I would also like to address a common response to the 'Google is doooomed' thesis, which is "Hey, look their revenues are going up and up, you would be stupid to think they were doomed!"
I've asserted for several years here that Google's core advertising business is rotting from the core outward. And I used as the basis for that tracking their reported "cost per click" erosion and their "paid traffic" growth.
The reasoning is pretty simple; At the most basic level Google puts an ad out there, and when that ad is 'clicked' [1] they collect a fee. And if you are familiar with the history, this was revolutionary where advertisers could pay only when someone demonstrably both saw their advertisement and indicated an interest in it with a click. And initially it was wonderful and then bad guys figured out they could insert themselves in the middle, selling advertising space to customers at a price per click and then buying the ads on google can clicking on them with their own computers and collect some ad revenue off the top. Easy riches except that the ads massively underperformed because most of the clicks on them were fraudulent. And that begat a war that continues to this day and has lead to a general disillusionment with online advertising.
So people get disillusioned, but as advertisers are setting the prices and budgets they show this by being unwilling to pay more for ads that they aren't sure will be effective. And since Google's system is auction based their pulling back results in the average price Google can get for a 'click' to go down. At which point Google has three choices, eat a revenue loss (not really a choice), give them more opportunities to buy ads, or increase the number of people who are getting a chance to click.
Google started by bringing more people in contact with their ads by "buying" traffic. The way that works is you find a web property where people are accessing the Internet and you pay the proprietor money to only show your ads to their traffic, or to move their traffic to your web site. So in the case of Firefox they paid mozilla to send their search traffic to Google, in the case of Apple they paid to have all of iOS search traffic go to Google by default. In all these cases you "lose" a bit by increasing your traffic acquisition costs (that money comes out of ad profits) but if you get more clicks from that traffic than you paid for it, you can cause your revenue number to increase. As the effectiveness of buying traffic has declined (and there just aren't that many sources out there and Microsoft is also paying for traffic so you end up paying more and more for less and less valuable traffic) and the cost has gone up, they moved on to stage two.
Stage two was to start increasing the number of ads you had exposure to when you landed on Google sites. You look at how hard people work to be in the first few search results on the page, well advertisers want to be in the first few ads. You can also create other ways to pay to play like, for example, charging advertisers a monthly fee in order to enable them to appear in a 'shopping box' result that appears on what looks like a shopping search query assuming they bid enough on their ad to show up there. Search engine result pages (SERPs) for shopping 'like' queries progressively got fewer and fewer organic results and turned into mostly ads. Google stopped surfacing organic web pages or web sites from their index that their algorithm told them were the best results for a query in favor of people who had paid to be on the page in the event a given query or keyword was entered. More ads to click, heck pretty much anything you clicked on the page was an ad, and revenue goes up.
Of course there are limits to how far you can take this, and the article pointed out that people are pretty fatigued from all of the advertising. As you block more ads and start avoiding things that are just ads in favor of perceptually "better" web sites, revenue should drop at Google. And I was surprised it didn't, but then I read this article : http://www.cnbc.com/2017/04/27/alphabets-google-unit-grabbin... The CNBC folks had noticed that Google's ad revenue from its own sites was growing faster than its partner sites. And that tells me that Google had a third strategy which was "reduce the share of ad revenues to partners." Which really sucks for partner sites and adds commercial sites to the list of people complaining that their AdSense revenue reduced to a trickle over the last 5 years.
Google burns though an extraordinary amount of cash. Its partly due to the nature of their business and its partly due to the way they are investing looking for new opportunities. The unanswerable question is what happens when Google runs out of ways to offset their ad profit margin declines and have to show a reduction in revenue Q/Q and then Y/Y? I don't think they will 'Collapse' any more than IBM collapsed when the computer market stopped carrying the load for them. But will be an extraordinary test on their leadership to avoid setting up a Blackberry like slow descent into near obscurity.
[1] Yes there are CPI models as well but historically they are both lower paying and a smaller fraction of revenue.