[avidiax]

This speech is anything but clear from an engineering or analytical perspective.

"Outsized success" is always in relation to the value of the allocated resources. "Conventional success" is making slightly more than liquidating those resources and following a passive investment strategy, like index funds yielding 8% APR.

So what "Satya" is saying here is "I want higher risk, higher reward strategies", but "don't be unlucky" (don't "make a habit of" failure).

I'm no exec, but I'm my opinion, this is a poor strategy in the consumer and especially enterprise software and services space.

"Satya" is wrong: there are 3 knobs. The missing knob is risk allocation. This works in two ways: some areas are intrinsically risky, and some areas you choose to take risk in. Either way, you allocate that risk with strategy and resource allocation (the other two knobs). If an area can't fail, you may spend more resources to de-risk or hedge, or strategically avoid change or the area. If an area has outsized opportunity, the strategy and allocation takes more risk there.

"Satya" is telling the whole room that their area should favor high-risk, high reward strategy and allocation, when he should be telling them that they must determine if their area should favor or avoid risk to maximize the entire company's return on investment.

The real estate team should not be a risk center. It should not fire 90% of the janitors in favor of "AI first" robot vacuums. It should not decide that real estate as a service (WeWork) is a good strategy over buying real estate in advance of need. It should avoid capacity crunches while putting holdings in LLCs to reduce downside risk. It should vertically integrate the data centers (land, connectivity, power, water, regulation/politics/PR).

Lots of Microsoft failures look like risk misallocation to me. Windows phone was allocated risk too late in moving away from Windows CE. Windows Vista took too much risk and failed to release until features were rolled back. Windows 11 (financialization of the OS) seems like they are betting the farm to secure some ad and subscription revenue, and keep their low margin PC maker partners happy, while they make the platform the least attractive it's ever been.

[rawgabbit]

In my mind, Satya is demanding outsized returns from the resources he’s allocated to them. He is delegating strategy to them. What we get and what Microsoft produces is a product of a dysfunctional organization without any coherent strategy. Slapping Copilot on existing products is not a strategy.

[cmp0]

You make a good point I think the blog post still works.

"The real estate team should not be a risk center." - Why not? Sure, I agree with you it doesn't make a ton of sense _to me_ to fire 90% of the janitors.

Satya said he wants his team to be "Intellectually honest".

I'd argue that a plan to fire 90% of the janitors today is not that realistic, but if some real estate exec thinks they honestly have a plan that could be successful, are tracking data, and change their plan if they're wrong - what's the problem?

Maybe they're crazy enough to have figured something out we don't know....or not in which case Satya says you'll be gone if you make it a habit.

[avidiax]

Because real estate innovation is not something that the customers care about, and the real estate team's outputs effect everyone else's outputs.

How does the rest of Microsoft perform if there's not enough office space or DC capacity is far below needs?

Risk is much more tolerable in the last mile than the first one, much more tolerable in the leaves of a dependency graph than the root.

[cmp0]

> Risk is much more tolerable in the last mile than the first one

Is that objectively true? I think you could also make the argument the other way that things compound when you innovate at the root.

I think it's your opinion that that's the case, and it generally makes sense to me as well, but I'm still saying it's a judgement call.

There are a lot of things / decisions at a company that don't strictly matter for the customer immediately but they still eventually matter for the customer - because they have a downstream impact. In this case, maybe you save that money and spend it somewhere else.