The World As Best As I Remember It : Software

Everybody Wants to Rule the World

philipsu — Sun, 19 Mar 2006 13:46:00 GMT

[An article a year ain’t bad, I suppose. After being Slashdotted on Show Me the Money and having several reporters call my home based on Brin and Bear It, it looked like a good time to take a breather. But I’m back, this time with a gripe.]

It’s become very hard at Microsoft to find good senior developers internally. Don’t get me wrong – there are still plenty of brilliant people who were once upon a time good developers at Microsoft. Now they’re just dev leads, dev managers, or vice presidents. (Either that, or they’ve gotten it into their heads to join the photogenic business majors – but that’s a whole other discussion.)

Make the Most of Freedom and of Pleasure

No self-respecting developer started off their computer career thinking, “Boy, my lifelong goal is to be a middle manager in a large multinational corporation beset by competitors, nearly collapsing under the weight of its own internal processes.” This thought is, in fact, so ridiculous that it reminds me of the eponymous Monster.com ad from the dot-com days (“I wanna claw my way up to middle management…”).

Developers want to code. Maybe design once in a while. Then code. Invent a thing or two. Code. Oh, and if things line up this weekend, perchance ask that new coworker in Accounting out to a movie... seriously, though – code, code, code. Occasionally, one of the younger ones will mutter something about changing the world, better living through software, blah blah blah. But it really all translates to, “I want to code all day.” Not a single developer starts their career wanting to be the pointy-haired boss. It’s the last thing on their minds.

For the first two decades of Microsoft’s existence, everything was in blissful harmony. The recipe for success was straightforward.

Software Au Jus
Prep time: 20 years

Ingredients

15,000 developers, preferably from college (or high school)

$285 billion dollars market cap

A metric gazillion computers (high altitude or Canada: use 1.5 gazillion)

Food

Directions
1) Spread developers on clean campus
2) Pour computers evenly over developers
3) Split market cap into portions (“stock options”), laying generously on each developer
4) Bake for 20 years, occasionally inserting food into developers as needed

What happens when you take a bunch of brilliant developers and make them millionaires? Well, nothing, really. They just keep coding. Sure, a handful starts driving to work in Ferraris. Another small set quit and begin lobbying for endangered species, or living like they’re looking for vacancies at Betty Ford’s. But the overwhelming majority of developers keep coding, because that’s what they truly love. By making them rich, you’ve just enabled them to keep on keepin’ on.

Nothing Ever Lasts Forever

But something happened on the way to heaven… that was then, this is now. What happens when you run out of a key ingredient of the original recipe? What happens when you can no longer pay developers ridiculous sums of money, the money that originally enabled them to pursue their true interests?

Well, they start thinking about their careers. Those of you not in computer science might miss the gravity of that statement. To typical developers, “upward mobility” means “finding the lever that raises the Aeron.” But developers who start worrying about their futures begin to do unnatural things. Things like asking about becoming a dev lead or dev manager.

This is now happening all the time. All the time. Just about every developer I know who is a rising star wants to be a dev lead, dev manager, royal highness, etc. People I manage. People I interview. I can’t exaggerate how outrageously often this happens. I kid you not, about 80% of the developers I manage or interview say that their goal is to be a dev lead or dev manager. Many of them – brace yourself – don’t want to take or continue a coding job unless they can be a manager. A manager! This, from the lips of real-world developers!

How do you staff and retain a team when 80% of the people you already have, and 80% of the people you interview, want to be managers? It’s unnatural.

Say That You’ll Never Never Never Need It

If you’re a developer reading this, please believe me when I say you should seriously think it over before setting your mind on becoming a dev lead or manager. Let me count the ways.

· You love coding. Sometimes you’ll lose sight of this, especially if you’re worried about your career. But ask yourself what you’d do if you were independently wealthy. If you’re like most developers, you’d rather code. Why not continue doing what you love?

· Most developers make bad managers. Typical developers excel in skills largely orthogonal to those required of managers. Typical developers also hate the things that comprise most managers’ days. Seriously.

· There is no pay difference. At least at Microsoft there is none. I get this one all the time, thinly veiled behind noble-sounding phrases like, “I want to have broader scope.” Most developers think that dev leads and dev managers get paid more than individual contributors. This is simply not true. In almost every team I have ever managed, there were always individual developers that earned more than me. Your salary has nothing to do with whether you manage people. It has everything to do with the perceived value of your contributions. Developers make the software world tick. They get paid.

· You are hot property. Teams need four times the number of developers than they do leads or managers. And yet about 80% of the folks that I talk to want jobs as leads. Do the math! It’s well-nigh impossible to find good senior developers these days. Your future is assured. Your choice of teams to work on, problems to own, etc, are virtually boundless. Seize the day.

Now that I’ve gotten that off my chest, if any of you reading this are great senior developers that know you want to write code, drop me a line. I’ve got plenty of great spots for you.

The End of an Era

philipsu — Sat, 24 Jul 2004 08:49:00 GMT

Software developers, I've got one word for you: www.yourjobisgoingtoindia.com.

That's right. Your job is going to India, princess. Anyone who tells you different is trying to sell you something.

Peek-a-Boo

It's a fascinating thing to watch children grow up. There are stages in life where their ignorance is an amusing reminder of our own. Every child goes through a phase where they think you can't see them because they're covering their own eyes. “It sure is dark around here! I bet no one sees me.”

It's the same with outsourcing. Sure, you read about it in the press. Alarmist titles like, well, “Your Job Is Going to India!” Or, “America -- Land of the Ji, Home of the Knave.”

For those of you playing along in the home game, Ji (“gee!”) is a Hindi word that's a title of respect. So, instead of Sir Charles Barkley, you might say Barkley-ji. A native of India is far more likely to say Gandhi-ji, of course, unless it's their fourth Guinness.

But no one I talk to in everyday life seems to take it seriously. We're like kids covering our own eyes, ignorant of our impending doom.

Your job is going to India. Period. Well, semi-colon. It could go to China. Or Borneo. But India's most likely. Let me count the ways:

They’re cheap. Everyone focuses on this. After all, eight developers for the price of one is hard to argue with -- it’s a veritable Wal-Mart of developers over there. A three-hour cab ride from Mumbai to Pune is $25. I don’t think a cab in Manhattan would even slow down for you for that little.
They're brilliant. I can't believe the amount of time I spend debating this with otherwise rational people. “Americans are creative! We invented the computer! Heck, we invented C++! Java! C#! Oh, and did I mention Logo? Or the mind-numbingly lame ML? Let's see Hindustan match THAT!”

Puh-LEASE! Americans are awesome at inventing technology and then being overtaken by other countries with more efficient production. Cars. Radios. Air Jordans. You name it. (And please don’t be the one person who inevitably writes in arguing that American cars are engineered better than Japanese ones. Don’t be that person. Not today.)
And just think about the sheer odds against us! The US produces about 1.3 million college graduates a year, from a population of 290 million. India graduates 3.1 million college students a year, culled from a population of 1.07 billion. They’re far more selective. Just the raw number of graduates is daunting! As if that’s not enough, more than half of them are working the local Hyderabad Starbucks waiting to be discovered. It’s as close to a geek L.A. as it gets!

They're entrepreneurial. Americans have traditionally been entrepreneurial, an advantage we’ve long enjoyed over many other cultures when it comes to capitalism. But India’s different. In general, Indian culture accepts entrepreneurial risk-taking much like the US. This is one of the main reasons outsourcing is so hot there.
They speak English. I’m bracing myself for the deluge of comments griping about Dell tech support. You’re missing the point. We’re not talking about tech support. We’re not even talking about talking. These are coders we’re talking about -- room full of socially awkward youngsters, fridge full of Mountain Dews, the works. But they can code in English. Compile in English. Bid your next contract $3 million dollars lower, in English.
They’re learning. This is the thing: even if you believe that India is currently not as good at software as the US, they’re learning from American-trained developers. Every year gobs of senior developers and software managers return to India from Microsoft alone. I know these people! I’ve worked with them! They’re great managers, and they will train a generation of highly capable Indian software developers. And don’t even get me started on just how motivated these young Indian graduates become when they’re chosen as part of the elite few to join a software company. It’s like they’ve hit the jackpot in Powerball Calcutta. You don’t know the meaning of motivation until you’ve paid a 22-year-old a bajillion dollars in local currency, contingent on his continued stellar performance.

Microsoft’s India Development Center (IDC) is growing like gangbusters – it’s like they can’t buy enough of Hyderabad. One of IDC’s main goals is to replicate the successful software development model of the Redmond campus. And Microsoft’s just one example – everybody’s doing this. The 2.3 IT companies that aren’t are going to start next month.

Be afraid. Be very afraid. And prepare to train your replacement on your way out.

Don’t cover your own eyes. Your job is going to India.

What Will Really Happen

I’ve exaggerated a little. Not all of us will lose our jobs, of course. Just most of us. Those left behind will have much lower salaries. The math is simple – the inevitability, manifest. You can choose to deny it or try to defy it, but it doesn’t make the laws of supply and demand any less true.

Most people think that they are paid (or should be paid) the value of their contributions to a company. This leads to whiny dribble I sometimes hear from Microsofties about how a portion of Microsoft’s $53 billion cash should be paid to the people who helped create that value (inevitably, they feel they’ve personally created a few millions’ worth). I totally agree with the spirit of that utopian dream – it’s a noble ideal that you should be rewarded according to the value you add to a company. It appeals to our inner sense of fairness. But the reality is that you’re paid your replacement cost. You are paid what it would cost the company to find another person who could add just as much value. So, in general:

developer.paycheck = min(developer.valueAdd, developer.replacementCost)

I’m not saying this is a good thing. I’m not saying this is a right thing. I’m saying this is a capitalistic thing. And I’m saying this is the truth. The equation of our capitalistic software development economy is getting cheap professionals added to it annually by the truckload. You bet this is going to hit your salary.

Brace For Impact

This is all fine and dandy, but it’s also horribly depressing. Stay tuned – next week, I’ll cover things you can do to brace for impact. For the pessimists: it’ll be a futile attempt to avoid certain doom. For the optimists: it’ll be a positive piece that will give you practical guidance and imbue you with hope.

Don’t ignore the issue. You need to prepare yourself.

Recruiting.com

philipsu — Fri, 02 Jul 2004 18:30:00 GMT

I'm constantly drawn back to the heady dot-com days at Microsoft. Not only because team meetings flowed freely with Dove Bars and Frappuccinos. But because the heady days created a fundamental shift in the culture at Microsoft -- a shift we may never fully recover from.

Above all else, the heady dot-com days decimated the talent pool at Microsoft. As you know from listening to execs like BillG and from reading books like Microserfs, Microsoft has always maintained that the one thing at the core of its success, the one key asset that distinguishes it from all its competitors, is its whopping $53 billion dollars in the bank.

Hehe. I'm only kidding, of course. Sure, the money helps. But it's about The People. At least that's what execs keep saying.

Two things happened to The People during the heady dot-com days:

If you were part of The Old-School People, the same people who wrote 3000 lines of code before breakfast and cracked jokes about SoftICE over the latest issue of Byte magazine, you left the company the moment your buddy, no doubt a member of The Good-Looking Old-School Business/Marketing People, phoned you up from the newly leased offices of youwouldntbelievehowgoodofafirstmoversadvantagewevegot.com.
On the other hand, if you were a member of The Talented Graduating College People, you dissed on SoftICE, grabbed a copy of Java In A Nutshell, and headed straight for thosefirstmovershavenoideawevegotapatentsubmarine.com.

I worked recruiting. I was there.

During the heady dot-com days, I often worked on a volunteer basis with Microsoft recruiters. I gave tech talks, attended career fairs, and interviewed students at several universities, and focused especially on University of Maryland, my alma mater, over a period of years. Working a Microsoft booth at career fairs in 1999 and early 2000 was like selling cigars in The Vatican. Some would stare at us with blank, incredulous looks. Isn't the DOJ thing enough already? Can you believe they're even here?! Others would walk politely by, ignoring eye contact. Maybe they'll think I'm a business major! [Before everyone jumps: for the record, Microsoft loves business majors. Two folks on my floor are Harvard MBAs. There's always room for business majors. We love you guys.]

Point is, a significant portion of the folks who even bothered applying to Microsoft during the heady dot-com days were mediocre at best. Couple that with the fact that we needed to fill positions like they were going out of style and you've got a formula for disaster.

disaster = probability(applicant[i].IsMediocre) * positionsFilled

The heady dot-com days produced a huge sucking sound at Microsoft. All the experienced folks were leaving. All the good college grads were busy working for Trilogy. Add to this the fact that we wanted to grow the company at record rates... I leave the conclusion as an exercise for the reader.

Culture Change

More than 60% of Microsoft's 55,000+ employees joined during the heady dot-com days or shortly thereafter. [Before you math whizzes pounce, remember that folks were leaving in record numbers during those years as well -- in net, Microsoft did not grow 60+% during that time.] This means whatever you've read in Microserfs, in Showstopper, in numerous magazines in the `90's about Microsoft developers, is at least 60% wrong.

This is not your father's Microsoft.

Those people you thought were here? They're gone. Well, some of them have returned, repentant, like prodigal sons and daughters. But they're mostly gone, replaced by The Next Batch.

Don’t get me wrong – we hired some great people during the heady dot-com days. We just had to work a lot harder to find the right folks, and often had to beg shamelessly. But I’m certainly not going to blow sunshine about how, against all odds, we redoubled our recruiting efforts and ended up with just as high quality folks as we’ve always had. That’s simply not true. That’s not even statistically possible. Truth of the matter is that good people were very hard to find during those years.

Microsoft Mojo was diluted because we had to hire during the heady dot-com days. There’s no doubt about it – the whole Bubble Phenom decimated Microsoft’s talent pool.

Now the Good News

The best thing to happen to Microsoft since the heady dot-com days wasn't settling the DOJ bit. It wasn't Windows XP. And it certainly wasn't the remix of SteveB screaming “Developers, developers, developers, developers!“

The best thing to happen to Microsoft since the heady dot-com days was the utter and complete collapse of the entire software industry. There were no jobs to be found anywhere for new grads, except for the thousands of jobs with the company that had bazillions in the bank.

You wouldn't believe the folks I interviewed last fall at the University of Maryland. One sophomore had taken over leadership of his dad's consulting business when he was fourteen. Another runs Menuocity in his spare time, and employs several folks during the course of business. Did I mention he was only a sophomore as well? I couldn't invent resumes better than some of the ones I saw.

Grab a Dove Bar. Snap open a Frappuccino. The heady days are returning for Microsoft.

Show Me The Money

philipsu — Thu, 01 Jul 2004 08:49:00 GMT

The Ever-Elusive Personal Finance Manager

I worked as a developer on the Microsoft Money team for a little over two releases (Money 99 - Money 2001). During my time on Money, it made its transition in the eyes of the press from an inferior Quicken wannabe into what reviewers would agree is a superior product (the dramatic improvement wasn't due to me, of course -- I just happened to have graduated in 1998). Despite this, I ultimately left the team because I was frustrated by the direction that Money was being led. Instead of making Money into what we knew users wanted, we were busy trying to match Quicken and meet MSN goals. This is the back story regarding what eventually led to my departure.

Time to take a break for a shameless disclaimer. What follows are my own opinions and perceptions of what happened. The facts I present may be completely wrong. I may discover years later that I had a grapefruit-sized tumor in my brain during my years on Money. And regardless of what I say below, the Money team was one of the most enjoyable, cohesive set of folks I've ever worked with. And I still love the product, warts and all.

The majority of consumers who buy computers claim that personal finance management is one of the top three reasons they are purchasing a PC. They've been claiming this for more than a decade. But only somewhere around 2% of consumers end up using a personal finance manager (“PFM”), with Intuit Quicken and Microsoft Money dominating the market. Both products have been around for -- you guessed it -- more than a decade. This dramatic disconnect between consumer demand and actual market penetration is mind-boggling.

Take a guess at what percentage of consumers launch Money ever again, after running it only once. You'll need to remove a digit from whatever percentage you're currently guessing. It's seriously that low. Granted, most copies of Money are actually pre-installed by the OEM on consumer machines, so you're not exactly dealing with a captive audience. But we're still looking at a huge discrepancy between expressed consumer desire and actual consumer behavior. There's a lot of money to be made if we solve this mystery!

From Money 99 to Money 2001, the product suffered from two major debilitations that ultimately ended up in disappointed customers:

Competition with Quicken. Don't get me wrong -- I'm all for competition in software. The dynamics of competition with Quicken, however, created an environment that ultimately did not serve customers.
Money's subjection to MSN success metrics. Money is part of the MSN organization. During the heady dot-com days, when all reason was tossed to the wind, Money's success was measured against the same metrics that other MSN properties (read: “websites“) used.

We'll need to take these one at a time.

Competition with Quicken

Before all The Haters^TM spam my inbox, let me go on the record saying that competition in software is almost always good for the consumer. After all, who can forget the BBS days, back when ProCOMM and QModem were suddenly assaulted by Telemate, a multi-tasking DOS-based modem program written by -- brace yourself -- one Japanese guy. No doubt in my mind that competition breeds innovation. (Whatever happened to Telemate's author, anyway? Unbelievable what he achieved. With Telemate, you could download one game, unzip another, and play hangman all at the same time -- I kid you not.)

Brief refresher on how the PFM market works: consumers don't buy PFMs for themselves. They don't. Well, maybe three of them do. But the bulk of the PFM market resembles two other markets:

Deer Hunter. What? Not familiar with the multi-platinum PC game title that has sold gazillions of copies after 5+ versions? It's a game where you shoot deer. Seriously. Stay downwind, keep quiet, the whole works. Anyway, the point is that you don't buy Deer Hunter for ~yourself~. Heck no. You buy it for your husband. For your dad. For some business client that just signed a billion-dollar deal with your firm.
Anti-virus. Your parents' Pentium 60 still has the same anti-virus program that originally came bundled with their PC. And it's probably got Money 97 on it. Enough said.

The reality of the PFM market is that it's completely driven by journalists' reviews of the software. Don't know which PFM to buy dad? Check the PC Magazine review. Want to know which PFM to pre-install with your new line of consumer PCs? Read the Wall Street Journal review.

Rewind back to 1998. Quicken led the PFM market, and had pretty much been spanking Money with thousand dollar bills. Intuit, Quicken's parent company, called the shots. The only chance Money had to survive was to win reviews. Reviews were everything. All hands were on deck to match Quicken feature for feature... there was a veritable cover-of-the-box bullet-point shootout at the PFM Corral. Direct from the Money 2001 Deluxe box:

“Track employee stock options!”
“Shop for insurance through MSN MoneyCentral!”
“Gain insights into IPOs!”

You'd think Money users were all upper-middle class technocrats.

Fact of the matter is that we knew from years of usability studies that what users most wanted was a simple system that helped them do the most basic of things. Track a monthly budget. Get out of debt. Figure out what's left in the checking account. Try as we might to improve these things, we couldn't invest seriously in fixing these basic user experiences. What if Quicken were to come out with support for calls and puts? Indeed, that too came to pass, and reviewers latched on to the fact that the new Quicken supported call and put options. What followed was a Reagan-esque Features Arms Race^TM. It was all you could do to close your eyes, hang on, and add yet another feature.

There was simply neither interest nor leeway to improve the usability of the core features of Money. For the business to survive, we had to match Quicken move for needless move, feature for futile feature. It's little surprise that users flocked to online bank websites in preference to using Money or Quicken. What? My current bank balance, without entering any transactions manually? Single-click bill payment? Simple budget categories? Sign me up!

Money's Subjection to MSN Success Metrics

Microsoft Money is considered part of the MSN family of products, and has been for many years. It was thought that the consumer products division, of which Money was a part, should belong in MSN. As a property of MSN, Money's success or failure was judged by execs using the same metrics as MSN's websites.

You might have missed that. After all, this article's kind of long. So let me say that again. Money's success or failure was judged using the same metrics as MSN's websites.

Metrics like minutes viewed per month. Like ad revenue. Like click-through. Stickiness. I am not making this up.

I sat through meetings where we were asked to research ways in which to increase the amount of time that users spent in Money. Increase the amount of time! Users always ask for the exact opposite. Users want a Navy Seal relationship with Money -- get in quietly, do the job quickly, leave no comrade behind, maybe smoke a little afterwards. We got busy making Money into a needy girlfriend. “Let's make it so fun and engaging people won't want to leave!” Users would rather be scheduled for a root canal than to spend another minute trying to balance a checkbook.

A goal was set to increase ad revenue by 600%. At this point in time, during the heady dot-com days, Money was already filled with banner ads. It even installed several icons on the desktop hawking E*Trade and other bygones. How were we supposed to increase the revenue another sixfold? Someone suggested showing the banner ads even faster -- essentially increasing the “frame rate,” if you will. Click-through infected the product similarly. Push MSN websites! Get 'em to click through! We get referrals!

It was 1999. Twenty-year-olds were millionaires. Anything was possible. Money was judged as a portal, so a portal it became. Thing is, no one wanted a sticky, ad-serving, time-consuming personal finance application.

Silver Lining

Fear not! The story has a happy ending! After all, Money still holds a special spot in my techie heart. I wouldn't leave it hanging.

Money is no longer laser-focused on beating Quicken. We realized that the PFM pie is much bigger than just getting upper middle class folks to read reviews and buy Money for their favorite uncle. There's now much more focus on getting simple things done. Integration with MoneyCentral is going well. The web forces you to simplify. And in personal finance software, simple = good.

Funny enough, the other reason the arms race with Quicken is no longer as interesting is that both sides have lost some motivation. Starting with Money 2002, Money has “won“ more reviews than Quicken. It's often thought to be the superior product by key reviewers. Turns out Quicken's not that interested in beating Money either. Intuit's business is founded on taxes and accounting. Just look at the 10-Ks -- TurboTax and QuickBooks dwarf any revenues that Quicken brings in. It's not at all a stretch to say that Quicken's raison d'etre is to drive TurboTax sales.

And what of treating Money like a web portal? Fortunately, the bubble burst and all sins were forgiven. Money is no longer evaluated on things like minutes of use per month. Sanity has returned and all is well.