3 ms·

> When I left in 2010 I shared in my exit interview that they should seriously consider embracing actually investing in customer support and leaning into the id

by Permit 5y ago

> When I left in 2010 I shared in my exit interview that they should seriously consider embracing actually investing in customer support and leaning into the idea of a "normal margins[1]" kinds of businesses. I speculated that as their advertising margin continued to deteriorate (it was already going down fairly rapidly by then) that without these sorts of 'base load' businesses to carry the operational costs Google would begin to struggle to maintain goodwill with their "free" services and "lavish" workplace perks. What would it have looked like if you were right about this? What would it look like if you were wrong? I ask because later you mention that you left some money on the table, but it looks more like you left most of it (85%+) on the table (which is fine and impossible to know ahead of time!). It just seems to me like you made some predictions but every metric points to those predictions being incorrect. Have perks been removed? Has compensation fallen?

ChuckMcM 5y ago
> What would it have looked like if you were right about this? What would it look like if you were wrong? Google's per ad click revenue (CPC) started falling like a rock off a cliff during the mortgage crisis and has continued to fall since. Google also sold ads through third party sites (AdSense for Content) where they shared a meaningful and proportional chunk of that revenue with the site provider. As part of their employee onboarding I had added AdSense ads to my personal web site to get a feel for how that part of the business worked. I also have a number of friends who ran web sites with AdSense ads and we all saw exactly the same pattern, for the same number of page views the "revenue" generated by the ads was tracking the decline in CPC. What that said to me was that if Google maintained the level of advertising on its own sites and continued to share the same revenue with third party sites, then in roughly 5 years it wouldn't be enough money to cover costs. At that rate in 2013 or 2014 Google would have its first layoff, and then by now it would be a much smaller company or perhaps bought by someone (Apple comes to mind). (I did say I was pretty pessimistic right? :-)) But "making more inventory" (putting more ads on a given page) has zero marginal cost and if you add more inventory than the cost per ad decreases, statistically you increase your revenue. Also if you're the #1 search engine you automatically get the best possible ad prices because you have the biggest audience. But what is #1? Well its the search engine with the most traffic going through it, and if you're not organically getting enough traffic you can pay people to send you traffic. When I started at Google they paid less and $100M a year to third parties to direct traffic to Google search, by 2015 they were paying a billion dollars a quarter for that traffic. Why? Because if they didn't pay that the traffic would go to other equally good search experiences (notably Bing). I've watched Bing too and seen their CPC values go up while Google's were going down. The most likely explanation of that is their increase in market share and the fact that a lot of Windows people didn't bother to change search engines and of course they offered fairly cheap API access for third parties who wanted to create a custom search experience but didn't want to invest in owning and operating a web crawler, indexer, and ranker. So if I had been right, Google would right now be a division of Apple or maybe Microsoft. Meta (nee Facebook) could have made an excellent play here and wiped out Google by integrating a search engine into the Facebook experience. They chose not to, largely because Mark just didn't "get it." He did not appear to understand search is social, after all when you're going to buy a new car who do you ask for recommendations? Your social network of friends. Blekko demonstrated a really killer implementation where we could return search results ranked by what you and your friends liked vs what random people like (which is what Google/Bing do) and it was pretty amazing (not great for privacy and had a bunch of other implications like finding out one of your friends likes web sites selling BDSM gear when you search for mask :-)) So as the search revenue declined, Google has compensated by adding inventory (using search for things like products is nearly all ads), buying ever more traffic through deals with folks like Apple and Mozilla, and cutting costs by killing off projects/products. Sometimes directly and sometimes by forking the people responsible into another group under Alphabet and then quietly selling or closing down that group. The mean time to live once you've been spun out into an Alphabet "other bet" seems to be pretty short these days. I've got lots of ideas and anecdotes about how they got there, but basically their behavior has now put them into a very difficult position. Starting a new service costs money, the market doesn't trust you'll stand behind your service, so adoption (during which you lose money on the service) is slower. As a result you lose more money than you might have with rapid traction and growth and this hesitancy masks the "is it good?" signal. So you run with the service, which is losing money, unable to distinguish between it losing money because nobody likes it, and losing money because people like it but are afraid to adopt it. If you then cancel the service from a lack of adoption, you increase the hesitancy and you don't really know if the offering would have been successful had it not carried the 'baggage' of people not trusting you. Obviously, even though I was wrong about Google "being dead" within 10 years I'm still not bullish on their future prospects.
- Permit 5y ago
  Fantastic response, thanks for taking the time to break it down!
- bitcharmer 5y ago
  Wow, what a great insight. Thanks for sharing
- ithkuil 5y ago
  True true. Also, this is a story about the curse of success. Many of their cancelled products were really things that nobody cared about, but since google itself was so high profile, that a relatively few disgruntled users had such a high resonance because reporting and echoing their woes generated clicks elsewhere.

by Permit 5y ago

ChuckMcM 5y ago

> What would it have looked like if you were right about this? What would it look like if you were wrong? Google's per ad click revenue (CPC) started falling like a rock off a cliff during the mortgage crisis and has continued to fall since. Google also sold ads through third party sites (AdSense for Content) where they shared a meaningful and proportional chunk of that revenue with the site provider. As part of their employee onboarding I had added AdSense ads to my personal web site to get a feel for how that part of the business worked. I also have a number of friends who ran web sites with AdSense ads and we all saw exactly the same pattern, for the same number of page views the "revenue" generated by the ads was tracking the decline in CPC. What that said to me was that if Google maintained the level of advertising on its own sites and continued to share the same revenue with third party sites, then in roughly 5 years it wouldn't be enough money to cover costs. At that rate in 2013 or 2014 Google would have its first layoff, and then by now it would be a much smaller company or perhaps bought by someone (Apple comes to mind). (I did say I was pretty pessimistic right? :-)) But "making more inventory" (putting more ads on a given page) has zero marginal cost and if you add more inventory than the cost per ad decreases, statistically you increase your revenue. Also if you're the #1 search engine you automatically get the best possible ad prices because you have the biggest audience. But what is #1? Well its the search engine with the most traffic going through it, and if you're not organically getting enough traffic you can pay people to send you traffic. When I started at Google they paid less and $100M a year to third parties to direct traffic to Google search, by 2015 they were paying a billion dollars a quarter for that traffic. Why? Because if they didn't pay that the traffic would go to other equally good search experiences (notably Bing). I've watched Bing too and seen their CPC values go up while Google's were going down. The most likely explanation of that is their increase in market share and the fact that a lot of Windows people didn't bother to change search engines and of course they offered fairly cheap API access for third parties who wanted to create a custom search experience but didn't want to invest in owning and operating a web crawler, indexer, and ranker. So if I had been right, Google would right now be a division of Apple or maybe Microsoft. Meta (nee Facebook) could have made an excellent play here and wiped out Google by integrating a search engine into the Facebook experience. They chose not to, largely because Mark just didn't "get it." He did not appear to understand search is social, after all when you're going to buy a new car who do you ask for recommendations? Your social network of friends. Blekko demonstrated a really killer implementation where we could return search results ranked by what you and your friends liked vs what random people like (which is what Google/Bing do) and it was pretty amazing (not great for privacy and had a bunch of other implications like finding out one of your friends likes web sites selling BDSM gear when you search for mask :-)) So as the search revenue declined, Google has compensated by adding inventory (using search for things like products is nearly all ads), buying ever more traffic through deals with folks like Apple and Mozilla, and cutting costs by killing off projects/products. Sometimes directly and sometimes by forking the people responsible into another group under Alphabet and then quietly selling or closing down that group. The mean time to live once you've been spun out into an Alphabet "other bet" seems to be pretty short these days. I've got lots of ideas and anecdotes about how they got there, but basically their behavior has now put them into a very difficult position. Starting a new service costs money, the market doesn't trust you'll stand behind your service, so adoption (during which you lose money on the service) is slower. As a result you lose more money than you might have with rapid traction and growth and this hesitancy masks the "is it good?" signal. So you run with the service, which is losing money, unable to distinguish between it losing money because nobody likes it, and losing money because people like it but are afraid to adopt it. If you then cancel the service from a lack of adoption, you increase the hesitancy and you don't really know if the offering would have been successful had it not carried the 'baggage' of people not trusting you. Obviously, even though I was wrong about Google "being dead" within 10 years I'm still not bullish on their future prospects.

Permit 5y ago
Fantastic response, thanks for taking the time to break it down!
bitcharmer 5y ago
Wow, what a great insight. Thanks for sharing
ithkuil 5y ago
True true. Also, this is a story about the curse of success. Many of their cancelled products were really things that nobody cared about, but since google itself was so high profile, that a relatively few disgruntled users had such a high resonance because reporting and echoing their woes generated clicks elsewhere.