The Hard Truth About Network Effects

We continue our series on critical but widely misunderstood concepts in venture capital (if you missed the post on product/market fit, it’s here).

Most entrepreneurs – and even some financial investors – confuse network effects and virality. They are not the same thing.

Startups that can build long-lasting network effects will most likely win in the end, because they create an unassailable moat. 

How can you spot network effects and build them around your business?

What Are Networks Effects?

Put simply, network effects are characterized by the fact that the user experience improves as more people join.

The quintessential product that benefitted from network effects is the phone. 

When Bell invented the telephone, he had to build two, otherwise the invention would have been useless. 

As more and more people started using the telephone, its advantages became apparent. As well as the need to have as many as your friends and acquaintances plugged in, too.

Users could call more and more people they knew, which avoided them the hassle of writing a letter, sending a telegram, or going somewhere in person. 

The Difference Between Network Effects And Virality

Another critical component of network effects is that your first users convince new ones to buy your product. 

It is this aspect that makes people confuse virality and network effects. 

Viral products also rely on word-of-mouth dynamics, but new customers you talked into getting the same product generally end up using them by themselves.

You don’t need them to join the same platform for your experience to improve.

Let’s take an example.

Calendly and Shakr Studio are two applications we’ve been using frantically over the last few weeks, and which were recommended to us by friends and acquaintances. Same as Canva, the online designing tool, and a bunch of other pieces of software.

When a new user joins Calendly, booking appointments becomes easier for that person. But does it add anything to existing users’ experience? Not really.

It’s the same for Shakr Studio, which helps its users make astounding social media videos in just a few minutes. 

These brands have existing users act as ambassadors, as we just did. These first-adopters are so engaged that they take time to explain to their friends why they should also use such tools.

It’s virality at its best and helps reach scale. But there are no network effects.

Two Case Studies: Hotmail and Dropbox

Hotmail is a poster-child for virality. Was it a beneficiary of network effects, too?

At the time, it was the media company that had created the fastest user base ever: 25 million active accounts in less than three years. That’s faster than two wunderkids of the media world, CNN and AOL.

Back in 1996, when Hotmail was created, you had to pay to use email. It’s hard to believe now, but it’s a testament to Hotmail’s disruptive idea. 

But even with a superior product (Hotmail was web-based, which allowed users to connect from anywhere), how did founders Sabeer Bhatia and Jack Smith reach so many users so fast?

After all, startup cemeteries are full of good products that nobody wanted.

The answer was, as is often the case, devilishly simple: each Hotmail message sent bore a footer telling the recipient that Hotmail’s service was not only web-based but also free, with a link to open an account. 

The exact sentence was: “Get your free email at Hotmail.”

No doubt competitor AOL was not happy about such tactics, but it worked. Hotmail, which was spelled HoTMaiL to showcase its HTML roots, was the cool kid in town.

Having users advertise your product to their networks several times a day is a clever virality ploy. In fact, this tactic became the posterchild of “viral marketing”, a terminology made famous by the VC Firm DFJ (Tim Draper widely credits himself with the Hotmail viral marketing idea.)

However, did Hotmail benefit from network effects? Not really: it was still possible to use different services, even other webmails, and interact with Hotmail users. RocketMail, which later became Yahoo! Mail, is a prime example. 

In other terms, there wasn’t any real upgrade in the user experience to have more Hotmail users, other than maybe a faster email delivery.

Dropbox: The Referral Masters

Dropbox co-founder Drew Houston is widely acclaimed for being the marketing wiz behind the startup’s success.

As was apparent in Dropbox’s S1 document when the company filed for its IPO, the growth in users was exponential between 2009 and 2016.

Let’s zoom on a portion of that curve, the period going from September 2008 to January 2010.

In just 15 months, the number of registered users went from 100,000 to a staggering 4 million. 

According to Drew Houston, most of that impressive growth was due to “word-of-mouth and viral” (in his mind, these seem to be two different things).

The referral program alone accounted for 35% of daily signups.

The referral program was quite simple and borrowed a page from PayPal’s playbook.

Instead of earning money, as in the PayPal program, both Dropbox referrer and referee were awarded storage space. It allowed the first ones to upload even more documents on Dropbox, and was sufficient for newcomers to test the solution. 

So-called two-sided incentive program are nothing new. American Express has been running one for years. The trick is to offer the same reward to both parties.

Admittedly, Dropbox took it to a whole new level: they began rewarding customers who performed specific tasks, beyond just the referral. 

They also allowed users to refer up to 32 friends (at least initially), and indicated how much space had been added each time a referee joined the platform. 

It made users crave for more space, thus pushing them to send more referrals out. The process was simplified to the extreme. 

However, we ask the question again: are there network effects at play? One advantage of having more users on Dropbox was to decrease bandwidth costs, but Dropbox could opt to keep these savings to itself. 

The customer experience was not markedly increased when more people joined Dropbox.

Facebook: The “Big Kahuna”

Now that we’ve established what network effects are not, let’s focus on what they are. 

One of the best examples around is Facebook. 

Every social network relies on network effects: every time a new user joins, it makes the experience better for everyone else (barring bad behavior). It’s true regardless of the service: think MySpace, Twitter, Quora, or Reddit.

The discussion tends to be more varied when more people join.

But what makes Facebook different is how fast it became mainstream, and how deep Zuckerberg and his team have gone into exploiting network effects.

What is striking about Facebook is just how viral it was from the very beginning.

Even before it was born.

As popularized in the 2010 movie on Facebook’s origin story, The Social Network, Zuckerberg’s first attempt was to create a website where Harvard (male) students could rate other (female) students.

Following the understandable uproar from many on campus, Zuckerberg decided to take Facemash down.

What is interesting for us here are two quotes from the 2003 article published in Harvard’s student newspaper, The Harvard Crimson.

Zuckerberg said that he was aware of the shortcomings of his site, and that he had not intended it to be seen by such a large number of students.

The Harvard Crimson (Nov. 19th, 2003)

This defense could just be a weak attempt to avoid being expelled. However, the article goes on to mention that the link to Facemash was initially sent to only a few friends but quickly spread on-campus. 

450 people used it in just the first day, and they voted an astonishing 22,000 times. The service was not only viral, but it was also very engaging. 

We can see the early buds of network effects here: not only did students want to tell their friends about Facemash because they thought it was cool and fun, but they also wanted to compare how they rated the same people.

In other terms, it made the experience better for them to have their friends on the platform, too.

** Let us just pause for a second here and emphasize again that we strongly condemn this degrading behavior towards women. Our purpose is only pedagogical, and the root facts are well known. **

From Dorm To Billions

When it launched outside of Harvard, employed two tactics that helped it exploit network effects. 

The first one is that it made it very easy for existing users to invite their friends by email. The Facebook team noticed that, on average, a person would sign up after getting seven emails from friends. They encouraged users sending friend requests out.

The second ploy was used to convince users of top campuses, which already had similar facebook websites, to leave their native app and join Facebook. The brilliant “growth hack” that Facebook used was to deploy itself on campuses close to the one they were targeting and have these students invite their friends at the target campus.

Our objective here is not to get into all the actions Facebook took to grow. It would take a whole blog post.

One key takeaway from what precedes is that network effects are embedded in usage: to exploit them, you must first understand your users’ core motivations and even belief system.

Curiosity, fear of missing out, practicality, and even stopping the nuisance are strong motivations pushing people to become users, which gradually turn into ambassadors.

Virality often relies on practicality, or “utility”, a concept at the core of modern economics.

Network effects, on the other hand, are embedded in our social fabric: we want to convince our friends to join a platform we are using because we want to re-create a sense of belonging, of being part of the same village.

At NfX they breathe, eat and sleep Network Effects

The Big Deal About Network Effects

Network effects are so prominent in VC conversations and analyses because they enable those who exploit them best to build dominating market positions. 

Especially in verticals prone to winner-takes-all dynamics.

NfX, a seed VC firm based in the Bay Area, calculates that network effects drive as much as 70 percent of value in tech. They have a network effects guide that anyone interested in this phenomenon should read and watch (see section below for additional references).

As made clear by former a16z VC Benedict Evans, the advantage with network effects is that even a better product cannot beat you once you’ve exploited them to the fullest. 

It explains why venture capitalists pay so much attention to them – and why fundraising entrepreneurs should, too.

Not many startups operate in and around network effects. But if you use the phrase while talking to VCs, you should know what you are talking about.

That’s what top founders do, as Intuit’s legendary founder Scott Cook makes abundantly clear in this NfX podcast.

Cook’s point is crystal clear: in competitive markets like today’s, the one durable source of competitive advantage is network effects.

If an accounting firm uses QuickBooks, Intuit’s star accounting software, it is likely to convince its clients to move to QuickBooks, because it makes it easier to exchange information. It’s what Cook calls a “mild network effect.”

The network effects analysis may help explain why Hotmail got beaten by Gmail, and why Dropbox is still fiercely competing against offers like Google Drive and Microsoft Azure.

But also why Facebook, for all the recent bad press on user privacy and calls for boycott, remains the king of the jungle.

Learn More

We’ve mentioned a few sources in this post that you should visit if you want to become a network effects expert. Here are some more:

Let us know if you know about other good sources here.

Understanding Network Effects and being able to spot them in startup investment opportunities is a critical skill for Aspiring VCs and active Angel Investors. 

It is why we included it in our VC Investing Track, the most thorough e-mentoring program to form professionals VCs anywhere in the world.

Related Articles

  • This is really interesting. I’m quite curious about the experience of companies that do freeware and freemium products (not in gaming or social media, but in applications). I’ve got a couple of interests:
    1. Ability to grow a user-base very fast
    2. Ability to commercialise with premium products (and recoup investment in 1!!!)
    3. Continuous value add
    Will read the academic paper on Theory of Information Product design

  • I find this very interesting. I will definitely add this to my evaluation criteria.
    I believe the network effect could be a very smart customer acquisition strategy. Although it comes at a cost, it would be interesting to know at what cost it is optimal or marginal.
    At its inception, Tesla offered a referral program/incentive link to prospective customers for a rebate and free charging at Super Chargers. This was extended to other products as time went on. Eventually, it was stopped when the organization attained a critical mass, at which point, I believe it didn’t make any economic sense to continue.

  • The example related to referrals was particularly interesting! Just because users get a temporary advantage due to other users joining the platform, there are no visible network effects.
    Another interesting case of network effects are cross-sided ones. For example, in Uber, an additional user actually worsens the experience for other users. However, they encourage more drivers to join, which in turn encourages more users to join.

    • That’s a good point Utsav. I had not thought about it this way. Also, at some point, if there are not enough users, drivers go to another platform I guess.

  • >