* Enterprises have a continuous ongoing need for new technology. Some they build in-house but like anyone else they are mostly busy running their current business and so they tend to buy important new technology from technology companies. I bring this up because it was in doubt a while ago -- and enterprises really slowed down buying new technology between ~2000 and ~2008 -- but I think it's a constant truth.
* Just like consumers, in enterprises there are early adopters, mainstream adopters, and laggards. The good news is that there are almost always early adopters for any interesting new idea -- we always tell enterprise startups to look for the first 5 customers -- if you can make them happy, then you can reference sell to the next 10, then the next 50, then the next 100, then the next 500, and then you are huge.
* Consumerization of the enterprise is real. The bar on usability and ease of adoption is being set by the consumer product industry; employees are increasingly bringing their own devices and services with them to work; there are more and more enterprise products that can be bought and adopted bottoms-up.
* However, it is hard to get a lot of MONEY from enterprises through only bottoms-up adoption. To get the money, at some point you have to strap on your big boy shoes and go in top-down and explain to senior executives why your technology is going to give them real competitive advantage or save them a lot of money. (This is the "enterprises don't have credit cards" principle -- you have to go get the money.)
* There are real constraints around what technology enterprises can adopt and how they can adopt it. A short and incomplete list: internal bureacracy, sunk costs, existing vendor relationships, regulations, data protection laws (e.g. in Europe), security requirements, need for integration with existing systems. At first these seem very frustrating to deal with. Later you realize that these can become a source of competitive advantage for you once you are in. Figuring out how to navigate that is really key and something the best enterprise entrepreneurs are very good at.
This doesn't get into specific insights/problems that they have at any given moment but this is a good general framework to begin an enterprise discussion.
The CIOs are now much more open to new products than before even in very traditional and conservative industries, while the leading existing technology providers are not in a much better position to offer a new product, as they would be starting from scratch because of huge technological advances and a shift in development practices, all thanks to consumer startups.
Given a backing from an A-list player provides the needed boost in initial traction for enterprise startups, basically a "rolodex" of previous acquisitions and a network of CIOs, if an idea has a market, the failure rate can be kept at a minimum.
The only thing I don't understand is how an exit strategy looks like, as it seems only logical that there should be a monster or two on the enterprise market (Microsoft and Oracle of the 21 century), who should either acquire those products that succeeded or develop their own alternatives. So I'm pretty sure that in 5 years time the great consolidation is coming and it will become more difficult to develop a business out of a product idea.
On your final point, I think that's certainly possible. At least that's the historical cycle. There are actually a bunch of plausible acquirers at scale -- of the new companies Salesforce and Workday are certainly candidates, and a bunch of the older companies as well -- it will be interesting to see if a new CEO of Microsoft embarks on a major acquisition binge of enterprise cloud companies (perhaps in conjunction with spinning off or killing some of their less-successful consumer efforts).
However, it's also possible this cycle plays out differently, or at least at much larger scale and over much longer. The three big arguments in favor of this are:
* New enterprise cloud/SAAS vendors may be able to sell into much bigger markets than traditional enterprise software -- since there are many more companies that can use online services than could ever install and use enterprise software on premise. In particular, going downmarket is far easier with the new model. Plus, globalization and the developing world could dramatically increase market size. So new entrants may be able to get much larger as independent companies than the last generation.
* New enterprise/SAAS vendors may exist in many more categories than traditional enterprise software. New development/adoption/business models make it reasonable to think about a lot more horizontal and vertical applications and services than old enterprise software was capable of addressing. So maybe new vendors get much larger than anticipated since they can cover a lot more functions and verticals, or alternately maybe there are many more new vendors than we can conceive of now. We are already seeing a trend towards enterprise use of a lot more cloud apps per customer at this point in the cycle than a lot of people expected.
* As a consequence of the prior two, the independent market caps of the new cloud/SAAS vendors may end up being a lot higher than you might anticipate from history. This could really remove the incentive to sell hot new cloud/SAAS vendors as quickly.
I think these are live topics right now, and the answers will determine a lot of what happens in the enterprise startup ecosystem for years to come.