Facebook is social media. Google is Search!
Last time, I said Facebook has proven the theory of platform business models right (it states that exponential growth would be possible due to the indirect network effects). Google has proven it even more (and a few years prior to Facebook).
Have you wondered how today’s world (ok, let’s say the internet) would look like without Google? The early homepages for millions of internet users were portal (or tabloid)-type pages, where search was only one of the features often buried among a lot of clutter. Other prominent pages also adopted the portal-style design (here Yahoo 1994, and 1997), or just look at the current AOL.com.
When Google emerged, their pages looked very bare. They focussed only on one value proposition: search results and reduced all search and transaction costs involved (it is said that this was due to the fact that they did not know enough about webpage design but seems it was to their advantage).
Portal-type pages are gatekeepers that channel traffic in desired directions through “internet traffic lanes” consisting of a system of links. A good search engine that is not prone to manipulation or bias can still be seen as a gatekeeper but is much less so than a portal-page.
“Our mission is to organise the world’s information and make it universally accessible and useful.” Google
Our focus will largely be on Google Search and the many tools closely related to it. The main exclusions are Google Cloud (see my article on Amazon Web Services to understand Cloud) and the related Internet of Things gadgets (see my Amazon Alexa personal assistant and IoT article on the topic).
Revenue by type (segment)
Though Google has a lot of products that generate revenue, they only split those out that start making some notable revenues (in some cases, they delay reporting on individual products to keep it as a trade secret). The bar for reporting revenues separately is high.
Google’s total revenues as of December 31 2019 are $161.9b. Major revenues are as shown below. Note, that the parent company is Alphabet but as you can see below, most revenues are generated by Google (who also own YouTube).
|Google Search & other ads||$98,115 m|
|YouTube ads||$15,149 m|
|Google Network Members’ properties||$21,547 m|
|Google Cloud||$8,918 m|
|Google other *||$17,014 m|
|Other Bets revenues||$659 m|
* Includes YouTube non-advertising revenues
Revenue by geography
Revenue by geography is determined by the address of the (advertising) customer. Geographically, Google (Alphabet) reports user metrics and revenue for:
- Europe, Middle East, Africa (EMEA)
- Asia-Pacific (APAC)
- Other Americas
Platform business models need to think about the value proposition to all sides of the multi-sided platform. You hear me say always that platform business models are about positive network effects as well as search and transaction cost reduction. Google is certainly the best example of the concept of search cost (as in search effort) reduction.
On a high level, the value proposition to users are (based on their own words):
- Organise (the world’s) information: through crawling and indexing and then matching with search queries
- Make information accessible: this includes making accessible things: search news archives, patents, academic journals, billions of images and millions of books (through scanning)
- Accessing information in most life situations: including on the go on their mobile (mobile-friendly pages will rank higher)
- Fast pace of search (and search result pages): millions of search results found within milliseconds. Their Chrome browser and tools are typically the fastest among peers including on mobile. Their desire for pace is uncompromising (compare MS tools that never get faster even if the underlying HW has exponentially gotten faster according to the old saying: “What Intel giveth, Microsoft taketh”)
- Clarity & simplicity: What may have started due to lack of HTML skills has become one of their biggest value proposition, clarity and simplicity of the start page and search results. Laser focus on usefulness
- User search experience: Fast, simple, clear, best ranking of search results without preference for ads which are marked as Ads
- Focus, esp no distraction through ads: Many search results don’t display any ads simply because no relevant ads are in the inventory. Ad quality scores aim to ensure display ads are relevant to the search
- Continued iteration and improvement: on various levels (see key activities)
- Democracy: “We assess the importance of every web page using more than 200 signals and a variety of techniques, including our patented PageRank™ algorithm, which analyzes which sites have been “voted” to be the best sources of information by other pages across the web” Google
Similar value propositions
There are over 39,000 businesses and startups on Crunchbase that fully or in part rely on Social Media functionality or have such an app. This easily exceeds the "games / gaming" category by 2x.
You could add to this related types of apps, such as user-generated content / media sharing, messaging, chat, search, vertical search, location-based search, review platforms among others and you'd probably end up at 100k+ businesses. Here's an article that shows 95 social media apps as the tip of the iceberg.
I have covered these types of platforms / apps in several long articles: Google, Facebook, Twitter, Snapchat, Pinterest, Groupon, Yelp. It might be the most extensive resources on the internet on the business model of these companies / apps. And I am being contacted by many who feel that understanding these types of apps, platforms and functionality can drive the growth of their business:
“The content you have created is super valuable and really makes it clear how those large players are earning their money” Malin Andersson, Co-Founder Worldsmarathons.com (Purchaser Super Bundle).
Business model first: I appreciate we all love to design and develop. But the view that I have expressed many times is that the business model matters most. Developing randomly can cost you years. There are typical elements that you will see across all of these apps. But each of these businesses implements them somewhat differently.
Comparing the differences, similarities and results has been crucial in understanding what drives success (and what doesn't). I have put all of this together in our social media / search super bundle. I am also providing learning aids that will help you to focus on the key parts in a way that can get you through it in a few hours time. Learn more here....
Google is a multi-sided platform with a large ecosystem of partners. Listed below are the “sides” of the multi-sided platform followed by the partners. The main focus is on the Google Search platform.
Sides of the multi-sided platform
- Users: The majority of users are searching and consuming content. A lot of users also use some or many of Google’s tools, such as GMail (1b users), Maps, the Android OS (2.5b users), etc
- Websites/blogs: Website owners create/curate content that Google crawls, indexes and presents as search results. Some sites stand out, such as Wikipedia who often get a featured box if Google thinks the user is looking for knowledge-type information. Other search-intents also lead to a tailored presentation of search results with some sort of content curation from relevant sites
- Brands: Large brands as such may not stand out as they do on social media. Their search results may appear higher in the search result if relevant, but there is no logo or anything. Exceptions are of course search-specific results presentation, think of shopping results if this anticipated as the search intent
- Businesses: Businesses – especially if registered in Google My Business – will be presented in a map
- Advertisers: Google presents ads up the top, provided it is considered relevant enough for the search. Ads on Google remain native, that also means they are very simple – often a simple text snippet. Advertisers can also add so-called ad extensions/call-outs (these can be additional links to relevant subpages, their phone number or other relevant information and more)
- Influencers, VIPs: influencers, VIPs, etc don’t as such get a lot of additional featuring. In some cases, you may find Tweetboxes being shown within the search results if this seems relevant to the search
- Media/News: News results also play an important role in the search platform. This becomes clear if you look at the search result autocomplete feature which can often feature keywords of breaking news. News has its own section with results often from the big media outlets or online magazines
- Creators / publishers: YouTube has a lot of creators who can make money from ads. On the Search platform, videos have their own section (it does not only show YouTube video in those results)
I see the above as relevant enough to call them a separate side. You can call them a key partner but I think a “side” of a multi-sided platform is actually a bit more than that.
Here are what we would more ordinarily understand as key partners:
- Customer service partners: Google partners provide services for the various aspects of marketing, some examples are (there is certainly an overlap with SEO partners):
- Search Advertising
- Video Advertising
- Display Advertising
- Shopping Advertising
- SEO partners: There is a large industry of Search Engine Optimisation partners that support website creators to attract more and better traffic. The industry is poised to reach $80b in 2020
- Developers: Google is highly revered among developers. They offer many ways for developers to get in touch:
- Thousands of communities directed towards students, professionals, entrepreneurs, women. Examples are: local Google Developer Group (GDG) chapters, Developer Student Clubs (also hundreds of them globally, Launchpad: global accelerator programs
- Groups organised around technologies (or consoles): Google API, Google Cloud, Google Play, Firebase, Chrome and more
- Roughly a dozen events per years centred around a technology
- Research partners: Google has an entire research portal with about two dozen research areas that they are interested in (which is a lot). A few examples are:
- Algorithms & theory
- Speech processing and natural language processing
- Quantum computing
- Security, privacy and abuse prevention
- Health and bioscience
- And a lot more
- Quality raters: Google has thousands of trained globally distributed third-party Search Quality Raters who help to evaluate the results of improvements to search algorithm upgrades using an extensive quality assessment guide (pdf here)
- News partners: Google has started the Google News Initiative, a network of partners to collaborate on digital news, industry-wide challenges (this includes fake news and fact-checking) and innovative programs. Google states: “Google cares deeply about journalism. We believe in spreading knowledge to make life better for everyone. It’s at the heart of Google’s mission. It’s the mission of publishers and journalists. Put simply, our futures are tied together.” They are also partnering with the International Fact Checking Network
The most important activities for platform business models are to enhance positive network effects and reduce negative ones. Google has significant positive network effects for its users but needs to constantly manage and improve these. Positive network effects come from good matching-making of search queries with search results.
Negative externalities can come from poor search results (e.g. due to inaccurate algorithms and manipulation).
- Crawl & index (organise information): The search process starts well before any users types a search query and it never stops. Google’s algorithms constantly crawl the internet by following links. Each page is being indexed for the words it contains (like the index of a book but for all words), analysed for numerous (>200) signals and organised in a search index. While far more complex, we can still imagine this a little bit like a good old library (if you still remember what that is)
- Search & match: The key activity that Google performs is to match what it considers to be the best result to any given search. This is a big feat due to the billions of pages (estimated ~8b, on around 1.6-1.9b web sites with 400m active ones, FYI: here is also the first web page ever!) available online and trillions of searches per year and the permutations across both. It happens through a “whole bunch of algorithms” (Google). Have a look at the infographics that explain the 5 steps involved (on a high level) at the end of this section
- Rank & present: As mentioned previously, Google is working on the best ways to present the search results. For starters, it ranks the results with what it considers to be the most relevant result at the top (here the original work of the founders on PageRank). As we have seen, they are now delivering different presentation elements that align with the search intent. I believe there are still many opportunities on the presentation side. It will also help them to achieve higher user engagement results (however, pushing this too far could put them at odds with website owners who at the end want the right users to come to their sites rather than reading a small snippet of their content on the search results page). You have seen some of the elements in the key activities section
- Organise knowledge (build knowledge graph): Google takes things beyond indexing by organising information and knowledge in graphs. This helps them in the presentation of results. The knowledge panel (a type of summary view) that (currently) is presented on the right hand side (when using a large screen) brings together information from various sources around the user’s search. But the ambition is much broader “This is a critical first step towards building the next generation of search, which taps into the collective intelligence of the web and understands the world a bit more like people do.”
- Improve value proposition: Ongoing innovation on their core value proposition is essential for any platform. For example, the first step of the search process is to understand what the user actually means. They state that it took them 5 years to develop search intent understanding which “involves steps as seemingly simple as interpreting spelling mistakes, and extends to trying to understand the type of query you’ve entered by applying some of the latest research on natural language understanding.” Their conclusion is that it significantly improved results in over 30% of searches. It has essentially helped to try to decipher what strings of words should be looked up in the index. Aside from the big innovations such as natural language understanding, the knowledge graph and similarly large improvements, there are of course the ongoing smaller improvements
- Build ecosystems of tools: Google develops ecosystems of tools. These can align more or less with the mission of organising the world’s knowledge. Many of their tools are covering devices and formats that people use for search or more broadly to access information. Android OS (in the supported GMS version), pre-installs Google’s core tools including the Chrome browser as the default browser. Or take Youtube which has the value proposition to provide knowledge in video format (rather than text plus image)
- Incorporate relevant ads: Ads can exert significant negative externalities when done poorly. Google has never allowed “loud” or overly interruptive ads from the very beginning. Initially, only text ads were allowed. But even therein they never allowed things like fully capitalised words, exclamation marks (“FREE!!!”), etc. Of course, they have long moved on from text-only ads But they have pretty clear rules around all types of ads, including display ads. If you compare Google or GDN-sourced display ads, you will certainly notice that there is significantly less clickbait-type stuff (compare to Snapchat who we have covered recently)
- Prevent low-quality search results, this includes deliberate misinformation, spam, clickbait and more. It is very important and I am covering it in more depth under the customer relationship section. Google manages this in the “Quality of content” step in their search algorithms
- Engaging users is mission-critical for any platform business. But this can mean a lot of things. Social media apps engage users through the news feed, another way are active prods such as notifications, messages from friends, active statuses of friends, etc. With Google things are a bit different. They provide tools for free. The Chrome browser opens to Google Search. It integrates the search function in the URL field. This was not the case in the early days of the internet. They are giving away the Android OS, GMail, Maps, etc for free. Another way is to make the search results page more engaging through the various rich results that we talked about
- Influence content creation on the web: Google has quite an influence over certain elements of content creation that is largely hidden to users. For many websites, Google is the largest source of traffic. And website owners follow the rules/recommendations on how to structure webpages. Good Search Engine Optimisation (“white hat” SEO) means to structure the site and content in a way that it increases the quantity and quality of traffic to the site (Google has their own SEO guides)
- Innovate: More broadly, innovation involves:
- Major innovations on existing core products, tools or algorithms such as the introduction of the knowledge graph. Important examples are “Hummingbird”, “Panda”, Penguin”
- The moon shots – they call out YouTube, Android and Chrome; add Waymo (the autonomous cars) and a lot of other projects underway
- Acquisitions and integration thereof into their portfolio
- Long-term research fields as shown before
- Continuous improvements on apps, websites, tools, hardware, algorithms, etc
- And more
- Marketing: sales and marketing make 12% of their revenues and thus a large item in their cost structure (and similar to Facebook with 14%)
Key Resources / Assets
The master resource (or asset) of any platform are its network effects. It is the resource/asset that needs to be built and nurtured. Google doesn’t share a lot of data on their user and business usage statistics. Thus, we will focus on this section in a qualitative manner.
- (Daily) active users:
- Active users: there are varying degrees of users. It depends on user habits. Some people’s browser directly opens to Google, others still use portal pages, others have (well-organised) sets of bookmarks to go directly to the pages they frequently use
- Chrome browser (and Chrome OS) users: those that use Google’s Chrome browser “share” typically more data with Google
- Android OS users: “share” even more data as most people have their phones with them 24/7 (depending on privacy settings)
- Users of their products: esp GMail, YouTube users and also lesser used tools
- Data assets: are Google’s currency and at the heart of everything they do. They source the data through crawling, their tools and many other ways:
- User Data: Google collects a lot of user data (no surprise here). The more tools and services people use, the more data they share
- Google Analytics tracking code: Website owners install Google’s tracking code on their site which helps owners to get traffic data. This, in turn, also provides all the website’s data to Google. Many millions of sites use this code which is an enormous data asset
- Algorithms: Google’s algorithms made them big, in particular their PageRank algorithm (which obviously has dramatically evolved over time and takes over 200 signals into account. By now they have many algorithms that work together to provide the different value propositions
- Intellectual property: Google has some 58,000 patents (23,000 granted currently but that is in part a function of the pace of the assessment process). This is astounding even for their size but relatively young age. We can clearly see the ambition behind these numbers
- The Google brand is synonymous with search. The term “Googling” has become synonymous with searching for something on the internet and has found its way into dictionaries. In the renowned Interbrand brand ranking, they are ranking second (to only Apple)
- Relationship to advertisers ranging from small/medium size businesses to large multinational brands. Six years ago, Google had 4m advertisers in 2014 and this number will be much larger by now
- Their technology, theirs Apps and all sort of other digital assets
- The website ranks 1st in internet traffic
- The Android OS has 2.5b Monthly Active Users (MAUs) as of May 2019
- G Suite (i.e. their office tools, such as GMail, GDrive, etc) has 2b Monthly Active Users (MAUs) as of March 2020
- The Google App: 5b+ downloads (number includes (re-)downloads after version upgrades)
- Advertising infrastructure: We are covering this in the advanced resources but it is a tremendous asset with several interconnecting platforms
(We are looking at this from the lens of the underlying needs, how they are served and the things that affect customer relationships. There is a natural overlap with the value proposition)
If you are a regular follower of my blog, you know that this is one of the most important categories for platform business models (in the long run). You can check out my previous articles on Facebook, Twitter, Snapchat on more introductory details and facets on this topic.
Imagine a web page that is deceptive but ranks really well on Google. It could get a lot of traffic on their pages and do bad stuff (e.g. misrepresent what they sell or just have a ton of low-quality high-ranking content and tons of ads in-between). And now imagine that this happened a lot because some users had figured out how to rank high on Google. Well, the aim of “black hat SEO” was just that: rank high without adhering to the intention of Google’s guides. There are/were many tactics. Google’s reputation and customer relationships depend a great deal on their ability to keep this at bay.
And then there are the difficult cases where improving the value proposition may come at the cost deteriorating customer relationships. Many platforms face this problem when it comes to personalisation vs privacy.
Personalisation vs privacy
Let’s look at this important example. Reduction of search and transaction costs are some of the biggest value propositions for platform businesses. While a lot of factors can fall under this, in the case of Google search results, the term “search cost” can be taken literally. But as a regular reader, you also know that I am always pointing out the post-transaction results. These are often negative effects that users could incur after a transaction is finished. Violation of privacy needs (which are highly subjective) can be an important post-transaction cost.
Basically, all channels, esp to users, are through the websites and apps. Most transactions are automated and self-serving. But the list can be very long, depending, of course, how many Google tools and services one uses.
Google seems to be using the same architecture and principles for user interaction across their ~200 products. This speaks for their unmatched software architectural prowess.
- Key interaction channels are:
- The URL field in most (if not all) browsers directly triggers a search unless a valid URL has been entered. This was not the case in the early days
- Being the default search engine: For most people and browsers, Google is the default search engine. In combination with the above (the URL field), it triggers trillions of searches annually
- Android OS, Chrome OS / browser, Maps, etc: Google gives users many free tools which trigger search activities (and data capture) or have Google Search set as default (e.g. Android/Chrome OS)
- Notifications: A number of Google tools have notifications, this includes GMail, News, their communication tools, the Search app among others. Many notifications are also sent to the user’s GMail account. Some notifications are based on subscription-notification features – most prominently YouTube
- Recommendations: A number of Google’s tools have recommendations, such as YouTube, PlayStore, PlayMusic and other in particular media related tools/apps
- And more – you know where.
Who is using search engines and what are the details of the customer segments? The answer to the first part is easy: almost every internet user (though not everybody is one – more in a second). The question regarding the segments is a bit more difficult. Unlike social media, Google Search is only poised in limited ways to find out the demographic aspects (language and location: yes; other aspects such as age, gender, education, income and interests: depending on user data share, online behaviours and privacy settings).
Determination of demographic profiles via search data is less desirable because it (a) is less reliable and (b) fraud with privacy concerns compared to data that users volunteer on social media.
There are more internet and search engine users among: younger, more educated and higher income users. These are segments for the US but also likely to be true more widely.
The sensitivity of daily active users to income and education were particularly strong. However, this was also likely an outcome of the innovation adoption curve, with the gap likely having closed somewhat by now.
Google’s cost structure consists of:
- Cost of revenues
- R&D expenses
- Sales and marketing
- General and administrative
The diagram show the cost structure on a high level as percentage of revenue. You can see that costs have inched up over the years (well, you would have heard this many times over).