Today, it is easy to look at Google, Facebook and VK and imagine that these companies were always huge.
Google is everywhere. People use Google Search, YouTube, Gmail, Android, Google Maps and dozens of other services without thinking much about the enormous infrastructure required to operate them.
Facebook became one of the world's largest social networks and eventually became part of Meta, alongside Instagram, WhatsApp, Messenger and other products.
VK became one of the most important social platforms in Russia and surrounding markets.
But none of these companies started with thousands of employees, enormous data centers or billions of dollars.
They started with small teams, ordinary computers, early versions of software and an idea that had to prove itself.
The interesting part of their stories is not simply that they became successful.
It is how they became successful.
How did they get their first users?
Who paid for the servers?
Where did the early investment come from?
Why did people continue using their services?
How did advertising become the business model?
Why does a company like Google make billions from a search engine that users can access for free?
Why does Facebook need so much information about its users?
Does Google actually sell your data?
Does Facebook sell your data?
And perhaps most interestingly, why did Google's attempt to compete with Facebook — Google+ — eventually fail even though Google already had an enormous user base?
This article looks at those questions from the beginning.
Before Google: Search Engines Were Already Everywhere
Google did not invent the search engine.
Before Google became dominant, people already had several ways to search the internet.
Yahoo, AltaVista, Lycos, Excite and other services were already competing for users.
The problem was that the web was becoming enormous.
Finding useful information was increasingly difficult.
Search engines could find pages, but determining which pages were actually useful was a much harder problem.
This is where two Stanford University students, Larry Page and Sergey Brin, became interested.
They were working on a research project involving the structure of the World Wide Web.
Instead of looking only at the words on a webpage, they investigated the relationships between webpages.
A link from one page to another could be treated as a kind of recommendation.
If many important webpages linked to another page, perhaps that page was more important.
This idea eventually became associated with Google's famous PageRank system.
The important innovation was not simply "search for words."
It was attempting to determine which results deserved to appear first.
That difference became extremely important.
Google Started as a Research Project
The early Google story is a good example of how a technology company can emerge from academic research.
Page and Brin did not initially set out with a simple plan such as:
Build a search engine, sell advertisements and become one of the world's largest companies.
They were experimenting with information retrieval.
The project became increasingly useful, however, and eventually turned into a company.
Google was incorporated in 1998.
One of the earliest important investments came from Sun Microsystems co-founder Andy Bechtolsheim.
The famous story is that Bechtolsheim wrote a $100,000 check to "Google Inc." even though the company had not yet been formally incorporated.
Google later raised a much larger $25 million financing round in 1999 from major venture capital firms including Kleiner Perkins and Sequoia Capital.
This is an important lesson about startups.
The founders did not need billions of dollars on day one.
They needed enough money to build infrastructure, hire engineers and continue developing the product until the business could support itself.
Why Would Anyone Invest in Google?
An investor does not normally give a startup money simply because the founders have a clever idea.
The investor is taking a risk.
The hope is that the company will eventually become much more valuable.
Google had something particularly attractive.
It was solving a huge problem.
The internet was growing rapidly, and people needed a better way to find information.
If Google could become the place where people started their internet searches, it would have something extremely valuable:
user attention and search intent.
That second part is especially important.
Imagine somebody searches:
"cheap laptop"
That search tells you something.
The person may be interested in buying a laptop.
Now compare that with somebody randomly seeing an advertisement while reading a newspaper.
The newspaper knows very little about the person's immediate intention.
Google's search engine can see what the person is actively looking for.
This eventually became one of the foundations of Google's advertising business.
Google Search Was Free — So How Did Google Make Money?
This is where Google's business model becomes interesting.
Google Search itself could be offered to users for free.
Google did not need to charge every person five dollars a month to use the search engine.
Instead, it could provide the search service to users and make money from businesses that wanted to reach those users.
Google began advertising programs around search queries very early.
In 2000, Google introduced an advertising program for text advertisements targeted to search queries and launched AdWords as a self-service advertising system later that year. Google subsequently moved AdWords toward cost-per-click advertising, where advertisers paid when users clicked their advertisements.
This was an extremely powerful business model.
A user searches for something.
Google displays useful search results.
A business pays Google to appear in an advertising position relevant to that search.
The advertiser gets potential customers.
Google gets advertising revenue.
The user gets a free search engine.
Everyone has a reason to participate.
The Difference Between Traditional Advertising and Search Advertising
Traditional advertising generally works by showing a message to a large audience.
A television channel might show an advertisement to everyone watching a program.
A newspaper might sell a full-page advertisement.
The advertiser hopes that some of those people are potential customers.
Search advertising is different.
If someone searches:
"buy running shoes"
the advertiser already knows that this person has demonstrated some level of interest in running shoes.
This makes the advertising opportunity much more valuable.
The system doesn't necessarily need to know everything about the person.
The search itself can be a powerful signal.
That is one reason search advertising became such an important business.
Google Eventually Became Much More Than Search
Google did not remain just a search engine.
Over time it built or acquired a huge ecosystem.
And many other products.
Each service solved a different problem.
But there was another advantage.
These services created an enormous ecosystem around the Google account.
A person could search the web using Google, watch videos on YouTube, use Google Maps to navigate, store photos in Google Photos and use Gmail for communication.
That ecosystem became incredibly valuable.
Why Google's Technology Had to Become Massive
There is a common misunderstanding about companies such as Google.
People sometimes imagine that a website becomes popular and the company simply adds more powerful servers.
At very large scale, it isn't that simple.
Google has to deal with enormous amounts of data.
Search indexes have to be stored and updated.
Queries have to be processed quickly.
Systems need redundancy.
Hardware failures must not destroy the service.
Data has to be distributed across many machines.
A single computer cannot perform all of this work.
This led Google and other large technology companies to develop distributed computing systems and technologies that eventually influenced the entire technology industry.
Google's infrastructure work helped popularize ideas around large-scale distributed storage, processing and data management.
Facebook Started With a Completely Different Problem
Google was primarily solving an information problem.
Facebook was solving a social problem.
The original idea was much simpler:
What if people could create an online profile and connect with their friends?
Facebook launched in 2004.
It initially focused heavily on university communities.
That limited availability actually helped create demand.
If your university was supported and your friends were joining, you had a reason to join too.
And once your friends were there, leaving became less attractive.
This is called a network effect.
What Is a Network Effect?
A network effect happens when a service becomes more useful as more people use it.
Imagine a messaging application with only ten users.
It isn't very useful.
But if all your friends use it, suddenly it becomes extremely useful.
Social networks are particularly powerful examples.
Suppose Facebook has 10 of your friends.
You join because they are there.
Then 20 more friends join because you are there.
Then classmates join.
Then organizations.
Then businesses.
The value of the platform increases as the network grows.
This creates a cycle:
More users → more connections → more useful service → more users.
This was one of Facebook's biggest advantages.
Facebook Did Not Start With a Huge Amount of Money
The early Facebook team was operating like a startup, not like today's Meta.
As the website grew, the company needed more servers, engineers and infrastructure.
Investors began putting money into the company.
Facebook announced a $25 million financing round in April 2006 led by Greylock Partners, with Meritech Capital Partners, Accel Partners and Peter Thiel also participating. At that point Facebook said it had more than seven million users.
Later, Microsoft invested $240 million in Facebook at a reported $15 billion valuation as part of an advertising partnership.
That kind of investment gave Facebook the financial resources to expand rapidly.
But investment alone does not make a social network successful.
The company still had to keep people using it.
Facebook's Secret Weapon Was Not Just the Website
The technology mattered.
But the product design mattered just as much.
Profiles.
Friends.
Photos.
Messages.
Groups.
News Feed.
Likes.
Comments.
Sharing.
Notifications.
These features created a constant stream of reasons to return.
The platform gradually became a record of people's social lives.
That created something Google Search generally did not have:
a detailed social graph.
What Is a Social Graph?
A social graph is essentially a model of relationships between people, accounts, pages and other entities.
Imagine that Facebook knows:
- Person A is friends with Person B.
- Person A follows a sports page.
- Person B likes a particular restaurant.
- Person A frequently interacts with technology content.
- Person B watches cooking videos.
- Person A belongs to several groups.
- Person B interacts with certain creators.
The platform can use these relationships to determine what content may be interesting.
This is extremely powerful.
It also becomes extremely valuable for advertising.
Facebook Introduced Its Advertising System
Facebook officially introduced Facebook Ads in 2007.
The system allowed businesses to create pages and target advertising to specific audiences.
Facebook described its advertising model as using the social graph to connect businesses with users.
This was a major change from traditional advertising.
A business could potentially say:
I want to advertise my product to people interested in cycling.
Instead of showing the advertisement randomly to everyone, Facebook could use information about its users to help identify an appropriate audience.
This is where Facebook's enormous amount of user interaction data became commercially important.
Facebook's Early Advertising Experiments Were Not Always Successful
One of the most interesting examples is Facebook Beacon.
Beacon was launched in 2007 as part of Facebook's advertising strategy.
Participating websites could send information about users' activities back to Facebook.
The idea was to use those activities as part of Facebook's social and advertising system.
It was controversial.
Users objected to the way information about their activities could be shared.
Facebook eventually changed the system to require clearer user interaction and later shut Beacon down.
The episode is important because it demonstrates something that is still relevant today:
Collecting information and using it responsibly are two different problems.
A technology company can technically collect information but still face serious backlash if users do not understand what is happening.
Facebook Learned From Its Early Advertising Mistakes
Over time, Facebook's advertising system became much more sophisticated.
Instead of simply telling advertisers everything about individual users, Facebook could keep much of the matching process inside its own system.
For example, an advertiser might say:
Show this advertisement to people aged 25–40 who are interested in hiking.
Facebook's system could determine which users fit that audience.
The advertiser does not necessarily need a spreadsheet containing the names, phone numbers and addresses of every person who sees the advertisement.
The platform can perform the matching internally.
This distinction is extremely important when discussing the phrase:
"Facebook sells your data."
Does Facebook Sell Your Personal Data?
This subject is often oversimplified.
Meta has repeatedly stated that it does not sell people's personal information to advertisers.
Its published advertising explanation says advertisers buy advertising space and targeting capabilities rather than receiving users' personal information.
Meta's current U.S. regional privacy notice likewise says that it does not sell personal information, while explaining that it may collect and process information such as activity, device information, location-related information and inferred interests, subject to applicable policies and laws.
So saying:
"Facebook takes your data and sells your name to advertisers."
is not an accurate description of the company's stated advertising model.
The more accurate explanation is:
Facebook collects information, uses it to understand and categorize audiences, and sells advertisers the ability to reach particular audiences.
That is a very different business model.
Then What Are Advertisers Actually Buying?
Imagine a company sells hiking shoes.
The company doesn't necessarily want to know the identities of every person who sees its advertisement.
It wants potential customers.
The advertiser can define a target audience.
For example:
- People interested in hiking
- People in a particular geographic area
- A particular age range
- People who have interacted with certain types of content
- People who visited a website
- People who previously interacted with the company's business
Meta's advertising system can then determine which users are eligible to receive the advertisement.
The advertiser pays Meta.
Meta delivers the advertisement.
The advertiser receives campaign performance information.
The user sees an advertisement.
This is the basic machine behind the free social-media business model.
But Does Facebook Collect Information Outside Facebook?
This is where the story becomes more complicated.
Meta has long used information received from businesses and websites through various business tools.
These can include technologies such as Meta Pixel, SDKs, social plugins, login systems and other integrations.
Meta's privacy documentation explains that advertisers, app developers and other partners can provide information about activity on websites and apps, including information about devices, websites visited, applications used and purchases, depending on the circumstances and applicable permissions and policies.
Meta has also explained that businesses can share information about actions people take on their websites and apps, and that this information can be used to make advertising more relevant.
This is one reason online privacy is more complicated than simply asking:
"Does Facebook know what I did on Facebook?"
The larger question is:
"What information can an advertising ecosystem associate with a particular browser, device, account or household?"
Meta's Advertising System Is Basically a Huge Prediction Machine
Modern advertising is not simply:
User likes cars → show car advertisement.
It is much more sophisticated.
The system can consider many signals.
For example:
- What content you interact with
- How frequently you interact
- What topics you appear interested in
- What advertisements you previously interacted with
- Which pages or accounts you follow
- What type of device you use
- General location signals
- Activity shared by participating businesses
- Your interactions with websites and applications
- Similar patterns among large groups of users
The system then predicts which advertisement is likely to be useful or effective.
Meta has explained that its advertising auction attempts to determine which advertisement should be shown based on factors including relevance, rather than simply allowing the advertiser who offers the most money to automatically win every time.
Google Uses a Similar Principle — But Search Gives It a Unique Advantage
Google's advertising system is different from Facebook's because Google has something Facebook traditionally did not have:
search intent.
If you search:
best gaming laptop under $1000
Google has a very strong signal about what you might be interested in.
Google's own advertising documentation explains that personalized ads can use activity to make advertisements more relevant, such as using a previous search or YouTube activity as a signal.
Google can also use factors such as search terms, browser information and general location for advertising, depending on settings and circumstances.
This is why Google can make money without selling your identity to advertisers.
It can sell access to advertising opportunities.
Google Does Not Need to Tell an Advertiser Everything About You
Suppose a company wants to advertise laptop computers.
The company could tell Google:
We want people searching for gaming laptops.
Google can determine which searches are relevant.
The advertiser does not need to receive a list saying:
Muhammad searched for a gaming laptop at 10:30 PM yesterday.
Instead, Google's advertising infrastructure handles the matching.
This is an important distinction between using data and selling raw personal data.
A company can derive enormous commercial value from information without handing the underlying information directly to advertisers.
So Why Do Advertisements Sometimes Feel Like They Are Following You?
This is where personalization becomes noticeable.
Imagine that you search for a particular product.
Later you visit YouTube.
Then you visit another website containing Google advertising.
You might see advertisements related to the same product.
That can feel strange.
But it doesn't necessarily mean that someone is personally watching you.
It can be the result of automated advertising systems recognizing signals associated with your browsing or account activity and determining that a particular advertisement is relevant.
Google explicitly explains that when ad personalization is enabled, information from activity can be used to make ads more relevant.
Cookies Were a Major Part of the Old Advertising Internet
For many years, cookies played an important role in online advertising.
A cookie can allow a website or advertising system to remember information about a browser.
For example, it might remember:
- A login session
- Preferences
- Which pages were visited
- Whether an advertisement was already displayed
- Whether a user clicked something
Google explains that cookies and similar technologies can be used for advertising purposes including serving advertisements, personalization, frequency control and measuring advertising effectiveness.
But the web is changing.
Browsers increasingly restrict third-party tracking.
Mobile operating systems have introduced additional privacy controls.
Regulators have introduced new privacy requirements.
Companies have therefore been developing other approaches to advertising measurement and personalization.
The Advertising Auction
One of the most interesting parts of modern online advertising is that an advertisement can effectively compete for an opportunity to appear.
Imagine that a person opens a webpage.
Several advertisers might potentially want to show an advertisement.
The advertising system can consider things such as:
- Is this user likely to be interested?
- Is the advertiser eligible?
- How much is the advertiser willing to pay?
- How relevant is the advertisement?
- What is the predicted probability of interaction?
- Are there policy restrictions?
An automated auction can then determine which advertisement should be displayed.
This can happen extremely quickly.
The user normally sees only the final advertisement.
Behind that simple banner or video is a huge amount of software.
Why Advertising Companies Need Machine Learning
There are too many possible users, advertisements and combinations for humans to manually decide what should be displayed.
Imagine an advertising platform with millions of advertisers and billions of potential impressions.
A human could never evaluate every opportunity.
Machine learning systems can estimate probabilities.
For example:
User A is likely to interact with Advertisement X.
User B is more likely to interact with Advertisement Y.
User C is unlikely to respond to either.
The system can continuously learn from aggregate outcomes.
This is one reason AI and machine learning have become so important to advertising platforms.
Now Let's Talk About VK
Google and Facebook are familiar to almost everyone.
But another major social network has a fascinating history:
VK, originally known as VKontakte.
The name roughly translates to "in contact."
VK was founded in Russia in 2006 and became one of the country's major social networking platforms.
Its founder, Pavel Durov, later became famous worldwide for creating Telegram with his brother Nikolai.
But before Telegram, Durov's major technology story was VK.
VK Was Heavily Inspired by Facebook
When VK appeared, social networking was already becoming extremely popular.
Facebook was growing internationally.
VK adopted a number of concepts that users would recognize from Facebook:
- Profiles
- Friends
- Photos
- Messages
- Groups
- News feeds
- Social connections
This does not mean VK was simply a copy.
The platform developed its own user base, culture and technical infrastructure.
But the broader social-networking concept was already proven.
VK benefited from entering a market where people were increasingly familiar with online social networks.
Why VK Became Popular
One of VK's advantages was localization.
A social network doesn't have to defeat Facebook globally.
It can become extremely powerful by dominating particular countries or regions.
VK became particularly important in Russian-speaking markets.
It also provided features that made it attractive to its audience.
The platform emphasized communication, media sharing and communities.
Like Facebook, it benefited from network effects.
If your friends were on VK, you had a reason to join VK.
If your school or university community was there, the platform became even more useful.
VK's Early Technology Was Not Magical
This is one of the most interesting lessons for developers.
The early architecture was based on relatively conventional web technologies.
A technical history of VK's architecture describes an early stack involving Nginx, Apache, PHP, MySQL and Memcached — essentially a LAMP-style architecture with additional caching and infrastructure layers.
That should sound familiar to many developers.
You don't need an exotic programming language to build the next major website.
A startup can begin with:
Linux + web server + database + programming language + caching.
The difficult part comes later.
The Real Challenge Is Scaling
Suppose you build a social network.
At first you have 100 users.
One server might be enough.
Then you have 10,000 users.
You add more resources.
Then 1 million.
Now the architecture becomes much more complicated.
Millions of users can generate:
- Database queries
- Image uploads
- Video traffic
- Messages
- Notifications
- Friend requests
- Search requests
- Feed generation
- Authentication requests
Eventually one database server is not enough.
One web server isn't enough.
One data center isn't enough.
The engineering problem changes completely.
Facebook Had the Same Problem
Facebook's early technology was actually remarkably simple.
Meta's engineering history explains that Facebook began in 2004 as a server-rendered PHP website and was built on the Linux, Apache, MySQL and PHP stack.
That's important for beginner developers.
The technology that powered the beginning of Facebook was not some secret programming language unavailable to ordinary developers.
The advantage came from:
- Product design
- User growth
- Engineering iteration
- Infrastructure investment
- Network effects
- Data management
- Scaling expertise
The technology became much more complicated as the user base exploded.
Facebook Used Caching to Survive Growth
One particularly interesting example is Memcached.
Facebook began using Memcached in 2005 because database queries were becoming increasingly expensive.
Instead of repeatedly asking the database for information that had already been requested, Facebook could cache frequently accessed information in memory.
Meta explains that Memcached became a major part of Facebook's infrastructure and eventually handled enormous request volumes.
The basic concept is simple.
Instead of:
User → Web server → Database → Result
you can sometimes do:
User → Web server → Cache → Result
If the information is already in memory, the response can be much faster and the database receives less work.
This simple idea becomes incredibly important at scale.
Facebook Eventually Needed Its Own Infrastructure
As the platform grew, Facebook could no longer think of itself as simply a website running on rented servers.
It needed enormous amounts of:
- Computing power
- Storage
- Networking
- Cooling
- Electricity
- Data centers
- Backup systems
- Monitoring
- Security
Meta's engineering history describes the progression from a small collection of servers to globally distributed infrastructure and large data-center regions.
This is one of the hidden costs of "free" internet services.
When you use a service without paying money, somebody still has to pay for the electricity, hardware, bandwidth, engineers and physical buildings.
Advertising is one way companies recover those costs.
Facebook's Technology Became Much More Advanced
Over the years Facebook developed or adopted technologies including:
- Memcached
- Cassandra
- HBase
- Thrift
- React
- GraphQL
- Relay
- Hack
- HHVM
- TAO
- Large-scale distributed storage systems
- Massive data centers
- Machine learning infrastructure
Meta has also open-sourced many technologies developed internally.
React is probably the most famous example.
This is another interesting business strategy.
Sometimes a company can gain value by releasing technology instead of keeping everything secret.
If thousands of developers use your technology, an ecosystem forms around it.
Google and Facebook Had Different Paths to Scale
Google's core problem was:
How do we search and process the world's information quickly?
Facebook's core problem was:
How do we store and deliver an enormous social graph and constantly changing stream of content?
Both eventually required massive distributed systems.
But their workloads were different.
Google's systems needed to index and retrieve information.
Facebook's systems needed to manage relationships, content, photos, messages and personalized feeds.
This is why large technology companies often invent different internal systems even when they use similar basic technologies.
Then Google Tried to Build a Facebook Competitor
Google already had:
- Search
- Gmail
- YouTube
- Android
- Google Accounts
- Millions of users
- Huge engineering resources
- Enormous amounts of infrastructure
So why not build a social network?
Google did.
It was called:
Google+
Google+ launched in 2011.
At first, it looked like a serious competitor to Facebook.
Users could create profiles, add people to circles, share posts, upload photos and interact with communities.
Google also integrated Google+ into other products.
This created a huge opportunity.
Google already had users.
So why didn't Google+ become another Facebook?
The Problem: Having Users Is Not the Same as Having a Social Network
This is one of the biggest lessons from Google+.
Google had enormous numbers of people with Google accounts.
But that did not automatically mean those people wanted to use Google+.
A social network depends heavily on active social relationships.
People need friends.
They need conversations.
They need communities.
They need reasons to return.
Simply placing a social feature in front of millions of existing users doesn't guarantee that those users will become active members.
Facebook Had Already Built the Social Graph
By the time Google+ arrived, Facebook had already spent years building social relationships.
People had:
- Hundreds of friends
- Photos
- Messages
- Groups
- Pages
- Events
- Posts
- Comments
- Years of history
Moving to another social network wasn't easy.
If you joined Google+, your friends might still be on Facebook.
Your photos might still be on Facebook.
Your communities might still be on Facebook.
Your family might still be on Facebook.
This creates a powerful barrier to switching.
Google+ Had Another Problem: Forced Integration
Google increasingly connected Google+ to other Google services.
At various points, users encountered Google+ features while using services they had originally joined for completely different reasons.
This could help Google expose Google+ to more people.
But exposure isn't the same as engagement.
A person being automatically presented with a Google+ feature does not mean that person wants to build a social life there.
This distinction became important.
Google+ Never Achieved the Engagement Google Wanted
Google eventually conducted a review of Google+.
In October 2018, Google announced that the consumer version of Google+ would be shut down.
Google said the service had low usage and engagement.
Its review found that 90 percent of Google+ user sessions lasted less than five seconds.
That statistic says a lot.
A company can have a product that technically has millions of accounts.
But if people don't actively use it, the network isn't healthy.
Then a Security Problem Made the Situation Worse
Low engagement was not the only issue.
Google also discovered a software bug related to the Google+ API.
The problem involved third-party applications and access to certain user profile information.
Google's Project Strobe review led to the decision to shut down consumer Google+ and accelerate the shutdown after another bug was discovered.
Google later shut down the consumer version of Google+ on April 2, 2019.
So the simplified story:
"Google+ failed because Google couldn't compete with Facebook."
is incomplete.
A better explanation is:
Google+ struggled to achieve meaningful consumer engagement, faced a difficult network-effect problem against an established Facebook, and then Google discovered privacy/security issues that accelerated the decision to shut down the consumer service.
Why Facebook Survived While Google+ Failed
This is one of the most useful comparisons for startup developers.
Imagine two companies.
Company A has 100 million users who occasionally log in.
Company B has 50 million users who constantly communicate with one another.
Which network is more valuable?
Usually, Company B.
The value of a social network is not simply the number of registered accounts.
It is the number and strength of active relationships.
That's why engagement matters.
Google Had Something Facebook Didn't
Google had search intent.
If you wanted information, you went to Google.
Facebook had something Google couldn't easily manufacture:
a social habit.
People didn't simply search Facebook for information.
They went there to see what their friends were doing.
They went there to talk.
They went there because their communities were there.
That is much harder to copy.
Why VK Was Able to Compete in Its Region
VK demonstrates another important concept.
You don't necessarily need to defeat Facebook globally.
If you can build a strong network in a particular market, you can become extremely successful.
Local language matters.
Local culture matters.
Local communities matter.
Local infrastructure matters.
User habits matter.
A social network can therefore succeed by becoming deeply integrated into a particular region.
How Did These Companies Actually Pay for Everything?
This is probably one of the most important questions.
A startup has expenses immediately.
Servers cost money.
Internet connections cost money.
Employees need salaries.
Office space costs money.
Legal services cost money.
Software and hardware cost money.
So how can a website that is free to users survive before it has millions of users?
There are several possibilities.
1. Founder Money
Some startups begin with the founders' own money.
This may pay for:
- Domain names
- Servers
- Hosting
- Initial software
- Food
- Basic operating expenses
This period is sometimes called bootstrapping.
The founders try to prove the idea before taking outside investment.
2. Angel Investors
An angel investor is an individual who invests their own money into a young company.
Google's early $100,000 investment from Andy Bechtolsheim is a famous example.
Angel investment can give a startup enough money to move from a prototype into a real company.
3. Venture Capital
Venture capital firms invest in startups they believe could become extremely valuable.
The startup receives money.
The investor receives shares in the company.
If the company becomes very valuable, those shares may eventually be worth much more.
But venture capital isn't free money.
Founders usually give up part of their ownership.
They may also have investors expecting rapid growth.
4. Advertising
Once a service has enough users, advertising can become a major source of revenue.
Google's search advertising model became one of the world's most successful examples.
Facebook also built its business around advertising.
Meta itself describes advertising as the business model that allows its services to remain free.
5. Premium Services
Not every internet company needs to rely entirely on advertising.
A company can also sell:
- Premium subscriptions
- Cloud storage
- Business services
- Software
- Enterprise products
- Payment services
- Developer services
Google, Meta and many other companies have multiple revenue streams today.
The "Free Service" Is Actually an Exchange of Value
When you use Google Search without paying money, that does not mean the service has no economic value.
You provide something extremely valuable:
attention and interaction.
Google provides:
information and useful services.
Advertisers provide:
money.
Google uses that money to operate infrastructure and build products.
The same general model applies to social networks.
Users provide attention and activity.
The platform provides communication and entertainment.
Advertisers provide revenue.
The platform uses that revenue to operate the service.
Are You the Product?
People often say:
"If you don't pay for Facebook, you are the product."
It is a memorable phrase, but it is not a technically precise description.
The actual product is the social platform.
The business model is advertising.
Your activity can help the platform understand what content and advertisements may be relevant to you.
Meta itself has argued that it sells advertising space and targeting opportunities rather than people's personal information.
However, that doesn't mean data is unimportant.
Data is central to the system.
The company needs information to personalize content, improve recommendations, measure advertising and operate its services.
So a better phrase would be:
Your attention and activity create economic value inside the advertising system.
Why Advertisers Want Targeted Ads
Suppose a small business has $100 to spend.
It could put the money into a general advertisement that reaches 100,000 random people.
Maybe 1,000 of them are interested.
Or it could try to reach 5,000 people who are more likely to be interested.
The second option could produce better results.
This is why targeted advertising became so important.
It allows advertisers to spend money more efficiently.
Targeting Does Not Always Mean Knowing Your Name
This is another common misunderstanding.
An advertising system doesn't necessarily need to know:
"This is John, age 34, lives at this exact address."
It can operate on groups, identifiers, signals and predictions.
For example:
Audience: people interested in mountain biking in a particular region.
The system can decide who belongs to that audience.
The advertiser gets the ability to reach that audience.
The advertiser doesn't necessarily receive the underlying personal database.
Google and Meta Make Money From Prediction
This is perhaps the simplest way to understand modern digital advertising.
The platform has information.
The platform builds models.
The models predict what users may be interested in.
Advertisers compete for opportunities to reach relevant users.
The platform chooses advertisements.
The advertiser pays.
The platform measures the result.
Then the system learns from the result.
This creates a continuous loop:
Data → Prediction → Advertisement → User action → Measurement → Better prediction
That loop is one of the foundations of modern internet advertising.
What Happens When You Click an Advertisement?
Suppose you search for a phone.
You see an advertisement.
You click it.
You visit the store.
Maybe you buy the phone.
The advertiser may send information back to the advertising platform about the conversion, depending on its tracking and measurement setup.
The advertising platform can then estimate:
This type of advertisement worked.
This is valuable because the advertiser doesn't only want clicks.
It wants customers.
Why Advertisers Sometimes Upload Customer Lists
Advertising platforms can also support systems where businesses provide their own customer information for advertising purposes.
For example, a company might already have a list of customers.
It can use a platform's audience tools to try to reach those customers or similar audiences, subject to the platform's policies and applicable privacy rules.
Meta's Custom Audiences documentation describes mechanisms for advertisers to create and manage audiences using customer information under specific terms.
Again, the important point is that the advertising platform is acting as the matching system.
The advertiser does not necessarily receive a list of every person who was matched.
Why Privacy Became Such a Big Issue
The advertising model is powerful.
But it creates difficult questions.
How much information should companies collect?
How long should they keep it?
Should users understand exactly what is being collected?
What happens if a company is hacked?
What happens if a third-party application gets access?
What happens if information is used for purposes users did not expect?
These questions became increasingly important as technology companies grew.
Google+ provides one example.
Facebook has faced many others.
The Cambridge Analytica Scandal Changed the Conversation
Facebook also experienced one of the most famous data controversies of the social-media era involving Cambridge Analytica.
The controversy involved data obtained through an application and questions about how information from Facebook users could be used for political purposes.
The incident demonstrated an important principle:
Even if a platform does not sell personal data directly to advertisers, third-party access to platform data can still create serious privacy risks.
That is why data security and access controls are just as important as the advertising business model.
Data Is Valuable Even If It Is Never Sold
Imagine a company owns a huge database but never sells the database itself.
That database can still be extremely valuable.
It can be used to:
- Improve recommendations
- Personalize advertising
- Detect fraud
- Improve search
- Train or evaluate systems
- Measure user behavior
- Improve products
- Predict demand
- Identify spam
- Detect abusive activity
The economic value comes from what the company can do with the information.
It does not necessarily require selling the raw information.
Google, Facebook and VK Also Needed Something Else: Trust
Technology alone cannot create a huge platform.
Users need to trust it enough to continue using it.
If a search engine constantly gives bad results, people leave.
If a social network becomes too annoying, people leave.
If users believe their accounts are unsafe, they may leave.
If advertisers don't get results, they stop paying.
If developers cannot build on the platform, the ecosystem weakens.
A successful internet company therefore has several different groups to satisfy:
Users
Advertisers
Developers
Investors
Employees
Regulators
Keeping all of them reasonably satisfied is extremely difficult.
Why Google Won Search
Google's biggest advantage was not simply having a search engine.
It created a habit.
People learned:
If I need information, I will Google it.
That habit became incredibly valuable.
Once a service becomes the default answer to a common problem, competitors have a difficult challenge.
The competitor isn't only competing against technology.
It is competing against habit.
Why Facebook Won Social Networking
Facebook similarly created a habit:
If I want to see what my friends are doing, I will open Facebook.
Then the service became part of people's everyday communication.
That made it extremely difficult for a new competitor to replace it.
Google+ had Google's engineering resources.
But Facebook had years of social relationships.
And social relationships are difficult to copy.
Why Google+ Is Such an Important Failure
Google+ is worth studying because it demonstrates that money and technology are not enough.
Google had:
- Engineers
- Servers
- Billions of users across other products
- Search
- Gmail
- Android
- YouTube
- Advertising
- A powerful brand
Yet Google+ still failed as a consumer social network.
The platform eventually suffered from low engagement and privacy/security problems, and Google shut down the consumer version in April 2019.
For startup founders, this is an extremely useful lesson.
You cannot simply buy your way into a network effect.
What New Startups Can Learn From These Companies
The stories of Google, Facebook and VK contain several lessons.
Start With a Real Problem
Google improved the problem of finding information.
Facebook improved the problem of connecting with people.
A successful startup should solve something users actually care about.
Start Small
None of these companies started at global scale.
The first version can be simple.
The important thing is getting real users.
Use Ordinary Technology at First
Facebook started with PHP and MySQL.
VK used conventional web technologies as it grew.
You don't necessarily need a futuristic technology stack to launch a useful product.
Architecture can evolve.
Scale When Necessary
Don't build a billion-user architecture for ten users.
Build something that works.
Then measure where it breaks.
Then improve that part.
This is one of the most practical lessons from Facebook's engineering history.
Network Effects Are Powerful
A service becomes difficult to replace when users have relationships and valuable content inside it.
That's why social networks can become extremely difficult to compete with.
Free Doesn't Mean Without a Business Model
A free product can make money through:
- Advertising
- Subscriptions
- Business services
- Transactions
- Cloud services
- Premium features
The important question is:
Who pays for the service?
The Technology Behind the Internet Giants Is Only Half the Story
It is tempting for developers to look at Facebook and think:
"They became successful because they used advanced technology."
But the order was mostly the opposite.
First came the product.
Then users arrived.
Then the company had to solve increasingly difficult engineering problems.
The technology evolved because the scale demanded it.
Facebook's engineering history is a perfect example.
The company began with a simple PHP application and later developed massive distributed infrastructure, caching systems, data stores, ranking systems and globally distributed data centers.
The same general pattern can be seen throughout the technology industry.
What Would Happen If Google or Facebook Stopped Advertising?
This thought experiment explains the business model.
Imagine Google Search remained free but Google stopped selling advertisements.
Google would still need:
- Data centers
- Electricity
- Engineers
- Networking
- Storage
- Security
- Research
- Hardware
- Customer support
The costs would continue.
Without advertising revenue, Google would need another way to pay those costs.
It could charge users.
It could rely more heavily on subscriptions.
It could sell enterprise services.
It could use other business models.
But advertising is extremely powerful because billions of people can use the service without directly paying Google.
The same principle applies to social networks.
Is Targeted Advertising Always Bad?
Not necessarily.
Targeted advertising can have legitimate benefits.
A small business may not be able to afford a television campaign.
But it might be able to spend a small amount targeting people who are more likely to need its product.
A user may also prefer seeing advertisements for products that are relevant instead of completely random advertisements.
The problem is not simply targeting.
The difficult questions involve:
- Transparency
- Consent
- Data security
- Sensitive information
- Data retention
- Third-party access
- Manipulation
- User control
The technology itself is neutral.
How it is designed and used matters.
Why Users Should Care About Their Privacy
You don't need to stop using Google, Facebook or every other online service to care about privacy.
A better approach is to understand what is happening.
Know what your accounts contain.
Review privacy settings.
Understand advertising controls.
Use strong passwords.
Enable two-factor authentication.
Be careful with third-party applications.
Don't give unnecessary permissions.
Understand that "free" services often have an economic model behind them.
Privacy is easier to protect when you understand the system.
The Bigger Picture
Google, Facebook and VK did not become successful because of one secret algorithm.
They combined several things:
Good timing
The internet was expanding rapidly.
Useful products
Each company solved a problem people cared about.
Network effects
More users made the services more useful.
Investment
Outside capital helped fund expansion.
Advertising
Advertising provided a scalable revenue model.
Engineering
The companies continuously rebuilt their infrastructure as they grew.
Data
User activity helped improve products and advertising.
Habit
Eventually people simply became accustomed to using the services.
That last point is easy to underestimate.
A technology product becomes extremely powerful when people stop thinking about alternatives.
From a Student Project to a Global Infrastructure Company
Perhaps the most interesting part of these stories is how ordinary the beginnings can look.
A university project.
A small website.
A few servers.
A handful of developers.
Some borrowed or invested money.
A basic programming language.
A database.
Then the number of users increases.
The system breaks.
The developers fix it.
The number increases again.
The company raises money.
More engineers join.
The architecture becomes more complicated.
Eventually the company owns enormous data centers and employs thousands of people.
This transformation is one of the defining stories of the internet.
Final Thoughts
Google, Facebook and VK all followed different paths, but their stories share a surprisingly similar foundation.
They started with relatively simple ideas.
Google wanted to make information easier to find.
Facebook wanted to make it easier for people to connect.
VK built a social platform that became deeply popular in its target region.
None of them started with the infrastructure they have today.
They grew into it.
Investors provided capital when the companies demonstrated potential.
Advertising eventually became a major source of revenue.
Users provided the attention and activity that made the services valuable.
Engineers built increasingly sophisticated systems to handle the resulting scale.
And the companies accumulated enormous amounts of information because modern internet services depend heavily on understanding how people use them.
But the most important lesson may be the story of Google+.
Google had money.
Google had engineers.
Google had users.
Google had infrastructure.
Yet Google+ still failed to become a major consumer social network.
Why?
Because a successful social network is not simply a website.
It is a community.
It is a collection of relationships.
It is a habit.
It is an ecosystem.
And once millions of people have built their social lives around one platform, another company cannot easily reproduce that network simply by building similar software.
The advertising side of these companies is equally interesting.
Google and Meta do not need to hand advertisers a copy of your personal profile in order to make money from information about users.
Instead, their systems can use data, signals and machine-learning models to determine which advertisements are likely to be relevant and then allow advertisers to pay for access to those advertising opportunities.
Google has the enormous advantage of search intent.
Meta has the enormous advantage of social relationships and activity across its platforms.
Both systems ultimately turn attention and information into advertising revenue.
That is one of the fundamental business models that built the modern internet.
So the next time you search for something on Google, scroll through Facebook or watch a video recommended by an algorithm, remember that what looks like a simple webpage or mobile application is actually the visible surface of a massive technical and economic system.
Behind the screen are databases, caching systems, machine-learning models, advertising auctions, distributed servers, data centers, networks, engineers and billions of pieces of information being processed every day.
And that is perhaps the biggest lesson from the history of these companies:
The internet giants were not built by one technology. They were built by combining technology, users, money, data, infrastructure and human behavior at enormous scale.
Frequently Asked Questions
Did Google invent the search engine?
No. Search engines existed before Google. Google's major contribution was developing a highly effective approach to ranking web pages and building a search service that became exceptionally useful and scalable.
How did Google get its first funding?
One famous early investment was a $100,000 contribution from Sun Microsystems co-founder Andy Bechtolsheim in 1998. Google later raised a $25 million financing round in 1999.
How did Facebook make money in the beginning?
Facebook eventually developed advertising into its primary business model. Its advertising system was formally introduced in 2007, allowing businesses to target audiences on the platform.
Does Facebook sell user data?
Meta says it does not sell people's personal information to advertisers. Instead, advertisers pay to display advertisements to audiences selected using Meta's advertising systems.
Does Google sell my personal data?
Google's advertising model does not require it to sell your personal identity information to advertisers. Google uses information and signals to personalize and deliver advertising according to its policies and user settings.
Why did Google+ shut down?
Google announced that the consumer version of Google+ would shut down primarily because of very low consumer usage and engagement. Google also disclosed security/privacy issues involving the Google+ API, which accelerated the shutdown.
When did Google+ officially shut down?
The consumer version of Google+ was shut down on April 2, 2019.
What is VK?
VK, originally known as VKontakte, is a major Russian social-networking platform founded in 2006. It became particularly popular in Russian-speaking markets.
What technology did Facebook originally use?
Facebook originally used a relatively conventional LAMP-style stack based around Linux, Apache, MySQL and PHP. As the platform grew, it developed extensive additional infrastructure and technologies.
Did Facebook really use PHP?
Yes. Facebook began as a PHP website. Meta's own engineering documentation describes Facebook.com as a simple server-rendered PHP website when it launched in 2004.
What is Memcached?
Memcached is a caching system that stores frequently accessed information in memory. Facebook began using it in 2005 to reduce the load on its databases and improve performance.
Why are Google and Facebook so difficult to compete with?
Their biggest advantages are not just software. They have enormous user bases, network effects, data, infrastructure, brand recognition, developer ecosystems and established user habits.
Can a small developer build the next Facebook?
Technically, you can build a social network.
The difficult part is not creating profiles and a news feed.
The difficult part is getting millions of people to use it at the same time and creating a network that becomes more valuable as it grows.
That is the real challenge.
Comments
Post a Comment