Andrew Chen, Advisor at Quibb

Andrew Chen shared:
How to build a "Distribution-first" startup


Read what Andrew Chen 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
Business Development, Appstore at Amazon

guess this is how Quibb is born? :)

Reply · Tweet ·
Advisor at Quibb

Haha- I'll let Sandi answer this herself :) But in many ways, Quibb is the anti-growth product :) Hipster growth.

Engineering at Pinterest

Sounds like your next blog post just named itself!

Reply · Tweet ·
Editor-in-Chief at Quibb

Yeah, not quite ;)
With Quibb, I think a couple characteristics make the product easier to spread and grow versus other types of consumer products (i.e. news is timely and therefore high-frequency, active discussions, leveraging existing networks, etc.). I didn't start by thinking about these characteristic, but they probably mean that in the future Quibb *should be 'easier' to grow than other products that are low-frequency, don't have any network effects, etc.
I definitely haven't ramped up growth as much as I could - like Andrew Chen mentions - 'hipster growth' (I think Mohamed Zahid actually coined that term a few weeks ago!). While the acceptance rate for membership has gone up recently as a lot of people are joining through invites, I'm still not accepting about 70% of all applicants.

Lead Analyst at HootSuite

I believe it was "indie growth" (less pejorative) :)

Reply · Tweet ·

To me this mangles Andy's statement a little bit. Even with a distribution first startup you still don't have any technical risk?

The danger I see from how you wrote this is that you end up in the local maxima of a company where you share photos or send free smses (or porn or torrents).

I may be a romantic but I think that you have the best chance of solving problems for customers you care about and that should be the root cause of starting a company.

Agree on this. If you're backing into a startup through the distribution channel, then where is the purpose, passion and drive behind the company? This sounds like it would only work for those entrepreneurs who are looking to copy successes from one category to another.

To be clear, I'm a big fan of tech, design and marketing all working together to make the most kickass product and experience. But I wouldn't go so far as to say that distribution is the key driver of decisions. The customer and his/her needs are always first and will always be first, because this person is the end user of your product.

Advisor at Quibb

Yeah, I'm just saying that if market risk is now the #1 thing, then you can alleviate that by solving the distribution problem first and foremost.

Ultimately though, you need a balance between lots of opposing forces- distribution, passion, defensibility, high margins, competition, etc., etc. But I think the "distribution-first" strategy is certainly an interesting lens on the problem of starting a company, though you quickly have to shift into thinking about the other problems. And if you decide that you don't want to do photos, texting, or porn, then that's fine- you can still build Dropbox, Skype, Yammer, Hotmail, Twitter, or many other fine products. But it's worth baking in at the beginning even if you quickly start working out the other issues.

But yes, I agree the balance is tricky, and people who think too short-term might find themselves running a porn site :) You still need values, after all.

Gotcha. I love your initial point about products being difficult to market. It seems obvious to me, but this realization hasn't hit the mainstream yet. I hear founders talking about it all the time, though, so I think it's only a matter of time marketing/growth/user acquisition is considered cost-of-entry to building a tech business.

Executive Producer at Behaviour Interactive

I agree with the fact that one shouldn't found a company just on the basis of applying viral tactics to a to-be-determined market/segment. I don't see that as a sequence, but as a system (which other comments support as well).

Trying to make a less-than-perfect parallel here: I don't think anybody in the 40-50s started trucking companies first and then tried to find what to transport. Industries transformed themselves (and were created) because modern mass and long-range transport was now possible at cheaper cost. One could argue that you might have a success building a smart distribution channel first but IMO the real disruptions happen when you start with a deeper segment knowledge, bringing a smart innovation to a problem and applying the now-must haves virality AND monetization strategies.

Reply · Tweet ·

The best way of expressing it is what Peter Thiel said about distribution (http://blakemasters.tumblr.com/post/22405055017/peter-thiels-cs183-startup-class-9-notes-essay). Essentially it's a product to be designed like the business is a product.

So essentially there is three products to a successful company:
1. The actual product
2. The product design of distribution and
3. The product that builds the product (company design)

Reply · Tweet ·

There are a lot of problems we can solve - some are tractable, some are not, and there are many potential solutions to a problem. Distribution-first a useful tool for focusing those decisions, for sure. But it's more than that.

I recall a talk by one of the Bump founders. Bump's distribution hinges on a specific moment where you need to give someone your contact info, and the product gets spread by fulfilling that need.

So, partly in reply to Niki Scevak, I'd say Distribution-first, is not necessarily in conflict with producing value for the customer. Done right, it forces you to understand the relationships between the stakeholders and build a product that adds value to those relationships.

Co-Founder & CEO at edshelf

Another way to think about it is, "the root cause of starting a company is solving problems for your customers" + including a core trait that makes your solution exponentially more useful to all of your customers if they share it with others. A great solution is awesome, but a great solution that has a natural "virality" to it is even better :)

Builder at Gametime

Super interesting. It'd be a valuable exercise to think of all current categories for which this works - file-sharing, communication, publishing - are good examples. Photo sharing (maybe a subset of file sharing) is arguably the big winner in the past year. Within communication, Chat apps have done well. What else was or will be good categorically?

Reply · Tweet ·
Advisor at Quibb

The best stuff is arguably areas that are big but untapped- taking social or chat or something and combining it with a product category where it enhances the offering.

It's part of the reason I really love what Yammer did in the market. In fact, I think there should probably be more social + professional versions of everything- Quibb is obviously an example, but I just got involved in a company called vsee.com that's like a "Skype for the enterprise" which has viral characteristics but can also still be sold B2B. Lots of potential innovation by thinking through the various categories.

Reply · Tweet ·
Founder at Fundamentum

Agree that it's powerful to build viral loops into products. I feel like many people focus on that and ignore paying attention to more traditional ways of bootstrapping audiences though. So how do you get that first 1000 right users into the system so that the virals kick in?
Sandi MacPherson probably has some thoughts on that too... :)

Reply · Tweet ·
Advisor at Quibb

Yeah, I think ideally you still have an incubation phase where you bootstrap an audience up to some point, before you go nuts on distribution. But it's good to have a plan for scalable distribution from day one if you can :)

Tinkerer at Amazon

Agree with the thesis here. In fact, I would take this one step further and say that you've got to think hard about not just distribution, but also monetization from day 1. I'm certainly not advocating for for every new consumer product to start monetizing from day 1, but you've got to have a clear plan for monetization and a strategy to test that monetization hypothesis without requiring massive distribution (in the millions of users). Products which require massive distribution to monetize are trying to catch lightning in a bottle, and the chances of success are significantly lower. The days when every social/mobile consumer app could raise a hefty seed round solely on the basis of user growth with a we'll-figure-it-out-later plan to monetize (most of 2011) those users well are gone (or dying). I think a lot of start-ups which start off only with the core customer problem and say "we'll just focus on solving this problem, and figure out how to acquire users and monetize them later" are shooting themselves in the foot. You've really got to have a clear plan for all of these initially, and constantly evolve that wholistic plan as you iterate on the product.

Reply · Tweet ·
Growth Marketing Manager at Pinterest

Reminds me of this article from Reid Hoffman: http://techcrunch.com/2009/04/19/reid-hoffman-my-rule-of-three-for-investing/

Had a lot of influence on how I thought about growth for GrubHub. You can either do it through product (Yelp with SEO), though other online networks (YouTube with MySpace), or through other offline networks (Facebook with colleges). People tend to focus on the first two almost exclusively now, but with social or referrals instead of SEO, and Facebook/Twitter instead of MySpace. But, an offline distribution-first strategy can work too.

Reply · Tweet ·
HubSpot, Coelevate

While you focus on consumer products with viral distribution I think this focus of Distribution first works (and should be applied) to most other types of companies. There are very few scalable distribution channels and strategies. They are all extremely competitive. A startup needs to have a built-in advantage in their company over the broader competition space in order to compete in those channels if they want to scale.

A lot of startups enter the market by using a new channel where competition is low. But that doesn't last forever. Competition eventually comes in and drives up noise/costs. That advantage can be in any part of the AARRR marketing funnel. Kixeye/Kabam is a great example in social gaming. They use the same channels as Zynga and other players. But they monetize so much better (Revenue piece of the funnel) that they can pay more in acquisition costs for their target audience.

In any case, I think the past 5 years of the startup ecosystem has been too focused on a "product is everything" mantra. I think the general sentiment is now swinging back in the other direction where focus on distribution earlier in the startup process is growing. Your post about "Distribution First" certainly captures this. Great job.

Reply · Tweet ·
Co-Founder & CEO at edshelf

I passed this article over to a friend who's an aspiring entrepreneur and doesn't come from the high tech world (e.g. isn't familiar with terms like "viral loops", etc). I summarized the article and comments like this:

After you've clarified the problem and the audience, when you think about your solution, think not just on how great your solution is, but also on how your customers will love to share your solution with others. I don't mean, "Hey, check this out, it worked well for me." I mean, "Hey, check this out. If we both use it, it will be even more useful for both of us."

Co-Founder & CEO at edshelf

Some more quick thoughts.

With physical products, distribution has always been a key initial consideration. Concepts like shelf space, packaging, signage, warehouses, etc, are all old hat. Those can be translated to the high-tech world as search results rankings, app store/directory rankings, social media mentions, ratings & reviews, customer referrals, etc. And like physical products, I totally agree that distribution needs to be a key initial consideration too.

> If you need a ton of traction to succeed in consumer internet, then you have to figure out your distribution model.

In some markets, I'd even posit that traction is king. Not product, content, or anything else, because having traction means having product-market fit, distribution, etc.

Reply · Tweet ·

This is a great point, Andrew!

I'd like to qualify, though, that while the internet is an equal opportunity network, when it comes to distribution, some startups are more equal than others.The key to distribution is to understand network-based distribution.

Network-based distribution is fundamentally different from channel-based distribution. The fundamental aspect of network-based distribution is that every node has the potential to be a consumer, a producer and a distributor as long as your product allows it to be. A producer is someone who initiates interaction, a consumer is the recipient of that interaction and a distributor passes the word on. Every node on a network can choose to use or spread your product. Hence, the key to distribution is to understand the motives of different types of nodes (don't just club them into the uber-term 'users') and convert them.

1. Network-based distribution works especially well for products built on communication and sharing: Hotmail, Skype, Napster allowed every node to be a producer and distributor. The producer has an incentive to get the word out and the other consumers subsequently become producers. The key here is that the producer and consumer roles are blurred so distribution works especially well. But in recent times, SurveyMonkey, MailChimp, DropBox, MeetUp and Ning have a similar dynamic. Every producer has a natural in-built incentive to invite other users to experience the value prop. SOme of those invited consumers become producers in another instance. Quibb also has a similar dynamic. A more detailed essay on this at: http://platformed.info/viral-loop-surveymonkey-mailchimp-ning-meetup-eventbrite-paypal-hotmail/

2. Network-based distribution works especially well for products that help self-expression or self-marketing (promote user's cause): A great way to ensure users keep spreading the word around without even explicitly having to talk about your product is by having your platform enable them to market THEMSELVES. This is what YouTube, Instagram, Change.org, KickStarter etc. do for their users. And along the way, some nodes become distributors and share the cause further. A more detailed essay on this at http://platformed.info/naturally-viral-how-to-design-your-platform-for-self-expression/

3. Network-based distribution enables some products to ride the growth of other products: When product A solves a key pain point for users of another product B, product A has the potential to leverage the growth of Product B. Paypal did this with eBay (allowed users to transact online). YouTube did this with MySpace (allowed users to host a music video). Flickr did this with the Blogosphere. This is a unique unfair advantage that can be exploited only on a network. Again, longer detailed essay at http://platformed.info/how-paypal-youtube-and-stumbleupon-gained-rapid-traction-through-piggybacking/

4. The internet is a network, not a channel. Think distribution of product functionality, not just destination: Destination is where users come to interact with your product (an app, a website etc.) Product distribution, on the other hand, allows your product functionality to be present in the context of the user (e.g. the browser extension for Evernote helps you capture content at the point of use rather than needing you to copy and paste back). Instagram is in-context, Flickr is a destination. You should think of ways to distribute your product functionality across the network in chunks (widgets, API functionality, browser extension etc.) rather than just focus on drawing traffic to a destination. Again, essay at http://platformed.info/destination-distribution-product-traction-widget/

Apologies for the multiple links, I wouldn't have peppered this comment with the links if I hadn't felt that they add direct value to the conversation here.

To sum up, I feel certain startups are more pre-disposed to network-based distribution (points 1 and 2) and certain startups are more aware of the importance of using it (points 3 and 4). The problem occurs when startups start treating networks like channels and start layering incentives completely unrelated to the product, and in doing so, end up with a cup that runneth over, but which leaketh over too.

Advisor at Quibb

Fantastic points here. And totally right, some products are well-suited to have network-based distribution, and some do not. I wrote a blog about this a while back specifically on B2B/SaaS products called, SaaS isn't viral*, exactly because so few have network-based functionality in the product. Some do, like Yammer or Dropbox, but that's the exception to the rule.

*link: http://quibb.com/links/saas-products-aren-t-viral-with-audio-track

PS. This needs to be it's own post! Maybe Sandi MacPherson can work on that, a "Share a comment as a link" feature :)

Reply · Tweet ·

I think that a solid distribution strategy is a competitive advantage nowadays, for consumer internet startups. As you said, that's where the market risk is. Plus, it's not a "sexy" part of starting a startup, and not one that hackers think through as often as they should, so it's an opportunity for startups to take advantage. I think you can point to lots (most?) of the big consumer products and argue that they largely won on distribution. LinkedIn is the perfect example of a product that won mainly on distribution, IMO.

Founder at Glyder

I 100% agree with what Niki Scevak and others have said about choosing the thing you work on based on the problem you want to solve rather than starting with the viral channel/distribution strategy and then figuring out what product it applies to. The reason I believe this is that distribution channels change frequently but your overall vision for your business should be something that you're excited to spend 5 years working on, or at least I think you have to go into it thinking on that time scale. Focusing your entire business on a distribution channel that may or may not even exist in 3 or 6 sets you up for a major "what now?" moment when the channel inevitably disappears or become less effective.

However, after you've picked the problem that you're excited about working on, you still have to answer the question "what do I build today/this week/this month?" to make progress against that problem you're trying to solve. And in most cases, especially in consumer products but also in SMB and increasingly in enterprise, figuring out what you can build to test/validate distribution is probably one of the most important filters to focus on when prioritizing your early stage product roadmap.

A brief example: I talked to two founders yesterday that are working on a niche two-sided marketplace in a space they're passionate about and have domain expertise. They know the general problem space they're trying to address (connecting buyers and sellers more efficiently in this niche market), have a vision for the product that builds the marketplace to solve that problem, and they need to start building a working prototype/MVP in the near future. Two sided marketplaces are tough - there's always that chicken and the egg problem, so what do they work on first? Applying the logic above to their particular case, I suggested they consider focusing their short term roadmap on building the consumer side of the product to validate that the users in their target demographic engage in the behavior they're banking on for distribution (sharing photos and expressing intent to purchase for a particular type of product) frequently and in high enough volume to build up the demand side of their marketplace. They hypothesis here is that if they get enough consumer interest, selling into the supply side of the marketplace is probably going to be relatively easy.

Reply · Tweet ·
Technical Program Manager at Expedia

This is a thought-provoking post. “Distribution first” is really just a particular instance of the general product development strategy of identifying and mitigating risks early. I agree that the distribution risk for a typical internet startup today is greater than the technical risk. There are so many powerful development frameworks and public APIs available that a small team can put together a highly functional product in a matter of days. Witness the impressive results from many hack-days and multi-day hackathons.

Getting a product into consumers’ awareness and considerations sets is challenging. Virality is often serendipitous and seemingly random. Your chances of success will be much higher if you can determine a target audience and understand its wants and needs. Your understanding will be fuzzy at best until you can start to observe your target audience’s response to the product, so I argue that:

--It will be valuable to have a minimum viable product (http://en.wikipedia.org/wiki/Minimum_viable_product) in tandem with a Minimum Viable Distribution Channel.

In a startup situation, where you want to be prudent with your resources, an Interesting corollary suggests itself:

--You shouldn’t expend startup resources significantly beyond the Minimum Viable Product without implementing a Minimum Viable Distribution Channel.

Technical Program Manager at Expedia

I mentioned serendipity in response above, in which I used the concept of a Minimum Viable Product in my argument. After posting the above, I looked at recent articles and lo and behold, Paul Kortman has a whole post around Minimum Viable Product. I promise that I didn't see Paul's post before I posted the above :)

Advisor at Quibb

Great comments! In terms of virality being random- I think that it's like 60% random :) Or, that is to say, you build a great product experience and you sometimes will get a bunch of random word-of-mouth traffic because people love what you do. Your product having "natural virality" from the product category is another factor here. The other 40% isn't random at all IMHO- it's a variety of factors, from the competitiveness of the marketing channel, the level of optimization of the viral loop, etc.

Reply · Tweet ·

This is an important clarification. Virality is often misunderstood. There is the random word of mouth that you did not engineer but the spread at every node as well as the conversion of nodes from receivers to transmitters can be given a significant boost through some engineering. (Of course, the nodes here are people who are part of some viral loop)

Co-founder and CEO at Denki

I think this approach you're describing is the one we took at Denki when we started building our word-game Quarrel back in 2008 - though it was less considered, more intuitive than you describe. But I recognise the logic retrospectively.

It didn't work out well for us back then for one simple reason - we chose Microsoft's Xbox Live Arcade distribution channel, which was gated. We'd shown them early prototypes and received confirmation that access wouldn't be an issue. However, in big companies strategies get reviewed and people change - a product that fitted perfectly last quarter suddenly doesn't fit the long-term vision today.

So all I'd add is to make sure to choose an open distribution channel if pursuing this strategy. Gated channels can seem very tempting - ready-made audiences, supported development eco-systems, unmatchable marketing support from channel owners, etc., etc., but they can disappear in an instant without warning.

We've yet to witness the fallout from a major, high-profile platform closing or fundamentally restructuring, but there have been a few small-scale warning signs: Facebook insisting on games using credits and taking 30%, Apple banning interpreted Flash code, Twitter limiting API functionality for 3rd party apps, and so on. Decisions like these can happen over-night and have catastrophic effects for startups whose strategies depend on these aspects of their chosen distribution channel.

So, in summary, I think distribution first is definitely a good strategy, though it's not without its own potential gotchas :-)

Reply · Tweet ·

As an "internet veteran" Andrew Chen has done an admirable job of fleshing out the distribution-first startup, which I think makes a lot of sense. One way to think about this is that as entrepreneur you are playing with a Rubik's cube or sudoku -- (pick your real-life example system of simultaneous equations) -- and you win (find Product-Market Fit) -- when you find a solution that satisfies all the variables -- a rough abstraction of these variables would be: technology, product, marketing, distro channel(s) and customer segment(s). So, for the distro-first startup, you hold the distro channel constant and quickly iterate through the other variables. There are more than a few startups embarking on such as mission de facto -- that is, they have committed to a platform and now must make the other variables fit with the constant. This line of thinking can also be taken to segment-centric, or tech-centric -- that is, you as an entrepreneur have some unfair advantage in how well you know your segment or technology, or marketing or product -- and then double-down on your chosen variable and make the rest align.

Sr. Director, Global Sales Ops at Spotify

I like the construct that Tom Tunguz has blogged about at http://tomasztunguz.com/2012/05/17/your-startups-top-3-priorities/ - Distribution, Engagement, Monetization.

Data Scientist at TripleLift

A bit late to the party but this is an awesome post. Something I've been observing is that there are more hardware startups being launched and I suspect they fall into the "high tech risk, low market risk" camp.

Sandi MacPherson, Editor-in-Chief at Quibb Sandi MacPherson
Editor-in-Chief at Quibb
Andrew Goldner, Advisor at Quibb Andrew Goldner
Advisor at Quibb