issue resulting from the strange fuzziness. Some posts that offer jobs in NYC just say “New York”, and we wish to be more relaxed in our observation. This should not make for a “Washington DC” results when you explicitly request “Seattle, Washington”, and we will need to continue our work in this issue.

    Alerts are available for users signing in with their Facebook account. We’ve fixed that in the FAQ, thank you!

    As you said, we are still in BETA stage, and are pressing hard at improving the algorithms, based on such feedback as yours. We hope and aim to provide with a good solution for job seekers by adding these very-hard-to-catch jobs offers on the net, which are usually only seen by first-hand friends and contacts.

    Regards,
    Shlomi Noach

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: