Mohamed Zahid, Lead Analyst at HootSuite

Mohamed Zahid shared:
Network Effects, Returns to Scale, and Barriers to Entry: An elaboration on Nir Eyal and Sangeet Paul Choudary's recent TechCrunch article


Read what Mohamed Zahid is reading for work

Quibb lets you share what you're reading for work. Use Quibb to share news about your industry, discuss what matters, and see what colleagues are reading.

Our mission is to connect professionals over business news and informed commentary — targeting every industry, profession, and geography.

Quibb was launched in 2013 and is based in San Francisco, California.

Apply with Twitter or Google


Likes Comments
Advisor at Quibb

I think there's still enormous network effects in social products because I tend to view them as a series of feedback loops, which means that network scale equals defensibility. I think that's the end point to all of this, although initially single-user utility is super valuable- I'll elaborate on this some more below.

The most important loop is the one that goes from "content creation to social feedback" - this basically makes people excited to create content because they know they will get great, authentic feedback as a result of their original action. What creates a high-quality loop? Well, you need every user to have a big network of people whose social feedback they care about- either folks from real-life or people they respect. In my focus on quality, you might consider me as being sympathetic to Path's point of view.

However, early on in the life of a social product you often don't have a dense enough network that whenever you post, people will respond. Right now, for example, people on Quibb with more than X followers are likely to be guaranteed some kind of social feedback, but not everyone has X followers.

As a result, early on in a user's experience, they need to have some single-user utility to the product- this is the part of Nir Eyal's essay that I do agree with. You have some single-user utility, for Quibb that might be to find interesting articles, and once you pick up enough followers then eventually you can get the main feedback loop around content creation. On the downside, that means that for people who don't find the single-user value prop interesting, they may churn relatively early.

Editor-in-Chief at Quibb

Alongside easier switching costs, there's also lower 'trying' costs. It's now easier to try a new website than before - just sign-in with Facebook and a lot of the previous experience of putting in your email & password, inviting friends, etc. drastically reduces friction. Because of that, building a network with 100M consumers is now much easier than it's ever been - but getting them to hang out is still just as hard.
All of that said, I'm sure Nir Eyal and Sangeet Paul Choudary would agree that if you CAN build a network of 100s of millions - that's still insanely valuable, and has huge self-reinfocing effects wrt brand, word-of-mouth, virality, etc.

Sandi MacPherson: You're right. As Mohamed Zahid also points out in his excellent analysis, we're not saying that the network effect is any less than it used to be and we are not talking about standalone value required to initially seed a network before critical mass hits in (although that is super-important). We're only talking about how formidable network effects used to be as a barrier to entry at one point and how single sign on and cross-pollinating the social graph can help new networks reach critical mass faster, especially those that aren't direct competitors (not the same product) but are indirect competitors in that they compete for the same finite user attention or time. (E.g. Pinterest and Instagram arguable take away time and attention that would otherwise be devoted to Facebook).

To Mohamed's point in the above analysis, we did club several things together but the common thread across those was the "competitive moat' and "switching cost" factor. Sunk costs are a pain that prevent you from switching, network effects are a gain that prevent you from switching. The way I see it, there are 4 types of switching costs: Sunk costs (pain), platform-owner created incentives (e.g. loyalty programs), network effects and enhanced user experience through stored value.

Hence, our point is not that network effects are any less important for building engagement on a product, but that they are not as strong as they used to be to prevent competition from coming in and taking users away. It's a more leveled playing field.

Excellent commentary. The following summary couldn't have been better "This means that network effects as a value prop have remained unchanged, but network effects as a barrier to entry have been reduced in their power"

Reply · Tweet ·
Seed Investor at Scout Ventures

Hey Mohamed, thanks for asking me to comment.

I learned this stuff in grad school from the people who pioneered it (Katz, Shapiro, their colleagues) so my pov is somewhat academic. Three comments:

- I feel that the original Techcrunch article focuses a lot on lock-in and very little on the most important / core value prop of network effects, which is that the value of the network increases as more people join it. The biggest networks are the most valuable and that's what keeps users on them. Lock-in is not core to the story.

- Re: The "cost of trying" that Sandi points out-- Multi-homing is when users can be on multiple competing networks. Costs (aka homing costs) are low to be on these online networks and as long as that's so, there won't be much pressure on the various networks to consolidate. Even so, to my point above, Facebook is by far the biggest and thus, most valuable network. So the key to all the players nipping at Facebook's heels is to figure out how to reach critical mass and max equilibrium.

- Final and most important point-- this is relevant to Techcrunch article as well as commentary I've read on Quibb. Thomas Eisenmann articulates it well: "Virality and network effects are often conflated and confused, so the distinction between them warrants clarification... not all products that spread virally exhibit network effects. Likewise, not all users of products with network effects are acquired through viral, customer-to-customer transmission mechanisms."

Read more here: http://platformsandnetworks.blogspot.com/2011/07/business-model-analysis-part-5-virality.html

Advisor at Quibb

Awesome notes.

Fantastic points, John!

Based on the discussions that I've had following this article, I feel that we could probably have articulated more on the fact that we were analyzing network effects as a source of competitive advantage only in the article and not really on the value proposition that it offers. Hence, we were aiming for a takeaway more along the lines of: "The Network Effect as a competitive advantage (or a barrier to entry) is less effective than it used to be". (Because startups do see it as a significant barrier to entry)

The third point is an excellent point and something that is often confused in literature around this topic. While virality helps products with network effects reach critical mass faster, there is no 1:1 overlap between virality and network effects.

On the topic of competition, the other interesting trend is the rise of the non-direct competitor, competing with a very different product but targeting the same 'job-to-be-done' (e.g. self-expression).

Thanks for weighing in! This was an excellent add to the conversation here.

Seed Investor at Scout Ventures

Definitely agree on that takeaway. As a corollary, I've said before and I'll say again, platforms are going to get bigger faster. I have to point out that this also includes hardware (which has up-front purchase costs): http://www.economist.com/blogs/graphicdetail/2012/02/daily-chart-13

Partner at USV

This whole thread is great. I cannot add much to it but will try:

1. Network effects and virality are two separate things, imo, related but not the same.

2. There is a real question as to whether networks are as strong (in say 2012) as they once were (say in 2010). This is because we may be entering a period of intense fragmentation (which I have written about) such that it may be much harder for a service even to get escape velocity where the network effects really kick in. Also, as my partner Brad likes to say, network effects are strong on the way up, but also on the way down. Anyway, if you believe in this fragmented world (see http://blog.aweissman.com/2012/08/the-great-fragmentation.html) this might also explain companies raising and spending tons of $$ (http://continuations.com/post/34757870523/the-return-of-the-capital-intensive-startup).

3. All that being said, there are numerous instances of companies with network effects that are growing strong as a result of those effects - it's just that they may be outside of the consumer service space. For example, Waze, classic data network effect. Factual - another data network. Marketplaces in genreal - Etsy - for example. Skype, of course, still rules with its network effects. Finally, finance marketplaces (Lending Club, Funding Circle). I am biased as an investor in some of these, but as a firm that focuses on networks and network economics, they are a bunch of different places to look

Lead Analyst at HootSuite

Andrew Weissman, I'd like to add a fourth point, dormant network effects: a large inactive user base that can be reactivated given the right application or external event.

Yesterday evening on Facebook I saw two consecutive wall posts with screengrabs of Google Hangouts from completely unconnected friends. While obviously anecdotal and a coincidence, it got me thinking.

I always thought Google set up its social network as a dormant player (thanks to all of its other highly used services such as GMail), waiting for the right app to re-activate (activate for the first time?) its users (or simply waiting for some massive screw up on Facebook's part, whichever comes first). Google Hangouts is the closest thing Google has come to a killer app (or MHX) for the Plus platform. As a result, I would imagine Facebook is hard at work on a Hangout competitor....

Thoughts Sangeet Paul Choudary and Andrew Chen ?

I agree with Andrew Weissman's point on data networks. I would consider Wikipedia to be the first 'data/content network' and Waze, Factual etc. are enjoying a similar network effect which runs in through a community-owned stored value model. More users create greater value which builds the product's value (breadth and depth). Users collectively build a central pool for the network and more users (producers) leads to greater stored value so that starting a follow-on network becomes that much more difficult with every passing day.

The key, of course, is the number of users in 'producer' mode. Given that this is a cross-side network effect with consumers benefiting more from more-of-producers, maintaining a high number of users in production mode becomes critical. There is very little same-side network effect. Hence, a possible threat to such a model is a new network with a much simpler and more frictionless production mode. A blog-to-tweet kind of reduction in effort could disrupt a strong existing data network if it follows early enough.

I am a bit wary about the 'dormant network effect' of G+. It sounds great as a retrospective thing but I doubt anyone would want to start off a network effects business wanting to be a dormant player. I'm sure Google wanted G+ to be the next Facebook when they launched, at least in terms of sharing and liking of web documents (to get all the social data into search), if not the communication part. The current dormant-with-a-chance-of spike model seems to be the next best thing when that didn't work out.

Advisor at Quibb

For G+, another way of saying "dormant network" is- we can get a lot of people to sign up for something but no one wants to use it :) They could get 100M uu/month to a blank page if they wanted to, just by linking to it everywhere from across Google.

Jordan Ng, Business Development Manager at HootSuite Jordan Ng
Business Development Manager at HootSuite
Taylor Loren, Education at HootSuite Taylor Loren
Education at HootSuite
David Thai, Marketing Automation Manager at HootSuite David Thai
Marketing Automation Manager at HootSuite
Mark Eijsermans, Software Engineer at HootSuite Mark Eijsermans
Software Engineer at HootSuite
Stephanie Wiriahardja, Community Celebratoire at HootSuite Stephanie Wiriahardja
Community Celebratoire at HootSuite
Michael Tan, Financial Controller at HootSuite Michael Tan
Financial Controller at HootSuite
Craig Ryomoto, Director, Pro User Growth & Revenue at HootSuite Craig Ryomoto
Director, Pro User Growth & Revenue at HootSuite
Su Ning Strube, Director, Business Intelligence at HootSuite Su Ning Strube
Director, Business Intelligence at HootSuite
Destin Haynes, Director at HootSuite Destin Haynes
Director at HootSuite
Mark Holder, Director, Platform at HootSuite Mark Holder
Director, Platform at HootSuite