[ChuckMcM]

This is exactly correct. If you have a printing press that is dumping a couple of billon dollars into the basement every quarter in spite of what you're spending, you have the luxury of trying all sorts of novel things from a business perspective.

That said, the margins on the search advertising business are shrinking, and have been for years now. You can cover for that by eliminating profligate spending and then eventually you can't. I think Ruth is the right person to handle that transition but its going to hurt in terms of people feeling that "the magic is gone."

When that happens, Google is going to have to start rewarding operational efficiency gains at least as strongly as they currently reward popular science project type gains. And then it will be very interesting to see how they change and if the era of "but Google offers its employees ..." will be over.

I hope the free sodas stay though.

[rsync]

""the magic is gone.""

Not only is the magic gone, but the boring, day to day fabric of functionality is gone.

Have you tried to really search for anything with google in the past five years ?

I don't mean meandering, aimless, consumer-centric searches like "new fitness watch" where any result tangentally related is useful ... I mean a specific, meaningful, concise search with multiple keywords, all of which are key words ... and the resulting pages do not contain one or more of those words.

I know all about allinsite: and quotation marks and so on ... those are, as far as I can tell, randomly interpreted by google. Who knows if they even work, ever.

A front page item on HN a few days ago asked what people would pay $1000 per month for ... I would pay that to have search with full control over search results with logical operators that actually worked.

[guycook]

I know exactly what you mean. "Half" "of" "my" "searches" "look" "like" "this" "these" "days". I can understand throwing out one of many keywords when I'm down on page 6, but when I search with only 4 terms and the very first result has 2 words s̶t̶r̶u̶c̶k̶ o̶u̶t̶ I start to wonder who exactly this search box is for.

[mojuba]

True, and one other thing I've noticed is e.g. "Half of my searches" can bring say 200,000 results, but "Half of my" yields 50,000. A few other weird things that's been there for years and nobody seems to care. As long as consumer-level searches work and ads are clickable, Google's clock is ticking.

[gniv]

I wouldn't put too much stock on those numbers. They are probably statistical estimates, not actual counts.

[lobster_johnson]

Under "Search tools", click on "Verbatim". It will cause the query to actually honor the words you typed.

[swang]

does this honor typos? i like spellcorrect changes but i hate all the other assumptions they make about my search terms.

[lobster_johnson]

It will show a "Did you mean:", but doesn't automatically rewrite it. Example:

https://www.google.com/search?q=gam+of+throes&tbs=li:1

[niels_olson]

I just searched for "Calcifying Fibrous Tumor", and got exactly what I wanted, the most recent review article on that tumor. What are you looking for?

[ChuckMcM]

"best free credit card" :-)

[paulcole]

> meandering, aimless, consumer-centric searches like "new fitness watch" where any result tangentally related is useful

I would wager that this describes a shockingly high percentage of the average person's searches. I know it's true for me.

[l33tbro]

This was a decent thread lamenting Google of yore.

https://news.ycombinator.com/item?id=9565464

[patrickk]

> A front page item on HN a few days ago asked what people would pay $1000 per month for

Somewhat meta, but I cannot find this item on https://hn.algolia.com or using a Google search ("site: news.ycombinator.com $1000 for") can you post a link if you have it?

[mkl]

I found it with Algolia using [ask hn pay]: https://news.ycombinator.com/item?id=13114954

[patrickk]

Thanks! I noticed it's "$1k/mo" so I suppose it's difficult to chalk this up to the google algorithm ;)

[AznHisoka]

i share the same experience. i have been using google to find articles that share a specific opinion so i do very specific searches in hopes of finding editorial opinion pieces. what i get back are lame how-to guides and wikipedia articles.

[knight17]

Earlier they had a blog search option which I used a lot to find pages written in a personalised style, but they shelved this. Now there is no way to find such pages under the current system other than trying yor luck with searching for site:blogspot.com and site:wordpress.com; but these two blog hosts are not even half the universe of all the blogs out there.

What I do now is to create a Google Custom Search Engine with all my RSS feed subscriptions; it resembles this earlier feature as I feed it lots of personal blogs that I read, but it is not even close to the earlier blogsearch.

Similarly there was a discussion/forum search which was also useful; that too was removed.

[Arjuna]

"I hope the free sodas stay though."

For those that are not aware of the reference here, you will enjoy reading Steve Blank's post entitled "The Elves Leave Middle Earth - Sodas Are No Longer Free."

https://steveblank.com/2009/12/21/the-elves-leave-middle-ear...

[technobabble]

Wow, thank you for the article, I've never read it before.

This happened at the Bike Cooperative I used to work at; everything was informal, but we were able to focus on weird types of bikes and recycling projects that were interesting to us and subsequently we'd dedicate hours to do them. Unsurprisingly, the management came in, and we were wrench monkeys like everybody else; not so say that we were whiny because we didn't feel like "special snowflakes. We lost freedom to focus on our strengths in the cooperative.

[nunez]

Awesome article.

[RandomOpinion]

>When that happens, Google is going to have to start rewarding operational efficiency gains at least as strongly as they currently reward popular science project type gains. And then it will be very interesting to see how they change and if the era of "but Google offers its employees ..." will be over.

One has only to look at Microsoft to see an example of what might happen. Things were very carefree there as well back in the '90s during their heyday when growth seemed limitless.

It's also interesting to contrast Google / MS with Amazon, where Bezos has kept a tight rein on operational efficiency all along. They probably won't suffer as much when they reach their growth ceiling.

[ChuckMcM]

Agreed, and studying companies that have come through the transition to the other side is very useful. Particularly telling is how people describe what they "like" about work.

[sdenton4]

I dunno, man; I already hear a lot about suffering at Amazon...

[_audakel]

http://www.geekwire.com/2016/report-amazon-employee-jumps-of...

[nir]

Do people in Google still think the magic is there? It seems like a great place if you want stability and a nice salary, but obviously (unless you're in the comparatively tiny X dept) you're going to be a very small cog in a huge machine?

[ihoz]

I know many people at Google Boston (Cambridge). They are all bored out of their skulls.

One guy says how he works 30 hours in a busy week and uses his 20% time to hit the gym.

But in another's words: "how can I leave when they pay me so much?"

[tytso]

I work at Google Cambridge, and I'm easily working 50+ hours a week. I work in Technical Infrastructure (on our storage systems in our data centers) and one of my 20% projects was to design and architect the underpinnings of what an Android Product Manager has claimed to be one the most important features of Android N (file based encryption), the open source incarnation of which (ext4 encryption) has turned into a generic file system encryption mechanism that has enabled encryption for f2fs and soon ubifs, and has inspired the btrfs and xfs folks to look at file system based encryption.

For my "official" 80% project, in order to get approval to go into the planning phase for a complex, multi-year project, we had to demonstrate how it would save hundreds of millions of dollars on an ongoing basis, and show how the TCO savings would be many multiples over the software engineering costs to implement said project. So in Technical Infrastructure, we do very much care about operational efficiencies --- that's how Google Compute Engine (which is driven by TI) has been able to cut its prices and inspire the market to follow suit.

And this is not new, by the way --- the whole time (coming up on 7 years) that I've been at Google, we've always been very interested in bringing costs down, and the team was amply rewarded when we rolled out ext4 across the production fleet, because we could accurately quantify the performance benefits, and how that translated to dollars saved to the company.

And so no, I'm not bored. In fact, one of the great things about Google is that I have flexibility to do things that help move the industry, both in Google products and in the Open Source world. This includes serving on program committees, working with a graduate student on a paper which we've submitted to the Usenix FAST conference (fingers crossed, we'll hear soon if it's been accepted), do presentations of my work at LinuxCon, etc. etc. Of course, I probably should be carving out some time to hit the gym, just for personal health reasons....

[ChuckMcM]

I just love the dissonance of "I'm easily working 50+ hours a week." and "one of my 20% projects".

When you say it "You'll be working 40 hrs a week and to be successful you'll need to spend an additional 25% of your time coming up with other stuff that helps us." it doesn't sound quite so magical :-)

That said, when I worked there I quickly realized that with no trouble at all I could spend all my waking hours in the office playing with the neat things that are made available to employees. It doesn't feel like overtime when its something you would do in your free time anyway.

The sticky bit is that you don't get profit participation in the financial upside of those extra projects.

[tytso]

Sure, but a lot of these things are things I was doing anyway when I was working for MIT and earning a five figure salary. When I moved out to a startup (VA Linux Systems), I more than doubled my salary; and when I went to Google I took a significant step up in compensation over my previous job at IBM --- and yes, I'm still doing the things that I love to do.

So it's not a matter of doing things because it's needed to be successful (although they won't hurt when I'm next up for Promo, I'm sure), it's the fact that I get to help influence the industry. As far as profit participation is concerned, I'm paid plenty enough to be comfortable, and I'm used to the concept of the financial upside coming indirectly from what my open source activities (e.g., ext4 maintainer, being on the program committee of ATC and FAST conferences, etc.) do for my personal brand. Yes, I'm not going to have the money of say, a Peter Thiel or a Larry Ellison. I also don't have to worry about having all of that money turning me into an *sshole, either. :-)

Am I aware that I'm trading off potentially being able to earn even more income in exchange for more security and the freedom to choose what I do (and the fact that I can find choices that help out the open source projects that I care about _and_ which result in benefits to my company is something I consider a feature, and part of the value I can bring to my employer)? Of course. Sure, I could probably earn more by working for a high frequency trading firm. But I'd lose a lot of freedom. More importantly, what working on moving the poles of the US financial system to the right-half plane would do to my soul isn't worth the extra money it might bring in.

[hobo_mark]

However, for those who are not aware: https://en.wikipedia.org/wiki/Theodore_Ts'o

In other words, very much an outlier, even at Google! :)

[tytso]

Last year, when I was attending our Storage Infrastructure roadmap/planning meeting, it was very humbling because there lots of people in the room that were way smarter than me. Which is the other really great thing about working at Google; unless you're someone like Jeff Dean, there will be lots of people smarter than you that you can learn from, and the default assumption that you can make is that your colleagues will be very talented.... so I'm not sure how much of an outlier I really am!

[vidarh]

Already 5+ years ago, I talked to a director at a recruitment company in London that loved Google, because they were easy pickings to find good candidates to place in the roles he had because they were all - in his words - disillusioned - and at the same time the "Google" on their resumes still made his clients all starry-eyed. He saw recruiting people out of Google as easy money.

[trhway]

Montessori for adults.

[mikelevins]

"the magic is gone"

I first went to work at Apple in January of 1988, at the end of the era that Woz described as making whatever they felt like and throwing it over the fence to ravening crowds who were throwing bales of money over the other direction.

Some time in the first few weeks I was there, my department head convened a staff meeting to explain that they were going to try a new concept that Apple had never tried before. "It's called a 'budget'," she said. "We try to guess ahead of time what our plans are going to actually cost."

At some point someone asked if that meant that regular meetings would no longer be catered. The answer was, "of course not. We have to have some way to get people to show up."

[dalbasal]

How/why are search margins shrinking? Is it getting more expensive to deliver search? Online ad spending is going up, not down.

[ChuckMcM]

What Punch said, basically its a maturing of the Internet Advertising market, the early enthusiasm for lots of metrics on engagement being dampened by a gradual understanding of click fraud. This is reflected in Google's financials as a falling 'cost per click' (what Google gets for an ad click).

The other factor is that Google search has been slowly losing its quality edge on its competitors and so services like Microsoft's Bing start developing more significant inroads, the combination of more market share and higher CPCs on Microsoft's service lead it to be profitable last year[1]. And the reason is in part because Microsoft can preferentially send search traffic to their properties, but it is also tied to the fact that consumers don't then switch their search provider later.

To counter-act that Google has been paying more and more money to third parties to send search traffic their way (this expense is called 'traffic acquisition costs') and Google was horrified when Oracle's lawyer disclosed that they were paying Apple a billion dollars a year for their search traffic[2]. They paid millions to Mozilla to send their search traffic to Google [3] and had a material dip in traffic revenue when Mozilla cancelled that deal[4]. What the author of the article in [4] did not realize is that all of Yahoo!'s search traffic is in fact served by the Bing index and servers. Yahoo! hasn't had a native search service since they agreed to send it all to Microsoft. So Bing+Yahoo! market share is really just Bing market share. That money paid, their TAC, comes right out of the profit margin for their ads

At the end of the day this is neither good nor bad, search is being commoditized. And as it commoditizes that puts pressure on margins since discrimination based on cost always puts pressure on margins. Eventually Bing and Google will be equally expensive from an advertiser point of view and the one with the best operational efficiency will make the most profit. Neither of them will be able to use "excess profits" to fund moonshots.

That is why it is so critically important for Google to find additional viable businesses to augment its search advertising revenue.

The current road leads to a very un-fun place to work, long hours (maximize work done for salary paid), few benefits (reducing costs), and little room for new development. They won't end there, it's like saying if you drive west in the US and never turn eventually you'll drown in the Pacific ocean, it's true but there is always a turn before that point. But if Google of 2003 was one end of the spectrum, Google is inexorably moving toward the other end. Ruth is there to make sure what's left is still a viable business that deserves a high stock price.

[1] https://techcrunch.com/2015/10/22/bing-is-profitable/

[2] https://www.bloomberg.com/news/articles/2016-01-22/google-pa...

[3] http://www.pcmag.com/article2/0,2817,2398046,00.asp

[4] http://www.geekwire.com/2015/yahoos-deal-mozilla-draws-searc...

[MarkMc]

Here is Google's net profit margin for past 4 years [1]:

2013: 22.9%

2014: 21.4%

2015: 21.7%

2016 (9 months): 22.0%

It doesn't look like profit margins are under strong downward pressure. But a better metric would probably be 'revenue per search query' - I don't have any recent statistics, but if Bing can close that gap it will reduce Google's margins when bidding for search traffic from Apple and Firefox (and maybe force Google to increase the percentage of AdSense revenue they give to website publishers). But even then, Google will still have three significant advantages to avoid the complete erosion of their margins: (1) Android; (2) Chrome; (3) The Google brand (I'm sure that more people try to change the default search engine on their new Windows computer than try to change the default search engine on their new Android phone).

Also, Bing would probably not be profitable if it was a separate company and had to pay Microsoft for the privilege of being the default search engine on Windows. Bing+Yahoo seems to be growing market share, but I wouldn't expect them to completely close the gap with Google in the next 20 years because Google's ownership of assets like YouTube, AdSense, Google Maps, Android and Chrome.

Google has all my search history, a list of all the AdSense websites I visit, my YouTube history, my location history, ten years of my email history and my Play Store purchase history. And for any given search query I perform, Google can access a larger pool of similar search click history from other users. How will Bing ever overcome all these disadvantages?

[1] Derived from https://www.google.com/finance?q=NASDAQ%3AGOOGL&fstype=ii&ei...

[ChuckMcM]

It is so much more nuanced than that. That number is essentially manufactured for the street (and before anyone jumps down my throat about being some sort of conspiracy nut let me explain).

Early in Google's life there was an article which asked the question "Is Google's gross margin 100%?" It went on to observe that Google had overhead, staff, facilities, and monthly recurring costs like electricity and phone bills, but every advertisement they 'served' cost them essentially no delta in their overhead, so their "cost of goods" when looking at the business through the lens of the goods economy[1] was zero. And any revenue they generated by selling a "click" (their stand in for a widget) was 100% gross profit. As a result, using old school goods economy accounting rules, profit is entirely a function of how many clicks they "ship" and there are many knobs to adjust which boost how many clicks they ship. They include, but are not limited to, putting more ads on their own pages and buying traffic to point at their pages with ads on them.

But when you evaluate their business using the principles of information economics[1] you see that they do have costs and their core product's value is eroding.

To understand that statement, you have to ask "what does Google actually sell to advertisers?" An 'ad unit' or an 'ad word' is the name but what is it really? My claim is that what they really sell is a piece of very valuable information that is hard to get. They sell "this person has just looked for this good or service, now who wants to respond?" They create that information by providing a portal that people can write in a question.

This "solves" for advertisers their biggest ask, that a consumer looks at their advertisement exactly when they need the good or service that advertisement is promoting. You can put an ad for a sports car in a magazine about cars, you can put an ad for a sports car on a television show about cars, but that pales in comparison to the idea of putting your ad for a sports car in front of someone who has just asked "What's the best sports car?".

I think that it is pretty easy to see the value there to advertisers, they would rather spend money advertising to people looking for their products then advertising to everyone and hoping that some of them want their products. What is more is that it works well and that is why it has become a major force in advertising. It is also an interesting measure of the value too advertisers.

Advertisers will constantly evaluate what it costs to advertise against product sales or market penetration or market visibility. They have a limited budget to spend and they want to make it count. That sets up the traditional economic forces which demands they allocate their ad spend capital wisely to get the most impact. Google, unlike a television network or magazine, brought into wide use the notion of an auction that happens at the decision point, between potential advertisers on what they would pay to have their ad shown, and where it would be shown, in the search results for a given query. In an information economic sense this gives us the fundamental value of the information that Google sells.

The key being that all of the knobs that Google has at its disposal will eventually be turned to their maximum point, and if CPC keeps falling their profit margin will too. Because Microsoft's CPC number is going up their margins will continue to increase. At Google's current rates they will eventually be buying so much traffic and putting so many ad slots on their own sites that non-"free" services will appeal to more and more people which reduces the number of eye-balls on a Google ad which reduces its value to advertisers still further. If Google reaches the point where they are no longer able to adjust their traffic mix to achieve their numbers, they will be in a very tight spot.

[1] The "goods economy" is the system of directing capital into the production of goods. The "information economy" is the system of directing capital toward the disclosure of information.

[MarkMc]

I agree with your analysis that Google basically sells "this person has just looked for this good or service, now who wants to respond?". But you seem to be ignoring the large, persistent information advantage that Google has over Bing.

Consider 3 scenarios:

1. Person A just searched for "What's the best sports car?".

2. Person B just searched for "What's the best sports car?". He's a 45-year old man who works for an insurance company in New York. He has a wife and three kids, but he recently started seeing a 28-year-old psychology grad student in Chicago twice a month. Two years ago he bought a $6,000 watch for his wife's birthday.

3. Person C just searched for "What's the best sports car?". He's a 23-year old man who lives with his parents and does not have a wife or girlfriend. He works for Walmart and has just been promoted to manager. Nine months ago his friend bought a Subaru WRX.

Now consider what types of advertisers will bid for each search. And perhaps more importantly, what types of ads and landing pages will be offered for each search. Clearly, having detailed personal information is a significant advantage when deciding which ads to show to a user. Google has that advantage and doesn't look like giving it up any time soon.

However, I can think of two pieces of evidence against this argument: (1) Bing was able to outbid Google to be the default search engine on Firefox; (2) Outbrain and Taboola regularly outbid Google AdSense to advertise on Time, Forbes, Bloomberg, etc. In both of these cases I would have expected Google to prevail, so perhaps Google's personal-information advantage is not as strong as I think. Or perhaps Google isn't taking full advantage it.

[ChuckMcM]

I don't disagree, but consider that person B is a 45 year old man working for an insurance company he is probably running Windows 10, and Microsoft knows quite a bit about him.

My point is that in my opinion you are over valuing Google's user information advantage. Most purchasing and buying habit information is collected using web page based analytics to an advertiser rather than coming from Google directly.

When Blekko was operating at a consumer search engine we got a great inside look at what tools advertisers have at their disposal to "qualify" their ad bids. If you send the query's source IP to the ad network when you request an ad they match that up with all sorts of cookies and beacons that have fired off that IP address. None of that comes from Google, it comes from all the trackers that are running all over the Internet.

[yuhong]

>The key being that all of the knobs that Google has at its disposal will eventually be turned to their maximum point

I wonder what this would exactly look like. As a side note, I wonder what would happen if for example an economic collapse happened and Google's ad revenue was significantly reduced.

[ChuckMcM]

I see early signs of knobs at the extreme in the news, harder to get projects approved, people being told they can't do 20% projects any more, groups losing people that don't get replaced, web properties that are hard to monetize (I'm looking at you finance.google.com) getting starved.

I would expect that once a quarter there is a meeting which goes like this, "Ok folks our profit margin is going to be 22% and our revenue is $xxx. That leave $yy in expense we have to get rid off or revenue we have to develop. Who wants to go first?" We've seen both efforts on the outside, more adds on search pages, pay to be in the 'Google Shopping' bar, Etc. for revenue enhancement, and cancelling things like Reader, Wallet, Glass, and other projects or ideas that haven't added to the bottom line. You can see the moves in Youtube (the latest being you can't keep a video running in the background tab unless you have "Youtube Red" as an example) to either get profitable or die.

As for cash flow? I can tell you that I was there when the Mortgage Crisis hit and it scared the crap out of them. That crisis killed TechStops[1] and that did way more damage than the financial crisis did.

[1] Techstops were Google's tech help on demand group that, when they were employees, was a really awesome service. When it got switched to a vendor it really lost its utility.

[gniv]

> This is reflected in Google's financials as a falling 'cost per click' (what Google gets for an ad click).

But what is not known is the cost of a _similar_ click. Since the number of clicks increases dramatically quarter over quarter, it could be that it's the new clicks that are bringing the cost down.

[spydum]

Chuck, just want to say your detail and insightful comments are a tremendous asset to the community!

I do have to disagree though: MSFT has office and windows as major cash cows, maybe azure will also start turning some big profit in 2017 as well (tbd). So they can afford those r&d style projects without nearly as much risk as "Alphabet". I do agree with your assessment of goog though: the heat is turning up and I don't doubt belt tightening is their future.

[ChuckMcM]

I agree, Google really needs their version of 'office' or to make their cloud offering competitive in order to survive long term. I see them trying to extract more and more revenue from Android as one of their strategies.

[PunchTornado]

a lot of it is going to facebook & others.

[bitmapbrother]

And yet Google is making record profits YoY. As for Facebook, it would seem they've been cooking the books.

[nunez]

I hope they strip the stupid tape over their fridges that cover the sodas "because a 200-calorie coconut water drink is SO MUCH BETTER for you than a ZERO CALORIE soda."

Who am I kidding; that will never happen