Phenomenal article, very entertaining and aligns with my experience as a prominent search "outsider" (I founded the first search intelligence service back in 2004, which was later acquired by WPP. Do I have some stories).
The engineers at Google were wonderful to work with up to 2010. It was like a switch flipped mid-2011 and they became actively hostile to any third party efforts to monitor what they were doing. To put it another way, this would like NBC trying to sue Nielsen from gathering ratings data. Absurd.
Fortunately, the roadblocks thrown up against us were half-hearted ones and easily circumvented. Nevertheless, I had learned an important lesson about placing reliance for one's life work on a faceless mega tech corporation.
It was not soon after when Google eliminated "Don't Be Evil" from the mission statement. At least they were somewhat self aware, I suppose.
I'm really glad the article came out though, it fills in some gaps that I was fairly confident about but didn't have anything other than my sense of the players and their actions to back up what I thought was going on.
I and a number of other people left in 2010. I went on to work at Blekko which was trying to 'fix' search using a mix of curation and ranking.
When I left, this problem of CPC's (the amount Google got per ad click in search) was going down (I believe mostly because of click fraud and advertisers losing faith in Google's metrics). While they were reporting it in their financial results, I had made a little spreadsheet[1] from their quarterly reports and you can see things tanking.
I've written here and elsewhere about it, and watched from the outside post 2010 and when people were saying "Google is going to steam roll everyone" I was saying, "I don't think so, I think unless they change they are dead already." There are lots of systemic reasons inside Google why it was hard for them to change and many of their processes reinforced the bad side of things rather than the good side. The question for me has always been "Will they pull their head out in time to recover?" recognizing that to do that they would have to be a lot more honest internally about their actions than they were when I was there. I was also way more pessimistic, figuring that they would be having company wide layoffs by 2015 to 2017 but they pushed that out by 5 years.
I remember pointing out to an engineering director in 2008 that Google was living in the dead husk of SGI[2] which caused them to laugh. They re-assured me that Google was here to stay. I pointed out that Wei Ting told me the same thing about SGI when they were building the campus. (SGI tried to recruit me from Sun which had a campus just down the road from where Google is currently.)
Did I mention I was more pessimistic? :-) I expect that today they could layoff 150k, keep the 30K that are involved with search and enough ads that are making business and that husk would do okay for a long time. I don't suppose you watched SGI die, that happened to them, kind of spiraled into a core that has some money making business and then lived on that.
One of my observations between "early" Google and "late" Google (and like the grandparent post I see 2010 as a pretty key point in their evolution) was employee "efficiency." I don't know if you've ever been in that situation where someone leaves a company and the company ends up hiring two or three people to replace them because of all the things they were doing. Not 10x engineers but certainly 3 - 5x engineers. Google starting losing lots of those in that decade. They had gone through the "Great Repricing" in 2008 when Google lowered the strike price on thousands of share options. And having been there 5 to 10 years had enough wealth built up in Google stock that for a modest level of "this isn't fun any more" could just do that.
But aside from your observation that "they have plenty of people" it is similar to observing that a plane that has lost its engine at 36,000' has "plenty of altitude" both true and less helpful than "and here is the process we're going to use as we fall out of the sky to get the engines back on."
Google has lots of resources. If you have ever read about IBM reinventing itself in the 90's its quite interesting to note that had IBM not owned a ton of real estate it likely would not have had the resources to restructure itself. I worked with an executive at IBM who was part of that restructuring and it really impressed on me how important "facing reality" was at a corporation, and looking at the situation more realistically. I had started trying to get Google to do that but gave up when Alan Eustace explained that he understood my argument but they weren't going to do any of the things I had recommended. At that point its like "Okay then, have fun." Still, at some point, they could. They could figure out exactly what their "value add" is and the big E economics of their business and realign to focus on that. Their 'mission oriented' statement suggests that they are paying some attention to that idea now. But to really pull it off a lot of smart, self-interested, and low-EQ people are going to have to come to terms with being wrong about a lot of stuff. That is what I don't see happening and so I'm not really expecting them to transform. Both not enough star bits and the luma are just not hungry enough.
Are you suggesting that Google fire all the engineers who work on Cloud? That would... be a very interesting business decision, likely closing any door for them working with enterprise in the future.
Here's a few more realistic changes that Alphabet could make:
- shut down X
- shut down Verily
- sell calico or shut it down if no buyers
- sell Fiber or shut it down if no buyers
- shut down Intrinsic, Wing, and all the other X spinoffs
- make Cloud be its own Alphabet company with Kurian as an actual CEO
That would show Wall Street that GOogle is really serious about not wasting money on crazy ideas. That would boost the price (along with reducing costs) giving them some runway. I think it would be a shame if Waymo was shut down but it has a long, long way to being highly profitable.
It looks like Alphabet wants to sell Verily or spin it out of the Alphabet family entirely (after decoupling Verily's infrastructure from Google's) but nobody wants to buy it.
I was suggesting that they fire all the engineers that work on things that don't make money. It was only last quarter that Cloud actually made a profit. That said, I think you make a reasonable restructuring case; Now you just have to figure out how to get leadership to buy in and execute on that plan. In my experience two things work against that.
1) If it isn't their idea that don't believe it will do any good and could not possibly be the "right" thing to do.
2) If they don't have a job after it happens, they will work behind the scenes to sabotage attempts to make it happen.
You can work around those, but you need "existential risk" level energy to create that sort of change in a company.
Here here! Google needs to trim the fat, desperately! They need to eliminate all of their non revenue generating departments, ban all internal discussion forums and such that aren't laser focused on the job at hand. Cut 30-40% of all engineers, and get rid of the free food and other benefits. Install vending machines and charge for meals at their cafeterias, run them like any other business and make a profit. Get rid of free employee health benefits, make the employees pay for them. And for god sake get rid of that ridiculous swimming pool! Anything that isn't directly in the service of delivering value for shareholders needs to be done away with, starting with those hair-brained cash burning crazy X projects.
That it is, but a more apt comparison would be Duck Duck Go which was a contemporary of Blekko and definitely out performed relative to Blekko's success. DDG still going strong and even buying TV ads, so yeah.
That said, how Blekko and Watson ended up squandering good technology in search of something else is also an interesting learning experience/tale.
I agree with everything you are saying, but the stock price is up 6x over the last 10 years and revenue is still growing 13% a year so nobody at google is going to listen to any of this.
Basically ad dollars have continued to transition from old media to digital media over the last decade+ and that mass migration has created enough revenue to cover up all of Google's core problems.
The reality is that this firehose of money is what allowed those core problems to grow & fester. That and the practice of using TVCs for as much as possible, to the point where nearly every process that's documented is outsourced, and often no one inside really understands how things work anymore.
Looking at financials, all metrics are improving. They haven’t even started to lose altitude - they’re still gaining.
We might not like what they’ve become, but the comparison to a plane that’s lost its engine seems rather odd. Why couldn’t they keep going indefinitely, without making the changes that some would like?
Is IBM a good example? Like GE, their saving-grace restructuring was basically turning into a giant corporate leach (one through financialization, one through consulting).
ChuckMcM, I just wanted to say, I really appreciate the long view you bring to HN discussions. When you've been in tech for a few decades you start to see predictable patterns. History may not repeat, but it often rhymes.
Piggybacking on this to also express my appreciation. If/when you write a memoir someday, it would be a valuable historical record. If not, your hn comments are a wonderful corpus too :)
A) I think it’s important to acknowledge that in many ways Google is actively trying to keep CPC low - what they care most about is total spend. A low CPC means an effective advertising network where interested consumers are efficiently targeted. Their position is complex thanks to their monopoly status over online advertising.
B) I don’t think it’s fair to characterize recent layoffs as some put-off collapse… criticize Google all you want for running a bad search engine, but right now they’re still dominant and search is the most effective advertising known to man. They’re raking in buckets of money: they had 54K employees on 01/01/2015, and 182K on 01/01/2024. Similarly, they made 66B in 2014, and 305B in 2023. The latest layoffs are them cleaning house and scaring their workers into compliance, not the death throes of a company in trouble — they’re barely a dent in the exponential graphs: https://www.macrotrends.net/stocks/charts/GOOG/alphabet/numb...
A) This is short-sighted. What you're suggesting is in fact a way to optimize short-term gain over long-term viability. It's pure MBA tactics.
Additionally, it's complete and total oversimplification. If you look at Google's earnings it's pretty damn clear that at least until 2020 they were not just going for maximum total spend, but for a steady, gradual raise in total spend. Not too slow, not too fast. They were NOT taking every opportunity they had, in fact they're famous for systematically refusing many opportunities (see the original founders' letter, but even after that). They were farming the ad market, the ad spend, growing it, nurturing it. Then COVID blew up the farm.
Maybe you're right now, but I do hope they're recovering their old tactics. Because if they maximize it you'd see nothing but scams ... wait a second.
B) Google was built by providing a vision, and getting out of the way of ground-up engineer efforts. "Scaring workers into compliance" IS killing the golden goose.
You can see this in AI. Every story from an AI engineer that ran away from Google is the same. They didn't run away for the money, they ran away because they were getting scared into compliance.
Now AI may make it, or not. I don't know. But this is happening EVERYWHERE in Google. Every effort. Every good idea, and every bad idea runs away, usually inside the mind of "a worker". Not to make them personally maximum money, but it's natural selection: if the idea doesn't run away, the engineer it's in is "scared into compliance", into killing the idea.
Whatever the next big thing turns out to be, it simply cannot come out of Google. And it will hit suddenly, just like it did for Yahoo.
Totally agree on the overall prognosis of Google - I am (also?) one of said engineers! Here’s a recent update from a tiny corner of the company: the rank and file is still incredibly smart and generally well-intentioned, but are following hollow simulacrums of the original culture - all-hands, dogfooding, internal feedback, and ground-up engineering priorities are all maintained in form, but they are now rendered completely functionless. I am personally convinced that the company is — or was, before ChatGPT really took off - focused on immediate short term stock value above all else. After all, if you were looking down the barrel of multiple federal and EU antritrust suits and dwindling public support for the utility you own and operate, you might do the same…
I guess I’m standing up for the simple idea that terribly inefficient organizations can prevail when they’re the incumbents, at least for significant periods if not forever. We can’t be complacent and assume they’ll fall on their own, esp when AGI threatens social calcification on an unheard of scale.
Why would Google's collapse be for the good of humanity? When was a power vacuum ever beneficial?
"Build a better search engine for the good of humanity", I can understand. "Kill a search engine for the good of humanity" is a reductive, childish take.
They've already killed it in essence, so that they are hurting billions of humans with it daily. But they can still run it because it creates more revenue in this harmful form than it did in its helpful form. Therefore sabotage against that revenue is justified.
Sabotaging the revenue of Google search will weaken them against honest incumbents. They are currently well funded enough to kill incumbents. That will start to change as they decline, aided by our boycotts and other forms of sabotage. The decline and sabotage of Google is necessary for a better search engine to have the space to succeed.
A power vacuum is often good.
Linux and open source exists in a personal and collective power vacuum that was created by proprietary knowledge and software.
Sometimes power vacuums are colonized by people with good intentions. And it's neither reductive nor immature to help create those opportunities.
I never said that someone shouldn't sabotage Google as well as create a better search engine. I myself am working on llm-driven knowledge retrieval systems, at the same time as advocating for the destruction of Google.
Good luck and do anything in your power that you think will help humanity have good search again.
Very much appreciate the sentiment and kind words! Reminds me of Yudkowsky’s line[1] about AI: “we should be willing to destroy a rogue datacenter by airstrike.” This kind of talk sounds insane in the Silicon Valley language game, but we’re talking about real people’s lives here and sometimes implied violence needs to be made explicit. And that’s what I see your suggestion as, ultimately —- but that’s probably because I got an American HS education, so the Malcom X vs. MLK Jr. debate was driven into my mind quite thoroughly.
Luckily/unluckily I left already due to factors out of my control. Regardless, for all of Google’s faults I will say that they were incredibly serious about data security and respecting consumer data protection laws with strict oversight, so I think “sabotage” in a direct sense would be incredibly hard + risky. The only solution I see is continuing to organize for government regulation. I would include worker organization within Google, but I recently learned they represent less than half a percent of the company…
> You can see this in AI. Every story from an AI engineer that ran away from Google is the same. They didn't run away for the money, they ran away because they were getting scared into compliance.
Companies this size die several years before the body hits the floor.
They're dead when everyone starts to hate them and someone says "no, look how much money they're making, they're fine." That's the fatal blow, because they think they're fine, and keep doing the things that make everyone hate them.
At that point you're just waiting for someone else to offer an alternative. Then people prefer the alternative because the incumbent has been screwing them for so long, and even if they change at that point, it's too late because nobody likes or trusts them anymore, and ships that big can't turn on a dime anyway.
You have to address the rot when customers start complaining about it, not after they've already switched to a competitor.
I remember running into Kodak engineers, at an event in the 1990s, and they were all complaining about the same thing.
They were digital engineers, and they were complaining that film people kept sabotaging their projects.
Kodak invented the digital camera. They should have ruled the roost (at least, until the iPhone came out). Instead, they imploded, almost overnight. The film part was highly profitable.
Until it wasn't. By then, it was too late. They had cooked the goose.
If they owned the digital camera space like they should have, who’s to say they wouldn’t have eventually released a smartphone. It probably would have been an absolutely incredible camera first, and some mobile internet and phone features second.
One can really dream up a fascinating alternate timeline of iKodak if they didnt shoot themselves in the foot.
I'm not a Steve Jobs fan, but one business-quote I do like: "If you don't cannibalize yourself, someone else will."
In other words, it could have been better for Kodak as a whole if they allowed their digital-arm to compete more with their film-arm, so that as the market shifted they'd at least be riding the wave rather than under it.
The mistake Adobe made was in canceling Flash instead of open sourcing it. Publish a spec and the let browsers implement the client side, then you can keep selling tools to make animations without everyone having to deal with the bug-riddled proprietary player Adobe clearly had no interest in properly maintaining to begin with.
It's kind of astonishing that all these years later we still don't have something equivalent in browsers. In theory they're Turing-complete and you can do whatever you want, but where's the thing that makes it that easy?
What makes you think people want easy? /s I mean, clearly that would be best for creativity, for cultural robustness, for accessibility. Unfortunately, there are a lot of incumbents in all the spaces Flash touched who were ecstatic (if in a schadenfreude-esque sense) to see the ladder pulled up after them. When you make it difficult or impossible for the peons to create, you make it difficult or impossible for them to bypass the professionals and the gatekeepers; when they can't tell their stories, their stories get told for them. Again, the professionals and the gatekeepers (and, now, the propagandists) find this ideal.
Suffice it to say, there are a lot of people who worked very hard to make sure that the 1998-2012ish period of openness and open-access and democratization was an anomaly. You got to see a mini-echo of this with the rollout and rollback of pandemic-era accessibility.
The just-so story about Kodak is one of those things that bugs me. Kodak did own the digital camera market, stem to stern, for years. They did not ignore it. They did, however, invent all that stuff a little early, before the semiconductor manufacturing technology had matured to the point where it could be a consumer good.
The company imploded because it spent all of its time, attention, and capital trying to become a pharmaceutical factory, starting in the mid-1980s.
Yeah, lots of things happened for a perfect storm of downfall…probably starting with the antitrust breakup of the film processing division.
They did indeed have a huge patent arsenal from all their research efforts that was very valuable. They were also really good at consumer tech - so it’s a shame it didn’t amount to more.
One of the problems was just how profitable film was. No amount of digital camera sales is going to be as profitable as being able to charge people $2 per photo (film+development).
Fujifilm survived by diversifying more into a chemical company than a consumer product company (whereas Kodak sold off those portions of the company as "not being core to consumer imaging" and focused on printers(??))
And yet even Fuji are now back to having traditional film photography being their single largest revenue generator (their instax instant film is now so popular it is chronically sold out and they are doubling factory capacity to keep up)
Any examples of this actually playing out with a company as established as Google? You can read comments like this on many companies... Microsoft (70B$ income), Meta (40B$), Oracle (8.5B$), IBM(7.5B$), SAP (6B$), yet none of them seem to ever actually enter the predicted death spiral.
And the internet isn't new anymore. There is no vast landscape of unexplored new technological possibilities, and no garage start up with an engineering mindset that will just offer a better solution.
IBM used to be bigger than MS, it's a 10th of it today.
But most importantly all the above listed companies with the exception of Meta are those that are heavily ingrained in large companies operations. IBM still provides mainframes, MS has Exchange and Windows domains and is successfully transitioning a lot of customers to Azure, Oracle has their databases and other products, SAP their ERP systems.
Once a non-IT company has their internal IT systems and some legacy working they're going to be very very slow in changing them out if it works, companies that provide those and get a critical are going to have very very long runways compared to regular b2c companies if a significant portion of their revenue comes from this.
Google has Chromebooks that are used in schools and some GCP usage but could that save Google long enough if search revenue was cut into a fraction? And GCP is kinda of an also-ran today, people looking at larger options usually look at AWS(nr 1) or Azure (Windows legacy).
In 2023 the revenues of Google Cloud, Youtube Ads and "Google Other" and Google Network Members Ads were 130B combined.
If they could reduce headcount and operating expenditures to 2019 levels without losing that, they would be roughly breaking even without any search. They also have 280B$ in equity to tide them over.
When Google actually sees its business failing, it will have many many many chances to turn things around.
Microsoft and Meta reinvented themselves a few times over. At this point Windows is just an legacy business unit for instance, and Meta literally changed name to mark the turn.
Oracle, IBM and SAP have the advantage(?) of being heavily business focused from the start, and I don't see them ever die a natural death in our lifetimes. As long as they have the money to outbribe the competition they'll be there, and it will require a small miracle to break that loop.
The one thing that has kept Microsoft afloat is their business oriented part. They are deeply entrenched in any company that needs to use Office and only ever hires Windows admins who won't even look beyond Windows. That is pretty much every non tech small to medium company. When things were shifting to the cloud they were smart enough to make sure it would be their cloud, locking customers even deeper into their own technology.
To add to this, Microsoft is really really good at understanding businesses, in a way Google will probably never be I think.
Having on premise hosting options for Exchange and all their core services is an example of that, even as they're also pushing for 365 in the cloud. I remember them being earlier than GCP to deal with GDPR and the in EU requirements as well but my memory might be failing.
People use Windows at home and at school and then employers use the same thing because they don't want to retrain people. But the home versions of Windows are becoming so malevolent that they're losing market share. Meanwhile all the things that used to require Windows are becoming web pages and phone apps. You go to a university and it's full of Macbooks and if you see a PC in the CS department there's a good chance it has Linux on it. These are the people who will be choosing what to buy in a few years.
But who cares about the clients anymore, right? They're making money from cloud services. Except their hook is getting people to use Active Directory and Microsoft accounts, which are the things for managing Windows client devices.
It's going to be a while before anybody convinces the accountants to stop using Excel, but for large swathes of employees Windows is no longer relevant, and if you don't need Windows then why do you need Azure instead of AWS or any of the others?
> if you don't need Windows then why do you need Azure instead of AWS or any of the others?
I don't have enough insight, but there's more to it than Windows/Microsoft services tie up. It's clearly not the ease of use for small customers, it could be the contract making, or something else that makes it better deal for businesses beyond just the cost bundling.
For instance I remember Apple hosting iCloud on Azure. And there's a few other big players going with Microsoft, especially retail chains who can't touch anything Amazon, and don't trust Google.
It's the ease of use for medium customers. Large customers have Linux servers with full-time staff to write custom code and do whatever they want because they have their own resources; Facebook doesn't use Azure. Small customers buy a Macbook or Chromebook or tablet and have a gmail address and host their website on WordPress or one of those awful (but easy to use) web host proprietary site builders.
Medium businesses are big enough to want to have their own email domain but not big enough to want to implement their own spam filter, so they turn to the likes of Amazon and Google and Microsoft. Then Microsoft's advantage is they can manage and integrate with your Windows devices. Otherwise they're just doing price competition with every other hosting company. People who aren't even using Active Directory start to wonder why they should pay extra for SQL Server instead of using Mariadb on Linux, and in turn why they shouldn't put that VM on AWS unless Microsoft cuts them a better deal. (Which is presumably what happened with Apple, but offering long-term discounts is not how you make a lot of money.)
Moreover, it's increasingly easy each year for companies to support BYOD and let employees procure whatever they want that meets IT requirements. My current employer gave all non-tech staff $2000 to buy themselves a laptop, which was then enrolled in some fleet management systems with almost a single click.
Frankly, I see very few people choosing Windows anymore.
Also, another point to add: Microsoft's Intune fleet management system is perfectly capable of managing Macs, and you can use AD as your IDM source of truth for just about anything, including SSO for Google Workspace & ChromeOS devices.
To your last point, Windows Server is a hard requirement in many enterprises because of legacy or procured software that requires it. That is entirely separate from end user computing.
(I used to run end user computing for an F500, and I also ran the Enterprise Apps org at the same time. This was from about 2008-2015, and initiatives including mass migrations aware from MS Office to Workspace, and replacing thousands of Windows laptops with Chromebooks.)
I think many of us are underestimating Microsoft because of how crappy Windows is and keeps being.
But as a business entity they've been ferocious from the start, and succeeded through sheer perseverance where Google gave up after some tepid tries.
Xbox would have been killed by Google in the first year. Exchange would have stayed in beta for a decade, and Office365 would have had no support if it was in GSuite.
If Google were to find a way, I think they'd need a radically different approach, as I don't see them ever fixing their focus problem.
I think that's a valid point. Maybe Google culturally will not be able to turn around. Not because crappy product, but because of lack of focus.
That said, Google is still printing money and increasing profits and revenues. Nothing like falling profits (or even losses) to create some pressure to focus. DEC would be the example of a company that failed to do so.
Reinventing yourself because you imploded your primary market is still an own-goal. If you can capture a new market then you could have had both. And what if the primary market collapses first?
AT&T, GE, AOL, Yahoo, Sony technology (they are a media company now, but they did used to make things that weren't a game console), Time Warner, SGI, Compaq, 3DFx, DEC...
Not only that, most of the other examples are just not at the end of their death spiral yet. Take a look at Windows market share, it's down 20% over the last 10 years:
And that's just desktop. Microsoft ceded the entire mobile market, which in turn now represents the majority of devices. The majority of the company's profits no longer come from selling Windows and Office. If they hadn't pivoted into a new line of business (Azure) they'd be on a trajectory to impact with the ground.
IBM has been bleeding customers -- and business units -- for decades. Their stock is flat, not even keeping up with inflation, compared to +300% over the last decade for the overall market. And they have no obvious path to redemption.
Oracle is kind of an outlier because of the nature of their business. Their product has an extraordinarily high transition cost, so once you're locked in, they can fleece you pretty hard and still not have it cost more than the cost of paying database admins high hourly rates for many hours to transition to a different database. Then they focus their efforts on getting naive MBAs to make a one-time mistake with a long-term cost. Or just literal bribery:
And even with that, their database market share has been declining and they're only making up the revenue in the same way as Microsoft through cloud services.
Meta isn't a great example because people just don't hate them that much. Facebook sucks but in mostly the same ways as their major competitors, they're still run by the founder and they do things people like, like releasing LLaMA for free.
All of the companies I cited are hugely profitable. They might not be as large as they once were, or as important, but a business that has non-declining net income in the billions is not in a death spiral. IBM has shrunk a lot, but except for the financial crisis in the 90s, they have been profitable every year, and profits are roughly flat since 2017.
This is certainly a completely different picture than Yahoo for example.
And your argument for Microsoft is that they are in a death spiral because they only have 70% of market share on the desktop, and are shrinking by 2% per year, so in, uh 15 Years they might only have 50% of the market share! Also, please ignore that they successfully diversified their revenue streams to other markets (Cloud).
And your evidence is that they failed to capture the mobile market. While you also argue that Google is in a death spiral when Google is actually the company that won the mobile market.
I think you might be using the term death spiral in an unconventional way here.
> All of the companies I cited are hugely profitable.
You cited them because they are hugely profitable, ignoring the ones that are already defunct. And the entire premise is that a company can simultaneously be posting profits while doing the thing that will ultimately destroy them.
> And your argument for Microsoft is that they are in a death spiral because they only have 70% of market share on the desktop, and are shrinking by 2% per year, so in, uh 15 Years they might only have 50% of the market share!
Platforms have a network effect. They're doing so poorly that the network effect from having 90% market share isn't enough to prevent them from losing market share. But now they only have the network effect from 70% market share, which makes it even easier for customers to switch. That's how you get a death spiral.
> Also, please ignore that they successfully diversified their revenue streams to other markets (Cloud).
Which are in turn dependent on customers using Windows so they need Active Directory etc. See also:
> And your evidence is that they failed to capture the mobile market. While you also argue that Google is in a death spiral when Google is actually the company that won the mobile market.
It is unquestionably the case that Microsoft lost the mobile market, which is the larger market. Android has the most worldwide market share, but Android is free to use and generates revenue for Google only to the extent that people want their services. If people stop wanting their services and switch to e.g. another search engine, how does it save Google from this even if they're using Android?
Yeah, it's a pain in the butt. It often shows you the graph and then you try to show the link to someone else and it tries to get them to swipe their card as if anybody is going to do that. Meanwhile it ranks highly in Google search results instead of some other site that contains the same information without the bait and switch, because Google has completely lost the ability to produce quality search results.
AT&T: 15B$ net income, world largest telecom company. #13.0n Fortune 500.
GE, while a reasonable example of a company that declined severely from its peak, was still generating 9B$ in income on 2023 before being split in better focused and profitable successors.
AOL/Yahoo were never dominant in a mature market. They were early to the Internet, but this was an uncharacteristically volatile time with an exponentially growing market.
Sony is also a leading manufacturer in several tech sectors (second largest camera, largest premium TVs). 6B$ net income and rising.
3DFx was never dominant in a mature field but, again, early in a nascent one. They collapsed quickly, not through some highly profitable extended death spiral.
Compaq was never dominant in a highly profitable field. Their market share peaked at 14%.
DEC might be a genuine example, they were never the top of the field, but they did not manage to adapt and turn things around when the world moved in a different direction. Compare to IBM who _were_ in a dominant position in the same field, and have leveraged that position into a sustainable and steady, if smaller and less groundbreaking, business.
Google might be in trouble (relatively speaking) if LLMs disrupt search, but they are not close to being in trouble from being outcompeted in search itself.
AT&T: today is not AT&T. The name was bought. It used to be Cingular.
GE: so your point is that it is a good example.
AOL/Yahoo: A 'mature market'? Are you making up rules so you can disqualify them?
Sony today only innovates in image sensors. They are a financial and entertainment company. Who cares if they sell the most 'premium TVs', this is the company created (off the top of my head) Betamax, CDs, DVDs, Minidiscs, Trinitrons, and made the best consumer tech in the world -- consistently.
3Dfx was the leader of an industry that is now lead by nVidia. That industry wasn't as big then, but everyone knew it would be and it was theirs to lose.
Compaq was the market leader in PC sales in the 90s.
I used the term "as established as google". In my mind this certainly meant the market has to be established. As long as an industry is brand new and rapidly developing, things are obviously different. Many early market leaders didn't make it in the internet. But in the last ten years, market leaders haven't been failing in the same way.
So no, not changing the rules, but maybe clarifying the point. Situations such as the rise of the internet in the late 90s and early 2000s are the anomaly, not the rule.
Operating Systems and Internet search are roughly the same they were ten years ago. 3d accelerator cards changed immensely in the years when 3dfx failed. Microsoft and Google are not in businesses where younger agile companies that read the changing tides better can quickly supplant them.
And that's why they get a thousand chances to turn things around while printing money with their "death spiralling" business.
Your question is effectively answerable as 'no' if you want to limit it to exactly google like market positions, because they haven't existed before. I was answering with examples of market leaders that fell due to bad top-down decisions.
While I totally agree that Atlassian products are terrible and steadily getting even worse, I'm not sure they are going anywhere anytime soon given their disconnect between users and customers. Most people who have to suffer their products have no say in the purchasing decision, and the company does a somewhat better job of appealing to the relative small group that does. Atlassian could very well have Oracle-like staying power.
You know how a chess player will say something like "mate in 6" because their experience of all the options left to their opponent are both easily countered and will not prevent them from losing? Companies, and tech companies in my experience, get into death spirals due to a combination of people, culture, and organization. Pulling out of one of those is possible but requires a unique combination of factors and a strong leadership team to pull off. Something that is very hard to put into place when the existing leadership has overriding voting power. You can look at GE, IBM, and to some extent AT&T as companies that have "re-invented" themselves or at least avoided dissolution into an over marketed brand.
I have a strong memory of watching a Jacques Cousteau documentary on sharks and learning that Sharks could become mortally wounded but not realize it because of how their nervous system was structured. As a kid I thought that was funny, as an engineer watching companies in the Bay Area die it was more sobering.
If you have read the article, I think Gomes was right and saw search as a product, whereas Raghavan saw it as a tool for shoveling ads. A good friend of mine who worked there until 2020 wouldn't tell me why they left, but acknowledged that it was this that finally "ruined" Google.
Their cash cow is dying, I know from running a search engine what sort of revenue you can get from being "just one of the search engine choices" versus the 800lb gorilla. Advertisers are disillusioned, and structurally their company requires growth to support the stock price which supports their salary offerings. There is a nice supportable business for about 5,000 - 8,000 people there, but getting there from where they are?
My best guess at the moment is that when they die, "for reals" as they say, their other bets will either be spun off or folded, their search team will get bought by Apple with enough infrastructure to run it, Amazon or someone else buys a bunch of data centers, and one of the media companies buys the youtube assets.
> You know how a chess player will say something like "mate in 6" because their experience of all the options left to their opponent are both easily countered and will not prevent them from losing?
As a chess person, saying "Mate in _" means it's a calculated inevitability. There is no mathematical way out of it.
It is not nearly equivalent to the outside judgement of a company with so many factors — it's just incomparable.
I don't disagree, chess is much more algorithmic and predictable. Maybe it is more like seeing your best mate of the last 20 years getting into their fourth or fifth relationship with the same kind of partner they failed with before and thinking, "Seen this movie before, it is not gonna work out." No algorithms, just you know how you're friend sabotages themselves and you also know they can't (or won't) look critically at that behavior, and so they are doomed to fail again.
But I can guarantee you that Google employees are reading these comments and saying "Wow, this guy is totally full of it, he doesn't know about anything!" and for some of them that thought will arise not from flaws in what I and others are saying, but in the uncomfortable space of "if this is accurate my future plans I'm invested in are not going to happen..., this must be wrong." I have lived in that space with an early startup I helped start, when I went back and worked on the trauma it had caused me it taught me a lot about my willingness to ignore the thinking part of my brain when it conflicted with the emotional part.
You have to do some of that to take risks, but you also have to recognize that they are risks. Painful lesson for me.
Yes, but there are other positions that do fit the comparison, like a couple of advanced passed pawns that can still be defended against with surgical precision, but most times are lethal.
Again, I think there is a misunderstanding of what the saying is used for.
In chess, it's specifically used for saying "even with the best defense possible, you will be mated no mater what in a maximum of X moves." Computers use this definition as well. If Stockfish says # in 6, that means there is an indefensible path to mate available, and with the best play of the opponent will take 6 moves.
I don't think so. At one of the Sun Reunion events a bunch of us sat around and talked about it. I suggested someone should write a companion volume to "Sunrise: The first 10 years of Sun" called "Sunset: The last 10 years of Sun." But as far as I know nobody followed up (if they did they didn't reach out to me for my take)
With Google, I always feel like the side hustles (waymo, X, etc.) Really exist to be sold off in the future to prop up the add/search business and ensure future profitability. Everything not adds/search is on that list, and anything shut down despite being useful isn't seen as future-sellable.
Google today is starting to smell of future financial engineering games, like when a car maker earns more through financing than selling core product.
fwiw, there are approximately 25,000 FTEs reporting up to Thomas Kurian, and I'm not sure how many thousands of TVCs. That's just for Cloud, and doesn't include the massive numbers of additional, relevant employees directly support Cloud from within TI. Part of Google's problem is that it's so big and so broad, and has always insisted on a monorepo for internal source code, and it's outsourced to vendors as much as possible, that it's nearly impossible to disentangle any one business unit from the next. I predict that if the FTC or the EU seriously try to break up the company, this will be there argument against it.
The majority of that revenue comes from violating data protection law and regulators and litigants are slowly racking up a series of wins which will gut ads margins.
There is no Plan B, they are just going to break the law until they can’t and there’s zero clue what happens after that.
They sat back and let OpenAI kick their ass precisely because ghouls like Prabakar call the shots and LLM are not a good display ads fit.
People would add sites for a particular topic (aka slashtag) to their list. That would build a virtual custom search engine within the search engine. And topic specific searches thrown at it would consistently out perform Bing and Google in terms of search quality. The meta "spam" slash tag (everyone got their own) would let you tell the engine sites you never wanted to see in your search results so if you were tired of your medical queries being spammed by quacks, add them to your spam list and they wouldn't be in your results.
Ultimately, lack of traffic. During Blekko's lifetime Google went from paying people less than $10M/quarter to send their search traffic to Google to over $4B/quarter to do that. If you are ad based you need traffic to serve ads to.
At some point a pay for search model might emerge that has a big enough audience to support a company but that time is not yet here.
1 . Does that mean blekko was something similar to millionshort ?
2. Was blekko capable of tackling seo sites or blogspam taht we have today or it had the advantage of low spam site count from the old web ?
3. Does it have a chance of coming back like how yahoo has been recently hinting a comeback ?
4. A stupid question : How much will it cost it build a blekko today ?
Not sure on #1, definitely mitigated SEO sites and blogspam (on an individual basis, if you added a spammy blog to your personal spam list it stopped showing up in results). As a result on slashtag searches there was very little spam.
Would it come back? As it was? no. The folks at Bing used some of the techniques to mitigate some spam in Bing results but didn’t implement slashtags.
It would cost between $3 - $6M to go from scratch to developing a 3 billion page index with a 10 billion page crawl ‘frontier’. You can seed the crawl with Common Crawl. If you can get $10 RPM’s ($10 per thousand queries) and roughly 10M queries/day (so $10k/day recurring revenue) you can run an operationally cash flow positive business. You would want to grow it organically to a 10 billion page index on a 100 billion page crawl which would cover 90+% of the english language queries. With clever optimizations (like a news sieve to only index pages about the news that made ‘sense’) you might improve efficiencies. You would also want to focus on reference applications (people who use search to get their job done) for paid subscriber growth, and simpler commercial partnerships for managing ad lead generation on commercial search (people looking for products or services).
Also you would need to be an advertising ‘primary’ (not taking feeds from networks on a revenue share model) So, for example, working directly with Amazon to both efficiently access their internal product index and to surface it on commercial queries. Note people like Amazon do their own advertising business on their own index so you compete with that to some extent and navigating that early is essential.
Certainly doable but not something that your typical venture fund would go for. It would have a longer payback time (lower internal rate of return) than VC’s look for.
The 2010-2013 timeline was also when the problem of ad fraud exploded. Google even acquired a company (or multiple, if I recon correctly: https://www.ft.com/content/352c7d8e-9acc-11e3-946b-00144feab... ). You had these companies popping up left and right that were snooping on Google and the emerging programmatic advertising environment to see if the websites and the traffic delivered were legit, and there were some scary numbers flying around.
The whole problem kind of got swept under the rug with most advertising ecosystems implementing a checkbox solution for clean traffic, and the web turned mobile user first.
My impression is that ad fraud never disappeared - it just got sanitized and rolled in with the other parts of the ad stack.
How much of (online) advertising is legit? Does any one know?
What would a "healthy" ad ecosystem look like? What should the the FTC (and advertisers) be working towards?
Eliminate any potential conflicts of interest? Bust up vertical integration (eg search & ads must remain separate companies)? Independent verification, as best able (eg like Nielsen does for ratings)?
Or maybe we determine (digital) ads based biz models are irredeemable, and we figure something else out.
I don't know what caused it but I suspected at the time, and still do, that it was simply business people getting more involved in order to drive growth.
The hostility was simply this. One day we had a dedicated high level Google engineer helping us out and giving us guidance (and even special tags) to get the data we needed in a cost effective manner for both Google and us. The next day, he was gone and we received demands to know exactly what we were doing, why and even sensitive information about our business. After several months of such probing, we were summarily told that the access we had was revoked and that there was no recourse.
We circumvented by setting up thousands of unique IP addresses in 50+ countries throughout the world and pointing our spiders at Google through them (same as they do to everyone else). These were throttled to maintain very low usage rates and stay off the radar. We continually refilled our queues with untouched IPs in case any were ever blacklisted (which happened occasionally).
As for what we did, we sampled ads for every keyword under the sun, aggregated and analyzed them to find out what was working and what wasn't. This even led to methods for estimating advertiser budgets. At one point, we had virtually every Google advertiser and their ongoing monthly spend, keywords and ad copy in our database. Highly valuable to smart marketers who were looking for an edge.
I enjoy reading this chap's stuff. It's not the way that I would write, but he's got a much broader audience than I do, so he obviously is meeting the needs of the reading population.
I do feel that I can't argue with his stuff, although it is very dark and cynical (and, truth be told, I have a lot of dark and cynical, in me, as well, but I try not to let it come out to play, too often).
How many companies have management consultants taken down? It's quite amazing how bad they are at anything. Peter Thiel's hatred for consultants is really legit.
The engineers at Google were wonderful to work with up to 2010. It was like a switch flipped mid-2011 and they became actively hostile to any third party efforts to monitor what they were doing. To put it another way, this would like NBC trying to sue Nielsen from gathering ratings data. Absurd.
Fortunately, the roadblocks thrown up against us were half-hearted ones and easily circumvented. Nevertheless, I had learned an important lesson about placing reliance for one's life work on a faceless mega tech corporation.
It was not soon after when Google eliminated "Don't Be Evil" from the mission statement. At least they were somewhat self aware, I suppose.