Cash cow businesses have a cash flow that they either return to investors or reinvest themselves. The only metric for success reinvesting the cash flow is whether the company's stock grows at a higher rate than investors could obtain for themselves if they invested the dividends in the market.
Microsoft has consistently failed this test. It doesn't matter whether their investments are purported to be defensive to protect their main businesses or offensive to develop new sources of cash flow.
Defending Windows from competition, for example, is only useful to the extent that it helps the company's stock grow. If the stock does not grow, it would be better to preside over a gradual decline while throwing off as much cash as possible so that investors could invest the remaining dwindling cash flow in better companies.
Google is an entirely different animal because their stock is much more attractive to investors. As long as their stock continues to grow, management are able to get away with much more "wasteful" attempts to get lightning to strike again. But the moment its stock plateaus, they will be subject to the same merciless metric from me, namely can they establish that they can manage the company's stock price such that reinvesting cash is superior to giving it to shareholders to invest for themselves.