Certifications are worthless. They don’t prove anything.
Etc. Bet you’ve heard those and a few more like them before haven’t you? Of the
people you heard it from, how many had taken the exams they were putting down?
Were those people really looking at it objectively? Or just based on their
exposure to someone who had taken the exams? Or maybe just read an article
(like this one!) that pointed out all the negative items?
At the time this article was written, I hold the MCP, MCSE,
and MCDBA certifications. I’ve taken a few exams! I’m no more objective than
the next guy, but I think the fact that I’ve taken the exams at least lends a
bit of credibility to the arguments I’m about to present.
I took the Windows 95 exam in mid 1998. At the time I was
doing help desk work primarily supporting Windows 95 and a ton of user
applications, taking 30-35 calls per day. After doing this for six months I
decided to venture into certification. As soon as I began to study I saw that there
was a lot of stuff in the book I didn’t know. Conversely, there was a lot of
stuff I did every day that wasn’t in the book. I passed the exam, but very
little of what I learned from the process helped me at work. Still, as you
recall in 1998 certifications were definitely hot, I decided to pursue MCSE, or
at least go as far as I could.
Soon after that I changed jobs, moving to one that focused
on databases, data cleaning, etc. Even though the MCSE wasn’t exactly
pertinent, though the rest of 1998 and most of 1999 I studied and passed the
remaining exams to become an MCSE. No boot camp, no brain dump, just me, a
couple computers and a lot of books. I learned a lot in the process.
Subnetting. Trusts. Permissions. Shortly after that Microsoft announced the
MCDBA track, so began studying for those and by the following year had passed
the two SQL Server exams. More good learning. About the same time my career
shifted a bit and I became a DBA.
Since then I’ve taken the SQL 2000 beta exams, the Win2K
Professional, and the Win2K Server exam – all with the goal of upgrading my
certifications to Windows 2000. As I began this round of exams, I started to
think more about what the certifications were testing. That’s natural, because
while there is a lot to learn about any new product, the first round of exams
had provided a fairly good base to stand on. I understood SQL well, I just
needed to focus on the new features (we’ll revisit this in a bit). When I began
to prepare for the Win2K Pro exam, one of the items I had to study was fax
support. Fax support? Maybe it’s a minor thing, but honestly – anyone who is
going to hire an MCSE isn’t going to be using the built in fax capabilities. Oh
well, can’t hurt to learn it, right?
The Win2K server exam was ok. As I started to study for
70-216 – RRAS, certificates, DNS – and then looked at what I had left, I
started to reconsider. Yes, I could learn the material and pass the test – but
should I? Maybe I was just tired, but it seemed to me that some of that stuff
was very useful (who can argue that DNS is not useful) while some fell into the
category of ‘things that will never get used’ like software routing. More
importantly, could I learn enough and be proficient enough to be employable as
a network administrator? I’d like to think that I could, but at what cost? I’m
weak in OLAP and Data Mining, .Net is coming out, XP is coming out, and Yukon
is on the horizon. What’s the most important? I’m still deciding, but I think
that for me the time has come to specialize and focus on the things I enjoy
most (and do for a living).
Ok, so now you know a little bit about me and my
certification background. Here are my reasons why you should not get certified:
#1
You WILL NOT get hired just
on the basis of holding any Microsoft certification. There was a time when you
could, but not anymore. The best you can hope for is that if it comes down to
you and another candidate, having the certification may tip the scales your
way. Depending on whom you interview with, it may work against you if they have
had bad experiences with certification holders. The premiums being paid to
certificate holders are mostly gone as well.
#2
Keeping reason #1 in mind, what will get you hired?
Experience doing the job for which you’re applying. Whether you get that
experience at work or from studying on your own at home doesn’t matter – can
you do the job? Say for example you want to be a DBA. Will becoming an MCDBA
give the experience you will need? It will not.
#3
The exams take up valuable time that could often be better
spent elsewhere. Suppose you budget 100 hours per exam to prepare – let’s say
400 hours to become an MCDBA. Now instead of 400 hours studying for the exams,
what if you used those 400 hours to read every SQL book and magazine you could
get your hands on and work through the examples? No you won’t have the logo on
your resume, but you’ll be current in your field and have some practical
experience.
#4
The exams are too general in an industry that is already
sharply focused. How many different kinds of DBA’s are there? Production,
development, the hybrid that Brian Knight has talked about here on the site?
It’s really much more specialized. Some jobs may require a lot of replication
experience, or maybe high availability, or OLAP. All of those require mastering
the basics, but it’s the specialization that makes you valuable. The exams at
best prove you have mastered the basics.
#5
The exams are heavily focused on getting you to use the
latest in Microsoft technology. For the SQL 2000 exams, XML is important.
Instead of triggers. Is it wrong to include these in the exam? No, it’s not.
But these are in addition to everything else you need to know – how can you
test all of it in 50 questions or so? Are these new features worth taking a new
exam?
We work in a fast changing industry and there are only so many hours available for study. Make those hours count. If you decide to pursue certification, do it after a calm and realistic appraisal of what you hope to
gain from it.
Please join the debate and tell us what you think.
Andy Warren
October 2001
Read the pro-certification article Certification
Should be Required
Return to the Introduction