Sat, 21/06/2008 - 22:20 — peterv

Below is my current vision for AllejobsinLeuven 2.0 => AllejobsinVlaanderen. I would very much appreciate your feedback, preferably as public comments below, but private comments are welcome too.

What is the problem?

I want to solve two problems for job seekers, initially focusing on the high-tech and ICT sectors:

  1. while we all know there are lots of jobs available, there is currently no way to get a clean, complete list of all relevant jobs and employers in a certain area (like a "telephone directory of jobs")
  2. and even when we do have a complete list, there is no organized way to know upfront (optimally before sending out the CV) how well the employer's company/management style will fit with the expectations of the job seeker.

In parallel, the employers also see significant problems in filling their job openings:

  1. they have limited ways to differentiate amongst the large number of other employers seeking additional people.
  2. placing job ads in the classic media (both paper and electronic) is perceived as very expensive and not yielding a lot of qualitative results. Particularly smaller/start-up type companies perceive it as unrealistic to invest these amounts of money (typically 1,000 to 5,000 EURO) in this media and are not expecting a lot of ROI.
  3. on the other hand, employers also receive large amounts of non-suitable CV's (relevance rates there seem not more than a few percents), even from places all over the world, but the limited supply of suitables candidates forces them to also sift through these large amounts of mostly non-suitable CV's (the "false positives" problem)

What to do about this?

I see the solution in 2 steps, similar to the classic technique of brainstorming.

Step 1: go wide in the initial search and make sure you catch "all" the ideas

==> Use different techniques to approach a "complete" and "up-to-date" list of all jobs in a region (tagged per sector, initially focusing on the high-tech and ICT sectors):

  1. to reach initial critical mass, do manual and automated scraping of published jobs on web sites of employers
  2. ask and allow employers to supply their jobs (zero cost) (either manually, through RSS, XML, self-service scrapers ...)
  3. ask and allow the community to hint new jobs, corrections and removed jobs for the site, to supply scrapers, etc.

Step 2: narrow down the target list to a few employers to which to send a CV

==> the job seeker uses the recommendation system of our site, combined with his network to check a possible match, optimally before starting the formal (and potentially risky) path of sending a full CV to the employer.

  1. ask people to publish recommendations for the employers they like and explain why they like this employer (this will be non-anonymous, with a verified identity).
  2. if people (both employees and people outside of the employer) want to recommend an employer under their full name, this will create quite some "goodwill" towards this employer and give a more finegrained view of reasons why it might be a good idea to apply for a job at this employer.
  3. this will increase the chances for finding "links" between the social network of the job seeker and the social network of the employer.
  4. for the employers that want to take part in this system, this will add a tremendous amount of visibility and will also allow job seekers to self-assess if they see a good fit with this employer.
  5. I don't expect this to be "neutral" towards employers ... certain employers will attract a lot more "network visibility" than others, and this will certainly help to yield visibility among job seekers ... if an employer feels disadvantaged, nothing prevents him from also joining the network and motivating people in his social network to write up recommendations for him. Companies with a very closed business culture will probably benefit less from our site, too bad ...
  6. maybe certain employers will even be prepared to accept "critical" reviews, as long as they are non-anonymous, stay within the lines of decency and there is a fair "right of answer". Friends have told me they would immediately consider working for any company that wants to be that open about its employment environment.
  7. [experimental] : to solve the problem of the large stream of mostly non-suitable CV's arriving at employers, I am also considering objective systems to allow job seekers to target employers in a prefered way. The first idea is that a job seeker in our system, may send 1 "VIP CV" per week to an employer in the system. This will force the job seeker to self-assess his best matching employer of the moment and will allow the employer to get a small volume stream of self-assessed (and hopefully much better matching) CV's. Probably better systems are possible, but this was one idea that allows objective selection by self-assessment (further refinement of this idea is certainly possible).

Where is the money coming from?

  • advertisements (standard, old-fashioned model, but it works to a certain degree). I need to focus more on text based advertisements, as these can be better targeted to search words and are less disruptive for the visitors (see the 10 things Google knows are true while some items on that list are debatable ...).
  • [NEW]: this is the main new element compared to version 1.0: getting paid (at cost) for all employer related work I do. That is:
  • try to make scraper development as much self-service as possible, so I have no work there
  • ask a small fee (at cost) for developing a scraper or setting up an incoming XML/RSS stream
  • ask a small fee (at cost) for setting up and maintaining the recommendations system for an employer

What are my hopes for the site?

  • become the community site about jobs. Hope that for a significant number people, this will become "our" portal site to jobs.
  • try a more "open business model" (strongly user driven). I come from an open source background and see little problems in beeing fairly open about these ideas and the way of realizing them.
  • having fun at programming/software architecting
  • lower the friction cost of job matching for our local economy
  • helping users find more fun, challenging and rewarding jobs
    • ==> more happy employees
    • ==> more productivity for employers without "cloning" people

    So, what was wrong with allejobsinleuven.be version 1.0?

    • after giving it some thought, I believe the fundamental reason I doubted was the imbalance between the timing of costs versus revenue. The concept of me taking up the initial investment of following up all jobs in a certain region was a nice idea to solve the "critical mass" problem, but it created a different problem: that of requiring a continued, major upfront investment for keeping the jobs updated that is much larger than I had anticipated. This was in practice the direct reason for putting the previous version of the job search on-hold and switching to a more static employer list that requires far less maintenance (but also offers less value to users). And taking the job search off-line for a certain time, gave me the "air" to think about the next version.
    • there was no Web 2.0 content (while I talked about Social Networking, a networking system was never implemented, because I did not feel certain enough about the right way of doing it). That is probably the major cause why I was yet unable to trigger a community effect and create a community around this.
    • limiting myself to "Leuven" area, while creating focus, also limited the possibilities for enthousiasts outside of the Leuven region to jump in and help build the brand (and in return profit from the brand for attracting attention to their company as an interesting employer). I still believe it was good to confine the first experiment to the Leuven area (suppose I had promised to track "all" jobs in Flanders ... that would have been _completely_ impossible at this limited level of initial investment).
    • I did probably deviate too much from "scratch your own itch". The initial problem statement resulted from my quest for a new job in Leuven, but I had rather envisaged a job in technology or ICT. I expanded that to jobs in all fields, and maybe that also was an overreach ...

    How could allejobsin* 2.0 look like?

    • align the timing of cost and revenue: to avoid my huge upfront investment for covering all jobs in a larger region at my own expense, I am developing a self-service for employers to configure their own scraper or will ask to pay me a small fee for setup and maintenance (if the fee correctly covers the handling cost, this resolves my finance timing problem), or allow other forms of supplying jobs that require less of my time (and any help from contributors is welcomed of course, even just to scrape the jobs of your own employer :-).
    • start with the believers: I will no longer take responsibility myself for assuring that "all" jobs in a certain region are in the system. This will be somewhat confusing with respect to the name "allejobsinVlaanderen", but because posting jobs is essentially free it is still "possible" to have all jobs and, as the relevance of the site grows, we will get closer to that ideal of covering really "all" jobs. One somewhat naive idea I had at a certain time is that really motivated people could take up the challenge to "adopt a letter" in their region and follow-up the jobs at employers starting with a certain letter in their certain region. Getting some mechanism of people adopting ranges of employers (equivalent to Debian package management, or Wikipedia) would certainly be a nice mechanism, but maybe I am dreaming here ...
    • User Generated Content: start publishing User Generated Content in the simple form of allowing people to write "recommendations" for an employer (which I still believe quite some people that are networked/open communicators would do, without receiving a direct compensation, just to help their peers make good choices and help the employers they esteem find well matching employees). Next to UGC, redactional content about an employer would also be valuable.
    • cover a larger area: open up the area to entire Flanders (=> http://allejobsinvlaanderen.be, but redirected from localized sites with filtering arranged, such as allejobsinLeuven, allejobsinGent, allejobsinAntwerpen, allejobsinAalst etc.). This will allow early adopters to join in, independent of specific location in Flanders. The concentration of "early adopters" that want to invest some time or money to have their jobs in the system will be small in the beginning (like in a "sparse matrix"), so a larger range is required to get enough volume. Also, from the viewpoint of the job seeker, for certain very specialized jobs, you really don't want to confine yourself to one city and search in a larger area.
    • focus on jobs in (high)-technology and ICT. I am personally most interested in this area, so following the "scratch your own itch" approach, it might be best to start here. We can always expand to other industry sectors later. Jobs in other industry sectors are not excluded, we can easily separate them out with simple tags.

    Based on these principles, I am quite convinced it will be possible to grow the concept in a way that is economically viable. At a certain point, additional investment will probably be required, but by using the above ideas I am confident to get as far as possible based on customer revenue and limited investments.

    Still, I have taken contact with a few serious investment funds and I am certainly open to any partnership proposal. But I actually prefer a community discussion first, as the users are the ones that will break or make this implementation of the idea.

    Many thanks upfront for your feedback :-)