The education as a service industry is booming these days. MOOCS (Massive Open Online Courses), online masters programs, bootcamps, and even YouTube videos offer today’s curious learner a massive menu of high quality options.
So it should be pretty easy to skill up and land a sweet job right? Wrong! Stacking certificates of completion and online degrees may feel great but it often won’t get you closer to your dream job – especially if that job is in another industry.
In today’s post, we will explore why despite the veritable explosion in high quality educational content, learners without a traditional degree (in the field they desire to work in) still face an uphill battle. We will also discuss ways in which we can get around this obstacle in order to offer new grads and career changers more opportunities to get their foot in the door.
Connections Matter… A Lot!
Because everything that matters in this world can fit in a 2 by 2 matrix (just kidding, only management consultants think this), here is one for slotting the members of the workforce into one of four categories:
- Connected but inexperienced – you lack experience but know a lot of people at the companies you desire to work for. Since you mingle and network well, employers like you, viewing you as high potential and worth taking a gamble on.
- Connected and knowledgeable – no job worries here, just keep vesting those shares in Google (or Facebook or Airbnb or Lyft).
- Hermit and inexperienced – you faceless endless rejection in the frosty, barren tundra that is cold applies (a referral-less job application). And when you do finally obtain the chance to interview, the hiring manager quickly shatters your dreams of gainful employment by zeroing in on your lack of relevant experience.
- Hermit but knowledgeable – your years of hard-won experience earn you a reasonable and steady paycheck as an individual contributor (IC).
So if you are stuck in the inexperienced hermit box, you have two primary ways of getting out:
- Meet relevant people, hit it off with them, and hopefully go work at their companies.
- Get educated and knowledgeable to increase your chances of obtaining gainful employment. Note that while experience trumps knowledge, there is a "chicken and the egg" problem where I need a job in order to get the experience required to get a job.
The Problem with Education as a Service Today
Today’s education as a service offerings primarily focus on up-skilling. They try to teach you the in-demand skills of the moment so that you move in the direction of the gold arrow below – from unemployed and inexperienced hermit to knowledgable and employable working fellow.
The problem with moving in this direction is that it’s really hard. It’s not enough to learn the stuff but you must also:
- Obtain a credential that signals to employers that you are knowledgeable – and it must be a credential that they have heard of and respect.
- Demonstrate the ability (preferably through actual paid experience like contracting or an internship) to practically apply your knowledge towards solving business problems. This becomes especially important if your credential is not top tier.
Do you see the problem? Everything is about earning credibility and there are currently only two primary ways to earn it – universities and companies. Since I write often about data science, let’s use it as an example. An aspiring data scientist trying to break into the field can earn some street cred in one of two ways:
- Attend a respectable university and obtain a degree in data science (preferably an advanced degree).
- Convince a respectable company to pay you to work in a data related role.
MOOCs, bootcamps, and other alternative offerings, on the other hand, are unable to earn you the necessary credibility. They may teach you the necessary skills (in my experience, they do a pretty decent job at this). But it’s kind of like that old philosophy cliche:
"If a tree falls in a forest and no one is around to hear it, does it make a sound?"
Even if I developed a deep understanding of machine learning and statistics through reading books, blogs, a MOOC or two, and a bootcamp, I would still probably get rejected from the job I want. Employers just don’t put much stock in a certificate of completion.
And it is a gamble from the employer’s perspective: given that you have no data science work history and no one to vouch for you that you truly know your stuff, they can only take you at your word. And unfortunately not every employer is willing to do that – people, including hiring managers, are generally risk averse and willing to pay a premium for the safe, "proven" candidate (hiring the wrong person is painful from both a productivity and financial perspective).
Yes I know. If you are full of drive and initiative, you can place in Kaggle competitions, meet industry insiders at conferences and meetups, build up an awesome portfolio of data science projects on your own, and work pro-bono for an early-stage startup (all while interviewing). But that is not for everyone. Bootcamps and their education as a service competitors, while cheaper than a traditional college degree, still cost a fair bit in terms of time and money ($15,000 to $20,000 in tuition and at least months of your time). They also emphasize in their marketing how employable you will be upon graduation. But without connections or a prestigious enough certification (which even the best bootcamps lack), your completion of the program is just like a tree falling in a deserted forest with no employers around to see it or care about it.
How Can We Fix This?
I agree that employers should be more open-minded. More opportunities to prove themselves should be given to career-changers and candidates from non-traditional backgrounds and self-acquired knowledge (and experience) should be viewed more equally to a prestigious degree. But sadly, wish all we want, employers probably won’t change how they hire (and if they do, it will be very slowly).
So it’s up to the education as a service industry, which is still reasonably young, to figure out how to deliver the most impact. Going back to our 2 by 2 matrix, it should be pretty obvious what I am going to suggest (there are only so many ways we can move):
Notice how the arrow now moves up and modestly right. Yes, the knowledge building part is still important but we (I will use the royal "we" from here on out because typing the education as a service industry over and over is tiring) need to find a way to provide non-traditional learners more and better pipelines to employers.
Potential solutions must meet the following criteria:
- Reduce the stigma and perceived riskiness of hiring non-traditional learners (from the perspective of employers).
- Increase the opportunities to work on credibility and knowledge building projects. And these projects must be ones that an average hiring manager would actually recognize as legitimate.
- Increase the opportunities for mentorship and network building. Also, the mentors and industry insiders should be incentivized to help the graduate. Relying on the goodwill of others can only get us so far.
There are several interesting models out there right now. The two I find most intriguing are the Insight Fellowship Programs and Sharpest Minds.
Insight is a prestigious and highly selective 7 week bootcamp (basically the Stanford of data science bootcamps) with great connections to employers. You can tell from its schedule how well-connected it is as Insight fellows spend all of weeks 5–7 (nearly half their time) visiting and presenting their data project (that they develop prior to week 5) to prospective employers. In addition, they have opportunities throughout the bootcamp to interact with mentors and alumni. So Insight meets all of our criteria. The catch is that it only accepts PhDs (for the data science fellowship), so its typical candidate is a highly traditional learner who probably needs little to no job assistance. In fact, I would argue that Insight is like a real estate agent in a hot market – just sitting between buyers (employers) and sellers (aspiring data scientists) and taking their cut of each transaction while adding little actual value.
Sharpest Minds is probably the better long term solution of the two. The company operates a platform for connecting mentors with aspiring data scientists. The mentor guides the student through a data project, provides interview prep, and opens his or her own network to the student. The mentor’s incentive to help derives from an income sharing agreement – if the mentor helps the student find a relevant job, the mentor gets a small percentage of the student’s first year salary.
What’s Still Missing?
But I think we can still do better. Insight is too selective and provides help mainly to those that don’t need it. Meanwhile Sharpest Minds relies too heavily on its mentors and their personal ability to open up opportunities for their students. It seems to me that there is still one relatively untapped resource – the companies themselves.
My proposal is a platform that matches aspiring data scientists with paid 3 month long data science projects sourced from actual companies. Here is a rough outline (my thinking on this is extremely preliminary) of what I have in mind:
- Learners (aspiring data scientists) who want to be a part of the platform must submit code for and a writeup of a data science project they completed in their spare time. It doesn’t have to be a perfect project but it must demonstrate some initiative, creativity, and passion.
- The leaner’s educational background and previous work experience are masked (revealed only after matching). Companies must select learners based purely on the project that the learner applied with or, if applicable, previous projects on the platform. This makes learner selection more merit based.
- Up to 7 people can work on a project. They can work as a team or independently depending on the company’s desires. Why independently you say? Well a company might want to see multiple independent approaches to solving a particularly vexing problem and pick its favorite one (less groupthink this way). Projects are paid upon completion of defined milestones and confidential.
- Companies are incentivized to provide projects because they can get data science and analytics freelance work for significantly cheaper than it takes to hire an experienced contractor (who charges a high hourly rate to make up for the lack of benefits), an even more expensive consultant, or a full time employee. This allows them to throw low-cost and hungry talent at high optionality problems. If the learners solve the problem then awesome, if not it’s fine too because the cost was so low. So it’s a low risk, high reward opportunity for companies.
- Companies and learners have a chance to rate each other as the project progresses keeping each side honest. If anyone’s rating falls below a minimum threshold, he or she is banned from the platform (including companies).
- Companies that provide highly rated projects or hire regularly from the pool of learners earn discounts in project rates and access to the learners of their choice.
- Similarly, highly rated learners get first dibs when it comes to new projects and interview opportunities (with companies on the platform).
- Some thought needs to be given to the algorithm for determining the rate that companies pay learners. It should be a function of company rating and learner rating (based on their application project and other projects on the platform) but must also do a reasonable job of balancing supply with demand. However, given that the value is tilted towards learners at least initially, there will most likely be significantly more learners than projects in the early phases (the rate should reflect this but not be so low as to discourage learner participation).
In this way learners can skill up and build their portfolios in a way that earns them real credibility (and money) both within and without the platform. Interacting with and presenting to actual companies should also provide valuable networking opportunities.
Companies, on the other hand, get the chance to assess potential candidates over the duration of a real project (or multiple projects) and access to a steady pipeline of passionate talent.
Finally, the inability of companies to see the backgrounds of learners guarantees that selection will be based on merit (names probably need to be masked as well prior to match or else companies can just look up learners on LinkedIn).
This random startup idea is definitely still a work in progress and I hope to continue to refine my thoughts on it in the near future. Thanks for reading and please share your views with me as well. Cheers!
More Data Science, Careers, and Education Related Posts By Me:
Are Data Scientists at Risk of Automation
How Much Do Data Scientists Make?
