Testing in Hell - The Real Reasons I (And Others) Left Microsoft - Updated

The Real Story Behind My Reasons for Leaving Microsoft

“The horror….the horror….”
Colonel Kurtz, Apocalypse Now

My first 2 – 2 ½ years as a Fox tester with the company went well after I joined in June of 2001. I was consistently leading in most test production statistics and had engineered some substantial procedural changes that had greatly streamlined the daily test process. Mike Stewart was the Test Lead for my first 18 months there and then stepped down for personal reasons.

Shortly thereafter, the Fox Team was merged with the VS Data team and I was asked to assume many of the lead functions; albeit without the actual title of “Lead”. This meant that for about half of the VFP 8 product cycle and all of the VFP 9 product cycle, I was responsible for contractor management, budgeting, resource planning, and served, with Calvin as Lead Developer and Randy as Lead Program Manager, in a sort of triumvirate making the day-to-day decisions about the product. I was the test project manager, but the Test Manager in charge of the full-time personnel remained unchanged.

The structure of the test organization was pretty simple: There was a Test Manager, a Test Lead, and then we test folks. After Mike stepped down, there was no Test Lead.

The work sounds fun, and in many cases was, but it was also very difficult. Since the company was trying to reduce contractor budgets and VFP was not exactly the star of the stage, I kept having to squeeze more and more out of the limited resources I had. There were many times when it was suggested that I work with Randy and Calvin to cut major product features, but I was unwilling to do that unless absolutely necessary. Fortunately, Randy had excess contractor budget on the books which he lent to me and we were able to get along without major feature cuts.

Starting not very long after I joined the company, I began to suffer mood swings and bouts of insomnia. These grew worse as the stress of getting a good product out the door mounted. Sometimes I would be late to work and other days I would leave early, exhausted. Management took note of this and it was often cited as a “minus” when my work performance was discussed.

In 2002 and mid-2003, my performance reviews were above average. My arrival and departure issues were noted but weren’t considered a serious detriment to my overall performance. On top of that, my individual contributor statistics still ranked at or above those my peers in spite of my additional roles as project and contractor manager.

Starting in the summer of 2003, the wheels started to come off. My wife traditionally spends every summer with the kids at her relatives in Miami and the summer of 2003 was no exception. However, her return was postponed over and over and it became apparent that she was prepared to sacrifice our marriage rather than return. She hated the Seattle area, had no friends, and did not like the person I was becoming as the stress mounted.

And that summer, our contractor lab manager was summarily dismissed for gross violations of company policy and we had to scramble to assume control of our own lab operations – which we did with some heroic hard-work by our contract personnel.

Also, in September, I suffered a freak ulnar nerve accident (the funny bone nerve) that left me with only partial use of my right arm for some time. I had to wear; at least I was supposed to, a cast that kept my arm out at an 80 degree angle. Thank goodness, after a few months, my arm seemed to be healing on its own (I was given a slim prognosis that this might occur so I was lucky) and I discarded the cast.

In February 2004 I was served with divorce papers. At that point, I knew I had to go down to Miami and straighten things out. It took every day of vacation I had plus all of my personal days, but I was down there for three weeks and we resolved our issues. She promised to return in the summer with the kids and try to make a go of it here again.

This is about the time when the budget strings were drawn tighter and I was constantly reshuffling responsibilities. It was a very stressful environment – I suppose it would have been easy to jettison VFP 9 features but all of us in the “triumvirate” wanted very much to avoid doing that. We felt that the cool new features were worth possible cutting a few test cycles and getting as creative as possible to spread limited test resources around.

Things were starting to look decent when, in June 1 2004, while in the middle of a “leads meeting”, I received a call that my father had unexpectedly passed away. Obviously, I dropped everything and rushed to fly to my parent’s home in Ft Lauderdale since I’m the eldest child and am expected to be in the lead when tragedy strikes (my great aunt and uncle has passed away within the prior three weeks, unbelievably).

I’ll never forget as I rushed down the hall to leave Microsoft that afternoon, my manager stopped me in the hall to remind me that I only had 4 days of bereavement leave .“5”, I told him, “with travel”, having just read the policies. This was just the typical warm, people-friendly response one could expect.

By the time mid-June rolled around, I felt that I was once again on top of things. The test milestones were being met and I felt that the team was humming along well, albeit with some escalating morale issues.

One problem that we all had on the test team was that we felt we were more or less adrift and that management didn’t seem to care or pay attention to what we were doing. When attention was paid, it was to “stop the presses” and some task was thrown at us that would completely stop VFP 9 testing dead. Don’t get me wrong; sometimes these were valid assignments but they pinched our already limited resources.

I fielded complaint after complaint in our weekly meetings, which management attended perhaps 25-50% of the time. I detected a growing “don’t give a crap” attitude and I didn’t know what to do about it: I was simply the project manager and not the employee manager – it wasn’t my job nor likely would it have been appreciated if I addressed career or morale issues.

There was a growing concern amongst the three of us that neither of us had been told of any career path after the shipment of VFP 9. Management either had no plan or hadn’t mapped it out well enough to divulge it to us.

In late June, I saw that we needed to play to our individual strengths and re-engineered test specifications and testing into a factory process and our productivity started to rise. It was simply a matter of having folks doing what they did best. The productivity breakthrough this achieved was, in large part, ignored by management. It’s hard to get kudos for work like this when lingering under a preconception that you’re dirt.

In July or August came utter disaster. All three of us full-time employees were told we were performing barely adequately and all three of us received poor to lukewarm reviews. We were told our metrics were poor (with no basis for justification), and were all compared unfavorably to our brethren on the VS Data side of the team….to be fair to them, they had nothing to do with this slight.

Each of us was told our bug statistics were deficient. This was due to many factors: The tools we were now employing caught many more bugs before they became issues and reduced the number of bugs we were likely to find in ad-hoc testing. Each of us had many more responsibilities than we had during the previous product cycle and, therefore, had less time to spend spelunking for bugs randomly. Finally, I and my fellow testers were pretty convinced that the product was just that much more stable and there were fewer bugs to be found as easily as before.

Of course, we were not supposed to know that each of us was dinged this way, but as in all organizations we shared information like this privately. It was pretty amazing the uniformity of the criticism we all had received.

In many ways, and not unusually for corporations, Microsoft’s review process is highly subjective. If a manager wants to make you look good, he’ll find ways. If he wants to make you look bad, he’ll find ways for that, too. I’ll leave it to the reader and simple logic which way management was predisposed to the Fox testers.

We all reacted differently. I have to admit that I simply “lost it” and ranted and raved to the point that I was reprimanded by Human Resources. Except for the time management issues I have previously discussed and an inability on my part to be consistent with status reports, I believed – and do to this day – that I had overcome immense hurdles to keep VFP 9 on track with as many features as we were planning to roll out. I also believed, strongly, that one of the other full-time testers had performed above-and-beyond average and I could not believe the outcome of his review as well, although my ability to intervene on his behalf was essentially nil since I was not his manager and I had my own injustices to deal with.

I believe at least one of the other two testers filed complaints with Human Resources, although I’m not certain about that and it wouldn’t be fair to divulge who.

In August, I began to have panic attacks. Not knowing what they were, I was hospitalized twice before it was accurately diagnosed. My doctor ran some tests and determined that I had acute Bipolar Disorder. It probably explained the earlier insomnia and fatigue to a degree but had been brought out into full latency – likely due to the extreme personal and professional stress I had suffered recently.

Starting in September, I went on a pretty nasty pill regimen. Funny – the first day at the company I made the mistake of taking all of my pills at once, after lunch. Fifteen minutes later I had to be driven home…I have never felt so stoned in my life!

I knew that the pills were going to become an issue. I tried getting some informal recognition from my management of possible disabilities but they wanted no part of any informal deal. Human Resources wanted to know the exact terms of accommodation I might need. I couldn’t give them exact terms because my condition and the pills caused my daily moods and energy levels to swing dramatically. One day I’d be a saint and the next writing the most ripping, snarling emails to management. The constant, often senseless interactions between HR, management, me, and my doctor was a joke and nothing was ever done to allot for my issues besides putting a burden on me that I couldn’t fulfill.

During all this time I was very open to management as to where I was in my treatment and what my current issues were. I communicated every word of doctors advice or insight I could offer on my ongoing treatment and prognosis.

Around this time, management started to require that I “clock in/clock out” and notify them by email of my comings and goings – to the point where I had to use sick or vacation time for doctor’s appointments. This is, as far as I know, unheard of there.

The pill regimen really started to get to me. I remember a meeting with my manager in which he asked me if I had kicked off a previously scheduled event that day (a “bug bash”) and I could nothing but stare at him. I had completely forgotten and it took me a bit just to remember what he was talking about.

And that was the nut of my problem from then on – I was being watched over and threatened with poor performance reports for issues that I didn’t have complete control over. It was almost as if they wanted me to fail and made life as difficult as they could – although perhaps that’s paranoia which is not one of my symptoms.

I grew increasingly desperate over my career as the VFP 9 product cycle was coming to a close. I could not, and still do not, understand how we testers could all be considered mediocre and yet we were maintaining a very high quality bar for what is one of the most feature-laden versions of Visual FoxPro ever. It seemed that everyone – peers, other Team members, and customers - thought we were doing a great job except for our immediate management. It’s like, “Well, John, things are looking good for VFP but you were ½ hour late today and we just can’t have that, can we?”

I made the mistake in late October of going off my medications completely to try to get a grip on things and get rid of the fogginess. The problem with psychotropic drugs is that when they’re titrated right and the mood swings are more-or-less handled, you tend to convince yourself that you don’t really need them. Also, a key element to my particular form of BPD is what is known as “hypomania” in which you are, essentially, judgment-impaired when it’s in full force and all ideas sound like good ideas including going off of one’s medications.

As an aside, if you want to read about a fascinating mental illness, look up “hypomania” on the web. In mild cases, it’s the only actually beneficial mental disorder.

By early November, my lack of keeping up with my drugs caught up to me and I suffered what I can only term a mild nervous breakdown. I was out of work for 3 weeks and probably returned before I should have just because it was a critical time in the product cycle. Believe it or not, in my next formal performance review, I was dinged for having left at this time and forcing the manager to, God forbid, actually manage for a time. Yes, this was in writing in my mid-year review; it seems so unbelievable.

I think around the time of the full July review Mike Stewart just stopped caring. I didn’t recognize it at the time although if I had been his manager perhaps I would have seen it. From that point on, I believe Mike was just waiting to be let go or to be given an excuse to make his own exit – which eventually happened in February 2005. Mike has already returned to the company as a vendor and is rapidly rising through the ranks of responsibility.

My other compatriot, Chandra, simply resolved at that point to get out as soon as he could. He managed to do so in November, 2004, and was hired as a developer in the Windows Division. He recently received his first review in this role and it was outstanding. It’s a private joke of ours that he’s well appreciated as a developer in a critical area of the company yet was kicked around as a tester on the relatively unimportant VFP product.

So, in early 2005, I was the “last man standing”. My contractor budget had run out at the end of 2004 and, so, there were no contractors. Mike had gone; Chandra moved to greener pastures. I tried to make light of it, in fact putting a sign on my office door reading “VFP World Test Headquarters”.

Around that time I began to make noise about being promoted to the formal title of “Lead”. It wasn’t a money issue; I just wanted some recognition of the responsibilities I had in what ultimately was a very successful product launch. Essentially, I was reminded that I was a piece of crap and then divested of VFP testing responsibilities. Another team tester came in to learn VFP and VFP testing.

For my last few months with the company, I busted butt to learn how to be a VS Data tester and the VB .Net automation tools and procedures. To their credit, management paid for whatever training or materials I requested. I never fully got into it, though, as there were just too many pre-existing VFP responsibilities I had that I couldn’t yet hand over to my replacement.

Don’t get me wrong – the tester who replaced me is exceptionally sharp and will do a great job in the long term and enjoys my utmost confidence.

In June, my wife became ill and was hospitalized. I needed to stay out of work to help take care of the kids – her parents came to help but they don’t speak English – and to taxi the family to visit her at her hospital daily, which was 40 miles away in heavy traffic.

Around the beginning of July I realized that I was in a lose-lose situation. My wife’s long-term prognosis was unclear and I hadn’t yet caught up to my peers in terms of VB testing. The track record was that, of course, I would be slammed in my next full review for not only the issues I’d already been hit with, but likely my output versus the other testers in VS Data since I was still battling a learning curve as well as vestigial VFP responsibilities.

So I tendered my resignation. I could not guarantee 100% participation in the future with my illness and now my wife’s illness, so the fair thing seemed to be to resign and free up the headcount for someone who could come to the team with skills more rooted in .Net.

On one of my last days there, when discussing all of the unpaid leave I had to take with my wife’s illness, my manager told me that if it was some consolation, I was going to receive a performance bonus since I had remained employed beyond the end of the fiscal year (June 30).

A day before the bonuses were to be paid, I received an email that I would not be receiving one. When I asked for further information, I was informed that they had told me only that I was eligible for one (ahem – my memory serves different) and that my performance and behavior made me ineligible. I have asked for details and specifics of where my performance did not meet my personal goals but no information has been received.

They’re nothing if not consistent, I guess.

So what is the complete truth? This is my side of the story, with omissions due to confidentiality agreements, and they’ll have their side of the story – which I doubt will ever see the light of day. Sometimes, when my mood swings downward, I begin to doubt my contributions and my role for the last few years.

But then I think of the hundreds and hundreds of good comments I have personally received on the stability and richness of VFP 9. I think of the confidence I enjoyed from my peers; Randy, Calvin, Ken, Richard, and Aleksey. In the weekly leads meeting I was Mr. Quality Assurance and I believed – and it was true – that I was personally responsible for the quality every single change for VFP 9.

I also believe that something over there began to go deeply wrong around two years ago. If I was sitting here writing this as the odd-man out maybe I’d feel differently – but not a single one of us remains from early 2004 to now. That should be telling.

Management would never say it, but VFP 9 owes a lot, if the face of an increasingly hostile environment, to Chandra, Mike, and to our primary contractors the last year: David Anderson, Garrett Fitzgerald, and Jay Jones (who is now an employee).

These folks pulled it off in spite of what it was costing them just because they happened to be where they were and doing what they were doing. All of us; all of us were (and most of us remain) passionate about the product and our customers. I personally participated with gusto in every community event or initiative I could: The team blog, MVP summits, code competitions, minority student events, and others. I single-handedly found and filed 200+ documentation bugs in my first few months with the company and led an initiative with the documentation manager (since departed) to create good code samples for language elements that didn’t have them or were outdated.

Naturally, my efforts in these areas were denigrated at the time of evaluation and not considered germane to my assessments. Sigh; so much for supporting company and departmental initiatives – no direct management support whatsoever. Community activities were sneered at; documentation bugs were not “real” bugs.

Of the three employee testers, all of us skilled, knowledgeable, and extremely enthusiastic as of January 2004….by a little more than 12 months later, one was (perhaps) fired, one left in disgust, and I believed resignation was my only sane (no pun intended) option.

To be as fair as I can, I know that I was a very difficult person to deal with considering all of my “issues”. But I met virtually all of my strategic goals.

I am left with no alternative than to consider that the ultimate destruction of the Fox test team was engineered for reasons that I don’t know and hesitate to speculate on. I could surmise it has to do with the high visibility of the VS Data side of operations versus the continual irritation of VFP, but it would be crass to think that we would have all been maligned simply for working with a less strategic product, wouldn’t it?

What is the effect on Visual FoxPro and what is the future for Fox? I can’t say. There are things I know and things I don’t know, but I swore confidentiality and these are issues I cannot and will not discuss. I also will not name names when it comes to management and who did what…it would not be fair. In some cases I’m not sure whose hand was holding the hilt of the knife in my back.

Not long after I joined Microsoft, there was a customer event. I was chaperoning a group of our customers around and two turned to me and said (I’m paraphrasing), “We appreciate you being here. We consider you one of us and not one of them and we know you’ll protect our interests.”

It brought tears to my eyes. Well, now I’m back to being one of “you” after a few years in hell. And I’m happy to be back.