Subscribe to Dr. Granville's Weekly Digest

This blog post criticizes Microsoft's hiring policies, but it also offers solutions to fix  the problems.

Microsoft has for decades asked the government for more H1B visas to fill their data engineering positions, but recently they've changed their rhetoric: according to their general counsel, they now want companies hiring H1B's to be punished by heavy tax penalties, and use the money to train local US employees to fill these positions, read this Seattle Times article for details.

We believe that Microsoft's argument is bogus. The real causes for not finding analytic talent are:

  1. Being very picky about culture fit, thus rejecting many qualified candidates.
  2. Offering low salaries, e.g. 95K for Ph.D. Statistician with 10 years of experience, $20/hr for a recent Romanian immigrant for a coder job. If you pay peanuts, you only get monkeys.
  3. Many people don't want to work for Microsoft: they've unsuccessfully applied in the past (MSFT receives 6,000 applications per day, the vast majority are never even read) and don't want to apply again, or they are worried that Microsoft is becoming a dinosaur like GM, with less and less job security, working on un-competitive products. Creativity at work is discouraged. Finally some very strong candidates fear for good reasons that by showing up at an interview, Microsoft is only interested in stealing their IP, not in hiring.
  4. Microsoft does not advertise on niche job boards (e.g AnalyticTalent) where most modern candidates look for jobs (this issue is easy to fix).
  5. Their online application forms frequently lead to nowhere (MSFT needs to do more QA on their job portal, or use a better vendor). 
  6. Their recruiters don't read great relevant blogs posted by leading data scientists living in their backyard: these blog posts (e.g. about fraud detection, advertising analytics, worst business practices)  attract recruiters from many hiring companies nationwide, to discuss job opportunities with the blog posters. But not Microsoft. (By the way, this is a reminder that posting a great blog post with us will boost your career).

The issue with low salaries is easy to fix. First tell applicants that there is no income tax in Washington state, and costs of housing are far below California or the East Coast. In addition:

  • Hire much fewer employees but hire more efficient ones and pay them better salaries
  • For coding jobs that can be done by drones, get them done by robots that write code, hire fewer drones and pay these drones even less money. (maybe start developing a programming language to write programs that write programs)

While Microsoft's workforce is much more diversified than many companies that attract top talent - you have more overweight, more older worker, more smokers, more slow speakers and more women than in many analytic departments in great companies such as Google, eBay or Amazon (I see these MSFT employees when they regularly show up in my favorite local restaurant) - it plays against them when recruiting analytic talent, perpetuating the same personality types.

My solution for Microsoft: create 50 small start-ups of 20-100 people, owned by Microsoft, and treat them as if these start-ups' clients were various Microsoft groups. Don't mention that these start-ups are actually Microsoft business units owned by MSFT. And get true successful start-up entrepreneurs to manage these entities and let them hire new employees using any way they see fit. All of a sudden you will be filled with these great analytic talents that you believe don't exist. And in addition, Microsoft will be able to do some fiscal engineering  and make a few more bucks by moving profits and losses around these start-ups. And also, avoid regulations that apply only to large businesses (>50 employees).

Related articles:

 

Views: 3357

Comment

You need to be a member of AnalyticBridge to add comments!

Join AnalyticBridge

Comment by Mike Kennedy on November 16, 2012 at 10:44am

Interest points but I can't really say I agree with much.

  1. Picky Culture. Considering Microsoft "being too picky about culture" is completely subjective.

  2. Do Data Scientists choose a firm based on higher salaries? Without data supporting this claim, who is to say top data scientists are even driven by the money? As Talent Analytics CEO Greta Roberts said in her Data Informed podcast, based on research, money isn't the biggest factor. To attract top talent organizations should look for a natural curiosity and focus candidates on the opportunity to solve challenging problems. Not saying salary is irrelevant, it's just not the largest factor.

  3. Based on this, it seems like Microsoft needs to improve their analytics employer brand. Analytics managers consider the brand to be a key ally in attracting top analytic talent, so it appears Microsoft has work to do.

  4. This may be the case but word of mouth, opportunity (not salary) and employer reputation are huge when attracting analytic talent.

  5. This is a process problem, not a data scientist problem.

  6. I agree with you on the blogs here! Lots of insight at my AnalyticBridge blog here.
Comment by Randy Bartlett on November 12, 2012 at 9:37am

This is an intriguing 'start-up' solution that other large corporations should consider.  Innovation is difficult for large autocratic corporations and that is what is wanted to keep them relevant. 

Comment by Lynne Mysliwiec on October 15, 2012 at 9:04am

Thanks, HURT - you didn't say anything wrong, and chances are that Mr Granville is probably my age, which is why his attitudes are so unfortunate.

My beef is with two statements:
- "you have ...and more women than in many analytic departments in great companies ... - it plays against them when recruiting analytic talent, perpetuating the same personality types.
- "Microsoft has probably been hit by various hiring discrimination lawsuits, so they must hire more older workers and more women I guess."
The secret to hiring good talent is to screen for and attach value to the attributes that get the job done - work ethic, training, creativity, experience, ability to work with others, and analytic thought process. PERIOD. To me, that's independent of gender -- either you can do the job or you can't -- what's hanging between your legs (or not) has nothing to do with anything.

Comment by HURT on October 12, 2012 at 2:41pm

Don't worry, Lynne, Vincent is probably much older than you :):):)...er...did I write something wrong ?

Comment by Lynne Mysliwiec on October 12, 2012 at 2:28pm

And what's wrong with hiring women, Vincent? (with right eyebrow raised with curiousity as she waits to find out how he's gonna tap dance his way out of a clearly sexist and highly irrelevant comment)

Comment by HURT on October 3, 2012 at 9:33am

Excellent :):):)

Comment by Oleg Okun on October 2, 2012 at 12:19pm

I think being obsessed with cultural fit of job applicants is a trait MSFT shares with many other companies. What is culture and why should it be static? To me, there are only two types of cultures: "stay in the line" (inhibiting innovation) and truly innovative. Everything else (e.g., hiring based on one's look or hobbies) is superficial. In order to be successful in attracting top talent, companies need to hire ONLY based on professional skills, thus making the rules extremely simple and straightforward for job seekers.

Regarding recruiters/hiring managers, I think that they must shift from the exact matching based on experience (we need somebody with Java skills; does an applicant have them?) to the fuzzy matching based on competence (we need somebody with Java skills; an applicant knows C#, hence, she can quickly master Java as both languages are quite similar).     

Comment by Vincent Granville on October 1, 2012 at 10:58am

A few more comments regarding my article:

Microsoft has probably been hit by various hiring discrimination lawsuits, so they must hire more older workers and more women I guess. Then the trend feeds onto itself. While they might save some money as salaries tend to be lower for some of these categories (In the US, is there is a negative correlation between obesity or female gender or smoker or slow speaker and salary, for the exact same job?), they must have bigger health insurance premiums to pay.

All in all, I didn't say they were more older / overweight / etc. than (say) GE and IBM - I believe that they represent the average US population. But certainly very different from Bay Area data scientist working hard and efficiently on truly promising and competitive technologies: these guys, unlike MSFT employees, do not represent the average American.

Comment by Amy on October 1, 2012 at 9:43am

Working for Microsoft is no longer viewed as being successful, it might even been perceived as a failure by some.

I don't understand why they want to train more engineers in US. How are they going to find the trainers if they can't find employees in the first place? Who is going to train the trainers? And what will happen to the trainees after graduation? Growing the rank of unemployed software developers, and getting hit with school debts?

Comment by Dr. Antony Browne on October 1, 2012 at 7:51am

However they do find talent here in the UK (obviously we Brits are special, although we drive on the wrong side of the road and drink warm yeasty beer) - in that they fund a big Machine Learning research centre in Cambridge.

 

Tony

© 2014   AnalyticBridge.com is a subsidiary and dedicated channel of Data Science Central LLC

Badges  |  Report an Issue  |  Terms of Service