"It's a bold step into the future - a win-win for employees, shareholders, and consumers of both companies. Bringing these great teams together will accelerate Microsoft's share and profits in phones, and strengthen the overall opportunities for both Microsoft and our partners across our entire family of devices and services," said Steve Ballmer, Microsoft's then-CEO, as he announced his plan to buy Nokia's phones business in September 2013.
"Given our long partnership with Nokia and the many key Nokia leaders that are joining Microsoft, we anticipate a smooth transition and great execution," Ballmer went on. Microsoft paid €5.44bn for Nokia's devices and services business, and as part of the deal 32,000 Nokia staff joined Microsoft. The idea: the combination of Nokia's design brilliance with Microsoft's software engineering could create a smartphone business to take on the might of Apple and Google.
And while Microsoft CEO Satya Nadella said this week he is "committed to our first-party devices including phones," it does not sound like a ringing endorsement. According to Nadella, Microsoft's strategy will focus on three areas: "management, security, and productivity" for business customers, "communications services" for value phone buyers, and "flagship devices" for Windows fans.
It's worth pointing out that two of those three elements don't necessarily involve Microsoft actually building Windows Phone devices anymore. Nadella's remarks suggest that Microsoft may reduce its involvement in phones even further in future.
"In the near term, we will run a more effective phone portfolio... In the longer term, Microsoft devices will spark innovation, create new categories and generate opportunity for the Windows ecosystem more broadly. Our reinvention will be centered on creating mobility of experiences across the entire device family including phones," he said.
In an email to staff on Wednesday, Microsoft COO Kevin Turner dropped some hints about the future of Microsoft's phone business. He said the company would have to make phones faster, but also said it would focus on the channels and markets that offer the best returns. "This is a similar approach to the one we have taken with Surface, which has been very successful," he said - but while the Surface hybrid PC has been succesful, at least if you're feeling particularly generous, it is a much more modest business than the smartphone empire Microsoft once planned to build.
So what went wrong? Part of the problem is that the tie-up between Microsoft and Nokia was based on weakness, not strength. Microsoft had been trying to persuade manufacturers to use its Windows Phone operating system on their phones, but with little success. Similarly, Nokia had found the competition from Android and iOS too strong - Nokia CEO Elop memorably described Nokia as standing on a "burning platform" by adhering to its Symbian OS.
Putting together Nokia's smartphone hardware and Microsoft's software and marketing power seemed to be a reasonable bet, but it was by no means a certain one.
But the phones' hardware and software was not the problem: Nokia's long history of creating elegant hardware meant that the Lumias won plaudits for smart design. And, after a few false starts, Windows Phone with its tiled user interface seemed far cooler than iOS. Nonetheless, together, they just didn't sell enough phones.
In its best-ever quarter Microsoft managed to sell 10.5 million Lumias (likely weighted towards the cheaper models). To put those numbers into context, Apple sold 61 million iPhones in its most recent quarter. Today, IDC puts Windows Phone market share at 2.7 percent worldwide. For Microsoft, that's just nowhere near good enough.
Partly that was because Android and iOS were well entrenched by the time Windows Phone had matured, but also because Windows Phone has never built up the same app ecosystem the other two operating systems did. This meant that some top apps would not arrive on Windows Phone for a long time, if they ever arrived at all.
Microsoft's mobile strategy has at times been hard to fathom too. It has favoured a scattergun approach, launching an array of different smartphones and then waiting to see which ones find favour - in contrast to the measured approach of Apple.
Lumias found most of their success at the entry level. Indeed, there hasn't really been a 'flagship' Lumia (a clear top of the range device) released for a while. In the meantime, Microsoft has been experimenting with putting its services onto iOS and Android devices instead (Nokia even testing out an Android device) which may have given some customers one less reason to buy its hardware.
So what of the future of Microsoft's mobile strategy?
Microsoft is, it appears, aiming to be device-agnostic here: it doesn't matter what hardware you own so long as it has Outlook, Office, Skype, Bing, or other Microsoft apps running on it. It's not a terrible strategy but at the moment consumers still prefer those services on handsets that aren't made by Microsoft.
The benefit of making your own hardware is that you can tightly embed your own software and services, making them the default options. When you are piggybacking on someone else's operating system, those close ties are much harder to achieve. If you own an Android, you'll end up using Gmail and other default apps, like Maps. On an iPhone you'll browse with Safari. It's just easier that way - which is why Microsoft bought up Nokia's smartphone business in the first place.
And mobile matters for the biggest tech companies. Even if you can't make money from mobile hardware (and not many companies can apart from Apple), that hardware matters because smartphones are - much more than PCs now - the gateway to our digital lives.
How? Well, smartphones control wearable tech like smartwatches. So to be a player in the wearables market it helps if you already make smartphones. And if you make wearables, it's that much easier to make a step into the consumer Internet of Things. All of these things are connected. Without hardware, you run the risk of falling out of the race altogether.
Surface can be deemed a success because, through it, Microsoft showed its hardware partners how to build a Windows tablet PC hybrid device that worked - and thus prodded them into making their own. The same strategy isn't sure to work with a smartphone, because the hardware and the software on it aren't the problem - it's the broader ecosystem. Building the world's best Windows Phone device for hardware companies to copy isn't going to help with that. Microsoft's adventure with Nokia hasn't succeeded, but that doesn't mean it can afford to give up on mobile.