The Critical Path #110: The Profit Algorithm

I go on and on about Google.

via 5by5 | The Critical Path #110: The Profit Algorithm.

  • Brian Anthony

    Horace, I have listened to several of your podcasts about Google and I have some understanding of their income that may help you. My information comes from a small business owner. The impression I get from my neighbor is that he pays Google to send people that searching for his product (in Austin TX) directly to his website. For this service he pays google based on the number of people sent. It seems that he can pick how many clicks a month and the fee he pays increases as this number of requested clicks increases. I also get the impression that the sales from his business more than double based on his relationship with Google and he couldn’t be a happier customer.
    It is possible that the reason Google doesn’t discuss this is they don’t want the public to know that the order in which links appear when we search depends on how much Google is being paid by the hoards of small businesses. I don’t think this revenue model is limited to the links that are clearly ads, I think it’s all the links. This also explains why so much of Googles revenue is based in the US and why they spend so much on SG&A.

  • Karthik

    Nice podcast on Google Horace…. Another point I would like to make (its a version of what I used to hear from my faculty / colleagues back in grad school is

    There are 2 types of researchers or scientists.

    The first type have an an idea what the answer might be and what ever work they do is sort of a connecting the dots kinda thing. It can still be a lot of good quality work thats needed to arrive at the conclusion (which they had an idea what it was to begin with)

    The second type totally start with a clean slate and keep working until they find something and then keep working to further understand that “something” and further refine the implications what they find before calling it a solution. The great scientists typically belong to this category.

    So maybe while approaching Google as a Research / Scientific company might be right to begin the understanding process of what Google is, its also important to properly categorize them on what sort of a research company they are…Do they have an idea of what hey want to be or do they keep testing waters (not just scraping the surface but probably some deeper work) till they get somewhere

  • kathy anderson

    “Being good in business is the most fascinating kind of art. During the hippie era people put down the idea of business-they’d say, “Money is bad,” and “Working is bad.” But making money is art and working is art and good business is the best art.”
    Andy Warhol

    A heartfelt thanks for your ground-breaking Google analysis from those of us who love business!

    Google Glass, self-driving cars, blood sugar eye lenses don’t distract from the unglamorous facts about Google. Its business model (if you can call it that) is the advertising business model that began with the newspaper industry almost two centuries ago.

    It’s a long and rich history, that led to radio, followed by broadcast television, then cable tv; though touted as the new pay subscription model, the Paywall of its day, advertising on cable tv still prevailed.

    But it may all implode with Google. Already the conglomerate’s algorithms decide what content is read and seen throughout vast stretches of the world that use the Roman alphabet. Soon their algorithms will decide what content is produced and distributed. People I know in the newspaper business report that news editors and media producers spend much of their time trying to orchestrate premium placement of their stories on Google News, for the benefit of their own advertisers, second-guessing Google’s placement algorithms.)

  • poke

    I think your characterisation of Google’s business as a “regressive tax” in this podcast has more explanatory power than the “research centre” hypothesis. To be sure, Google clearly wants to be seen as a research centre, but I find taking that self-image seriously explains very little about what is anomalous about Google. Consider the following anomalies:

    1. Strange acquisitions, such as Motorola and Nest

    2. The repeated lacklustre attempts at entering the hardware market

    3. Unwillingness to discuss business models

    4. Emphasis on “blue sky” projects

    Does the “research centre” hypothesis really explain these? It explains 4 directly, but only explains 1-3 indirectly – i.e., the assertion that, seeing itself as a research centre, Google simply acquires business models doesn’t really explain why they’d acquire Nest in particular. I’d argue that Nest is the opposite of what they’d be looking for if they were looking for a new cash cow to support research. Surely they’d be looking at services that produce regular revenue or at things they could do with their cash (banking, financial services, etc) instead. It seems like we have to ascribe a large dose of irrationality or incompetence to make this fit, but a dose of irrationality and incompetence that large can explain the data directly, without the hypothesis that they’re more interested in research than business.

    Consider a counterfactual reality where Google hadn’t invested time and effort in creating its current image. What does that look like? If Google was just the company that made money from ads and selling search terms, the “regressive tax” question might be at the forefront of our concerns. Privacy might be making headlines with regard to Google and not just the NSA. The seediness of many of the businesses that contribute to Google’s revenue might be a matter of discussion. Instead, we have Google the innovator, the Willy Wonka’s Chocolate Factory of tech; fascinating enough to make a bad Hollywood movie about (The Internship). This image problem explains anomalies 3 and 4 and at least explains how they talk about 1 and 2.

    No doubt Google likes to think of its failures as experiments, but how are we to decide? Was the Nexus Q an experiment or a failure? The Chromebook Pixel? The dozens of online services shuttered over the years? The Motorola acquisition? Microsoft bought WebTV, Skype, AQuantive, etc, and has produced many duds over the course of its life. You could call those “experiments” too. Is the only difference that Larry Page thinks in terms of experiments whereas Gates and Ballmer thought in terms of business models? Isn’t Samsung the ultimate experimentalist with its dozens of models? How do we differentiate between experimenting and the more mundane business practice of throwing stuff against the wall to see what sticks? The failures and acquisitions are, to my mind, best explained by Google’s being rather more normal than their image would have us believe.

    Certainly Google’s founders have broad interests but I wonder if those interests are dictating what happens at Google or are merely being exploited in the name of the image Google projects. If Page or Brin had wanted a research centre, they could have created one. They’re worth $30 billion a piece. At some point, I think, somebody suggested putting their quasi-philanthropic interests under the Google brand since it was obvious that Google would have an image problem otherwise (there was also the issue of attracting CS talent; maybe the self-driving cars are better thought of as belonging in the same category as the free cafeterias). This is easier to reconcile in the Schmidt era, since Schmidt appears to have none of the naivety or enthusiasms of Page and Brin, being a more Machiavellian figure. Under Page, it’s perhaps best seen as cognitive dissonance.

    I’d argue that the inexplicable part of much of what Page and Brin do, isn’t that they do it, but that they do it all under the Google brand. There’s Google Ventures,, Calico, etc. The odd thing about these efforts is not that they exist but that they have the Google brand attached to them. Larry Ellison has an institute similar to Calico, but it makes no mention of Oracle. Google’s image problem explains the choice of branding.

  • Claude Hénault

    I view Nest as an iceberg calved from the Apple glacier, smaller but similar in its structure of innovation and process of product creation. I think Google strongly desires to understand what makes Apple tick, but is stymied by that company’s culture of secrecy. Nest, with its Apple-like products made by large numbers of former Apple employees, is something Google can observe as a substitute, in hopes of solving its inability to develop and manufacture physical products that are simple, appealing and widely useful. A $3.2 million dollar substitute for a peek into Apple. Now to see if Google can keep it’s hands off, and just observe for long enough to learn something that justifies the huge price.

    On the larger issue of understanding Google, I am in awe of your attempt.

    • Absolutely. Page’s infatuation with Apple is well known and self-confessed.

      Nest buys Google some of Apple’s magic. We’ll see how Fadell stomachs the two vastly different cultures in a few years. I assume it will go smoothly, but the Nest acquisition could be a seminal event in Google’s unmasking.

  • Bruce_Mc

    Summing up, we have the well publicized public side of Google as a company full of engineers who are trying to make people’s lives better. Call that Google Research.

    Then there is a more private, secretive, and probably larger side of Google full of sales and administrative people. That side (call it Google Corporate) makes money by enabling others to make people’s lives worse. For nearly every dollar Google Corporate makes, there is some person scratching their head saying, “Did I buy this? It’s not what I want.”

    It is possible to argue that a lot of Google Corporate’s money comes from legitimate advertising. However, I think that’s a lot like arguing that a liquor distributor can get by if only casual drinkers use their products; they don’t need alcoholics to stay in business.

  • Horace: this is an amazing start to the discussion, but I would encourage you to continue reading between the lines and building out the algebra equation. For example, we can see that Google paid out $280M in traffic acquisition costs (TAC) to Mozilla in 2012. That was about 2.5% of total TAC.!tQN5p

    Even after accounting for further payments to Apple, other OEMs, and telcos (such as Verizon), we can see a large gap in accounting for TAC. That means that a large share of total TAC ($10 billion in 2012, projected to be over $12 billion in 2013) was a result of their network affiliate share. That piece of the business has completely different margins from search, since there is a significant pay out to publishers. Google discloses the base rates for display ads (68%) and search ads (51%) for publishers, although big clients presumably can negotiate even more favorable revenue shares.

    • Walt French

      I wish I understood better what your point was… seems like it’s more than what jumps out to me.

      But one thought occurs: at the 68% rate that you cite, an advertiser is paying $1.47 for every dollar that a blogger gets. It’d seem that alternative ad networks, maybe even one with inferior quality of associating ads and readers, could undercut Google. As many dollars as Google makes on non-search ads, I’d think there’s a lot of money to be had from successful competition.

  • neutrino23

    I just read that Google sold Motorola to Lenovo for $2.91B. Of that only $660M is cash. $750M is in Lenovo stock and about $1.5B is a promissory note.

    Amazing. These guys really don’t care about cash. They frittered away $10B or so on this deal.

    • JaneDoe12

      No, it was a good deal because of the patents. Here’s what I read at RealMoney(dot)com from commenter “Appleman”:

      Google did not lose 9 billion on the motorola deal.

      Breaking down the admittedly messy math shows that Google didn’t exactly lose nearly $10 billion on the deal. Here are some back of the envelope calculations.

      When Google bought Motorola, the hardware maker had about $3 billion in cash on hand and nearly $1 billion in tax credits. So that brings the original deal’s effective price down to about $8.5 billion.

      Then, Google sold Motorola’s set-top box business to Arris for nearly $2.4 billion. That lowers the effective price to roughly $6.1 billion.

      Now, Google is selling Motorola Mobility — primarily the handset business, along with a few patents — for $2.9 billion. So we’re at about $3.2 billion.

      It’s worth noting a few more things. In a regulatory filing in 2012, Google disclosed that it valued Motorola’s overall “patents and developed technology” at about $5.5 billion. [1]

      1. DePorre J. Things Are Different Now. RealMoney(dot)com. January 30, 2014.

      • Tatil_S

        Motorola had liabilities, as well as cash. For example, if my memory is accurate, its accounts payable was larger than its accounts receivables. Thus, some of that $3 billion cash was already spoken for. It also lost a few billion more after the merger was announced.