12 Things I’d Do if I Were Satya Nadella

Satya Nadella, incoming Microsoft CEO

Satya Nadella, incoming Microsoft CEO

With the appointment of Satya Nadella as CEO, Microsoft has an opportunity to turn itself around, get back on track. But fixing things will be a formidable challenge… so I’d like to help out. Here, roughly in order of priority, are a dozen suggestions: things that Nadella ought to do within his first few months in charge.

1.      Dump the Surface and WinPhone hardware lines.

Sell the name and the (minimal) goodwill if possible, but kill these boat-anchor hardware operations one way or another.  Leave the tablets to Lenovo, or ASUS. And realize that if no sensible OEM wants WinPhone, you’re fools to build it yourselves.

Emulate Google’s wisdom in ditching Motorola. Software companies don’t belong in the hardware business. You bring absolutely nothing to the party. You’re pissing away billions of dollars, pissing off your hardware OEM partners, and pissing on your billion once-happy users. Give it up.

As an ex-VW exec once told me: “Once you have to start advertising, it’s already too late.” Save your money on all those ludicrous product placements, and put it into promoting something people might actually want.

2.      Stop trying to make Windows act like a tablet.

Windows has a pretty good desktop UI. It could be improved, to be sure… but it cannot be scrapped. Could this possibly be more obvious? Windows without its ‘legacy’ UI and ‘legacy’ apps is GNU/Linux. Only not open, and not as good.

You need to immediately announce that Windows 8.2 will make ‘Metro’ an optional component, akin to Media Center, and restore the Start Menu. Then stand back and watch PCs start selling again.

While you’re at it, feel free to add some actual value – a new version of Solitaire. A half-decent mail client. Something. Give users a version of Windows 8 that actually does something Windows 7 didn’t.

3.      Make your phones and tablets act more like Windows.

‘Metro’ is a reasonably interesting mobile OS, but in order to succeed it needs the key advantages of Windows: full multitasking, true file system, the ability to run third-party firewall and security software, and the ability to plug-and-play with Windows with no drivers to install.

Without those, ‘Metro’ is just another doomed Android wannabe, with high hopes and no app library. You’ve tried to buy market share and you can see by now that won’t work. Use your colossal Windows ecosystem to add value (but not encumbrance).

4.      Announce a proper version of ‘Metro OS.’

Next, ditch the code bloat and build a standalone version of ‘Metro OS’ based on a modern kernel – either stripped-down NT, Windows CE, or a new Linux-like kernel. License this OS out to hardware OEMs at terms more attractive than Android. A tough challenge, yes – but that’s what it will take to steal market share in the merciless tablet business. If necessary, offer them cash.

Above all, stop trying to pretend that ‘Metro’ is not an OS, with its own APIs, its own UI, its own ecosystem. And give the #($@$* thing a name! Okay, so ‘Metro’ was taken. Don’t you have lawyers who can fix these things? Or English-lit graduates who know some other suitable words?

My own humble suggestion is ‘Mondrian.’ It’s an ‘artistic’ name, free of embarrassing connotations, that should work in most every language. It perfectly suits the rectangular look of the UI. Best of all, Piet Mondrian died a half-century ago, so there should be few legal entanglements.

5.      Announce a modular version of Windows.

Modular Windows has been on the to-do list long enough. Stop procrastinating and just do it. Call it something like Windows NG (for ‘Next Generation’). Sell it as a more-advanced version, parallel to Windows 9, rather than replacing it. As with Windows NT, allow users to migrate at their own pace.

Creating this new OS will take time. But that’s not a problem. What’s really important is that you announce it right away, and put your best people to work on it. You’ll be amazed how much buzz that will generate. People are waiting for this.

If you don’t have the right team leader for the project, hire one. It needs someone with a strong gut-feel for architectures,  and no stupid UI agenda. The UI is Windows. Simple.

6.      Announce a more-traditional successor to Windows 8

Windows ‘NG’ is nice to look forward to, but Windows 9 is the make-or-break version for your company. Fortunately, building it will be easy.

First: address every bug, glitch and inconsistency in Windows 8. Make networking simple – without introducing new standards like HomeGroups. Clean up and consolidate all those stupid Vista whitespace dialogs. Make Windows Explorer settings persistent. Fix it all. That will be enough to make it a best-seller.

Second: gently tweak the user interface. Make it more configurable. Add some small but nifty new tools. Streamline multi-monitor support and extend it to ‘n’ monitors. Add virtual desktops – already a standard feature of every other OS!

None of this is difficult. You could go to beta testing by fall. Want some other ideas? Just Google (or Bing, if you must) “Windows utilities.” Or browse sites like Softpedia or Download.com.

And remember: less is more. People just want a better Windows 7, not some totally new concoction. Your delirious visions don’t help me get my work done. Keep them to yourself.

7.      Evolve Office.

Again, easy as can be.

First, clean it up. Ditch the arcane DLLs and bizarre registry structures. And the copy protection – most of your revenue comes volume licensing by big organizations that are much too exposed for piracy. Make it a five-minute install. The goal should be: one application, one executable. (Or at worst, one folder.)

Next, clean up the UI. Dump the gruesome new monochromatic styling, and go back to something that’s usable, attractive and familiar. Reinstate Toolbars.

Remember, you’re competing with LibreOffice, and LibreOffice is currently way less of a pain in the butt than MS Office, in every conceivable way. Clone it if you have to. After all, they cloned Office.

Finally, look at adding some meaningful new features and capabilities. Not sure what those would be? Ask your users! Look through the endless wish-lists, the huge libraries of add-ons and hacks all over the Internet.

Personally, I know Word best, so here’s a start: make the Outliner useful, and improve structured-document creation. (Have a look at NetManage Ecco and Lotus Manuscript for ideas.) Allow multiple documents to be combined into workspaces, as in Excel. Rebuild the idiotic layout system, and make images actually go where you want them to. Support multiple synchronized columns, so that scriptwriters can finally use your software. Embrace open file formats, like HTML and ODF.

I also have a suggestion regarding PowerPoint: scrap it. It’s a dead loss and everybody hates it. Start over with a clean slate. Die-hard users will thank you.

8.      Open up your APIs, file formats and other standards.

Make Windows ‘proprietary but open,’ like it was in the 1990s. In the long run, it’s your only hope of competing with GNU/Linux.

Open-source your developer tools. Your version of Visual Studio should always be the best version. But as a company, you’ll do better in the long run if you let users ‘mod’ these tools, and extend their usefulness onto other platforms than Windows. Get behind the Mono initiative, help.NET become a universal standard.

It’s too late to pretend you’re a universe unto yourselves. Android is a reality. iOS isn’t going away. GNU/Linux will continue to make gains. Stop obsessing about being the only option; that’s never going to happen again. Concentrate on being the best option. That’s a goal you can actually attain.

9.      Sell or split off the Xbox division.

Software companies don’t belong in hardware. (Where have I heard that before?) Sell Xbox, or spin it off as an independent venture. It’s not making any real profit, and it’s not helping with any grand strategy of taking over the living room. (Right now, only Android and Linux have any chance of that. The TV content makers and distribution companies will never accept your proprietary ‘solution.’)

Once it’s separated, continue to act as the system software supplier for Xbox Inc. Work on a major revision of the software that offers more continuity with Windows. As soon as possible, offer this new Xbox OS to other hardware OEMs, like Valve is doing with SteamOS. If you had two companies competing to build the best Xbox, you might have a chance of a decisive victory over Sony.

10.  Rebuild the Windows games division.

This probably seems like a less-vital task, but the Microsoft games division has a lot of sentimental value. It was once a leader in its field. It should be again.

It also has a lot of franchises that deserve to continue. Age of Empires. (I dream of a baseball-themed RTS spin-off called Age of Umpires…) Links. Close Combat. Crimson Skies. Dungeon Siege. And above all, Flight Simulator! It’s an industry institution, and desperately needs to be revived. I’m sure you can get back the source code, and most of the developers, from Lockheed Martin’s Prepar3D.

The new games division will be profitable, but it’s even more valuable as a way of establishing Microsoft’s credibility. The leading software company should also be developer of some memorable game software. On the other hand, gaming shouldn’t be a dominant activity for you. Keep it small, isolated… a ‘boutique’ studio. That way EA and Ubisoft won’t feel miffed.

More importantly, the games division should be a vital resource, that informs all of Microsoft’s platform developments. It should help in the revitalization of DirectX, making proprietary developments like AMD’s Mantle unnecessary. And it should help renew the commitment to Windows – still the most successful gaming platform by far, which you’ve been failing to cash in on.

11.  Restore trust.

Open-source what you can, so users can assure themselves they’re not being spied on. Move your cloud services offshore, if you have to, to some regime people may be willing to trust. (If you can find one.) Make Bing the search engine that doesn’t abuse privacy. People will let you sell them stuff if they trust you not to take unfair advantage.

More importantly, flex your muscles. The US economy needs you more than you need it. Tell the government that you won’t participate in unconstitutional invasions of privacy. Go to court if you have to. Become the champions of total privacy. (What’s the worst that can happen? Nobody’s going to send the Marines to Redmond.)

This will cost you something in legal fees, and something more in the restraint to not engage in privacy shenanigans of your own. But it will build universal support from potential customers, including the global enterprises and foreign governments who should be your most lucrative market.

This is possibly the toughest challenge you face. But if you don’t do it, you will lose the battle with GNU/Linux and other open-source developments. Because they can promise total security, and you can’t, ever again, unless you change the entire political system.

12.  Above all, start listening to your users.

Start listening, and do it conspicuously. Create user-suggestions forums for Windows, Office, Metro and so on. Make a point of adopting those suggestions as often as possible. Forget ‘the cloud,’ or ‘mobile’… ‘crowd-sourcing’ is the buzzword du jour. Let your user base do the work!

Put a team to work scouring your user forums, old and new, for unanswered questions. Resolve them all! Commit totally to the ideal of never, never, never leaving a user unsatisfied. If it’s a reasonable question, either find an answer, or rebuild the software to create an answer. And then go back and tell them you’ve done it.

And That’s It!

Satya, once you’ve worked through that list, get back to me. We’ll do lunch, and I’ll quote you an hourly rate for my next 12 suggestions.


(My companion piece, “Perspective: Microsoft Then and Now,” has been posted to WifiHifi.ca. It’s a look back at the things Microsoft has done right, and wrong, over its near-40-year history.)

About fung0

Long-time tech journalist. Contributor to numerous periodicals, online sites and TV productions in the US and Canada. Author of several books on computing topics. Graduate mechanical engineer (Queen's U).

Leave a Reply

Your email address will not be published. Required fields are marked *