2014/03/18

Microsoft, IBM and the Price of Success

Microsoft's business is now clearly "under performing" with even the Board admitting it. Steve Ballmer has finally been pushed out. The MSFT Board could've followed IBM in 1993 and appointed an outsider, but they chose an long-term insider.

The imminent "death" of Windows-XP (8-Apr-2014, free security updates end) shows, in my mind, a clear misattribution effect by Microsoft: they've confused cause and effect, that people bought their brand, not just a "standard" commodity product.

Windows XP, with 29% of PC O/S market share vs 47% for Windows 7, is still a large product segment.
By orphaning it's second largest user base, Microsoft is making two very strong statements:

  • they aren't interested in building incrementally on success, they could easily get $5/year from Win-XP home users for updates (paid for 3-4 years in advance) OR create "son of XP" as a cheap, lightweight supported option for XP owners. They could make $25-$50 from those millions of current customers, with almost no investment by MSFT.
  • they believe people's primary purchase decision factor was "Microsoft" not a standard commodity platform.
How can successful companies confuse the causes of their success? These are smart, informed and fiercely rational guys (mostly still) aren't they? With a whole lot riding on their decisions: their careers and future income/reputation. Skin in the game, in spades. All pointers are they should be getting this right.

But from IBM, we know that success blinds marketers, managers, CEO's and boards. In 1981, 20 years after the launch of the IBM 360 mainframe, IBM had 60% by revenue of the entire market: they were 50% bigger than all their competition combined. "Successful" doesn't come close to describing them.

The IBM PC released that year became the standard PC, not just in offices, but everywhere. One small factor annoyed IBM, "clone" makers had copied their design and were stealing sales.

IBM, in 1987, responded with a known, effective strategy: they changed the product architecture, expecting to force those many clone makers to pay license fees.  The new PS/2 came with a new video card (VGA), new mouse & keyboard connectors, new backplane (MCA - Microchannel Architecture), new proprietary networking (Token Ring) and a new, much better Operating System (OS/2). Even new and better connectors to disk drives were added: a single connector for both signals and power.

The plan was new hardware & software together, but OS/2 was late and they shipped PS/2's with MS-DOS, undermining the strategy and costing sales along with market share.

IBM, like Hewlett-Packard, design and build great hardware. The PS/2 would've been a delight to use and great economics for large enterprises needing reliable PC "fleets". But great hardware takes time to design, build and test. Smaller, more agile hardware vendors responded in months, not the 3 or more years it had taken IBM to roll out the PS/2.

Only, customers didn't follow the IBM Marketing's script. Vendors kept building "ISA" (Industry Standard Architecture) PC's, not MCA, but upgraded them to use the PS/2 keyboard/mouse connector & interface and VGA. Ethernet won the networking wars and it was another 15 years before we saw laptops with single connectors for disks.

By 1993, IBM was haemorrhaging money: $16 billion in losses over 3 years, half that in 1993 alone. Things had gone bad for them, very quickly. Louis V. Gerstner was brought in from outside IBM and didn't just stop the decline, but put them back into profitability. He stayed 10 years, leaving in 2002. 

How could a company of 350,000 of the best computing and marketing people on the planet, led by seasoned professional managers with impressive records, get it so wrong?

My thesis is that successful companies must get disruptive decisions wrong - our minds are wired that way...

It isn't just confirmation bias, group-think, self-deception or perceptual bias/filtering, it's a basic misattribution effect: at some point successful enterprises (not just business) confuse Cause and Effect.

The very act of being successful in the medium-term renders the leaders and administration of an enterprise, be it business, military, political, religious, or academic, to believe their success is more than luck or serendipity. We are now seeing aircraft pilots crash because they believe their mental model "the automation/instruments are always correct" and discount the evidence of their senses.

Our brains are wired to make sense of the world and look for causality. We form very strong, powerful mental models that generally work for us and, when correct, create successful outcomes for us. This cognitive ability has allowed us to survive almost anywhere, alter the large-scale environment to suit our needs and create everything around you.

But it comes at a cost. Sometimes we confuse Cause and Effect. We can see it in other times & cultures and their predisposition to believe in magic or superstitions. For those people and cultures, this was a rational, evidence-based belief-system. Everything they observed confirmed their mental models of how the world works.

So to IBM and Microsoft and thousands of successful enterprises before them that stumble and fall. It took the Roman Empire a thousand years to implode... But fall it did.

You only get to find out your mental models are incorrect, which are very refined and have served you long and well, after a resounding collapse. It's not obvious to anyone when they start failing because "that's just a natural variation" or "we're facing challenges, but we've overcome obstacles before".

It took IBM around 5 years from the release of the ill-fated PS/2 to fail more massively than anybody could anticipate. How long will it take Microsoft to hit the ropes after abandoning 30% of their user base?

Ordinary people bought Windows-XP because "it was there" and "its the standard".
Hardware vendors provided Windows-XP because there was consumer demand and no affordable alternatives.
ISV's, Independent Software Vendors, wrote software for Windows-XP because it was the dominant Operating System. Serving that market alone was a correct strategic decision.

Being successful in the market created more success. Higher PC sales were caused by other factors, like falling hardware prices, aggressive competition amongst vendors and the Internet revolution. All of a sudden, PC's were a lot more valuable networked than standalone. Microsoft benefitted by being in the right place at the right time with a good enough product. They beat out all the early competition because they took some massive bets that paid off. We might all be running one of a dozen different O/S's if their bets had failed.

Microsoft, and Ballmer in particular, never recognised this. Ballmer and the Board, in my view, believed something they did led directly to their outrageous success. The very obvious contradictory evidence that they'd been unable to replicate their success, despite tens of billions spent trying, was simply ignored. The massive wake-up call that their engineering processes were deeply flawed, the 2004/5 "Longhorn Reset" - a 30,000 man-year "FAIL", was also ignored.

Humans are amazingly good at ignoring disconfirming evidence: IBM and Microsoft executives aren't unique or special.

We have absolute proof that a sizeable fraction of the population would rather die than change:
smokers continue to endanger their lives in spite of every warning while healthy, and an alarming proportion continue smoking even when diagnosed early enough to do something.
Ballmer and his board, like IBM's many CEO's and board, aren't evil, mad or bad in missing the obvious. They're just human with the same frailties as the rest of us and countless leaders before them.

The trick would seem to be having someone powerful around that will tell you "It ain't so".
My guess is that Bill Gates saw the writing on the wall and after his view was rejected, stood aside. But we'll never know because he'll never say.

No comments: