Growth Hacking Network Effects: or how to launch a new platform
Updated: Aug 25, 2019
Launching a new platform is no easy task. Whether you are a solo entrepreneur raising funds, you are launching a social project or you are launching a new platform initiative within an enterprise, you will need to show some validation.
Validating a platform business model requires showing evidence that a certain solution can facilitate the value creation process within an ecosystem.Matching successfully demand and supply implies a certain amount of scale to start with, hence the challenge. This is the famous chicken and egg problem of launching a platform.
Solving this challenge means: getting to the right design for product-market fit, (or platform-market fit), and the right scaling, having a (growth) strategy in place… all at once.
As product design and growth are so intertwined in this process I will touch on the design phase before we look in more detail at scaling strategies. After all the boundaries between the two processes are blurring.
I also want to introduce the concept of liquidity, which essentially means that there is a sufficient level of supply for the demand side to experience a valuable service. Our focus will be on this: reaching liquidity. In “Why Uber Won”, Greylock’s Simon Rothman said that in markeplaces:
Producer to Consumer (P/C) ratio“Liquidity isn’t the most important thing. It’s the only thing.”
To be pragmatic, (as in certain cases it’s the only way to be), it’s about having the right producer to Consumer (P/C) ratio. Having the right P/C ratio is the ultimate goal in the initial life stage of a platform. So the question is: what is the right P/C ratio for my platform? That depends and varies massively.
The answer could simply be: there needs to be the right amount of producers for consumers to have a positive experience….and vice versa. Let me explain with an example. In case of Uber could be that:
There need to be enough taxis for people to be able to hail a ride within 5 minutes at least 85% of the time (these numbers are just hypothetical)
There needs to be enough demand for the service for drivers to make on average £30 an hour (again hypothetical numbers); Airbnb found that in order for them to offer a good experience to travelers in a certain location they needed 300 listing for that particular city.
There is, unfortunately, no set rule to know what this metric should be for other platforms, but here is a way to put a line in the sand:
find a time unit that is as short as possible but big enough to give you a statistical insight (could be earning per hour for Uber driver or nights booked per month for AirBnB host);
get to know your producers and understand what *good* looks like for them, what competitors are offering or what their opportunity cost is:(how much does Lyft pay drivers? how much money a month is worth the hassle of renting out your spare bedroom?)
Your P/C ratio can be a hypothetical one at first, but it’s important that it’s eventually validated with your audience. If you hit your P/C ratio you want to be certain that it means your customers are having a positive experience on your platform.
Now let’s go through a few examples of how different platforms reached liquidity (or the right P/C ratio).
By far the biggest reason new start-ups fail is the lack of a strong product(platform)-market fit, as Reid Hoffman, founder of Linkedin, states in his recent book Blitzscaling. No matter how innovative, exciting or obvious an idea may sound on paper, we must first answer the following questions:
does our platform meet that demand better than existing alternatives (in case of a new market, a non-platform solution i.e. playing your game alone)
Is there a market for our platform? Is that market big enough?
While growth (or growth hacking) is often seen as a series of tactics, increasingly we have seen designers join the growth team within companies, just like Angel Steger, who is growth design lead at Dropbox managing a team of 20 . To reinforce this trend I particularly like a recent interview published by NFX, between Pete Flint and Andy Johns: Growth Hacking is broken . The key takeout is that the next frontier in Growth is about doing work that is guided by deep consumer insight. He makes the point that today many practitioners are familiar with a plethora of automation tools, but are not familiar with Steve Blank’ s work.
For developing a new platform the process can be summarised in three steps:
map the players currently populating your target ecosystem
understand deeply the context of the players in the Ecosystem, including their goals, performance pressures, potential to be leveraged and gains sought
create a holistic hypothesis of solutions that reach/solve goal and challenges
The goal is to design a solution that enhances and amplifies the ecosystem existing interactions, one that improves the lives/work of the existing players.
For steps 1, Platform Design Toolkit set has recently been integrated with the Platform Opportunity exploration guide: the Ecosystem Scan, coupled with the Value Chain mapping tools by swardley could offer the platform shaper a substantial head start in understanding the context, how the value is perceived and what elements are crucial.
Step 2–3 are essentially embedded in the traditional Platform Design Toolkit methodological approach, described in the User Guide, as a process which is more “emergent” than hypothesis-driven: the tools really help you “listen” to the ecosystem potential before rushing to hypothesize how you “solve the problem”.
Multiple tactics can be implemented to increase trust and they broadly fall in one of two categories:
Humanisation: making the players more personable, real-life like and accountable for their actions. On Linkedin, we are shown how we are connected to people and who do we know in common. Dating site e-harmony tackles this by asking people to fill in a long questionnaire so time wasters are excluded and even after that some profiles are not accepted: this leads to a higher quality of the whole user base. Community building is also extremely powerful for this purpose.
Risk dilution: putting in place services and processes that mean that part of the risk is taken by the platform owner like when there’s an insurance in case of failed delivery/payment; or like in the case of AirBnB’s where hosts are insured in case of house damages.
In some cases, particularly for one-sided platforms where the same agent can choose to play the role of demand and supply depending on the day, virality and network effects overlap. In such cases, all efforts around design and growth, (which will also overlap), should be focused on this. The chart below, courtesy of Sangeet Paul Choudary, gives us some good examples.
In cases where virality isn’t the be all end, it is still extremely valuable. Unfortunately achieving a high virality coefficient is extremely difficult, but aiming at achieving an average one should always be on our agenda. Some of the main strategies for this are:
Desining an experience that requires same platform communication e.g. Facebook
Designing a platform that facilitates same platform collaboration e.g. Houzz
Designing an experience that incentivizes sharing (e.g. discount on a restaurant if the table is booked for four or more; Groupon where deals are unlocked only once enough people join the deal)
Designing a platform for the creation of awesome content. This naturally leads to sharing and the creation of viral loops.
Andrew Chen recently spoke about this giving two examples. Platforms like Faceit that naturally generate video content and platforms that have a high visible offline experience, (which may be then shared online), like with the scooter company Lime.
Delivering growth (or product platform fit)
While there is no secret formula, (except passion, speed, and creativity), there are a number of tips and tactics that can help solve the famous chicken and egg problem.
This isn’t always the case, but it’s usually like that for 90% of platforms out there. There are many reasons why this is and the book explains that in detail. My personal insight on this is while sellers/producers might be looking to undertake a certain action, and that requires planning as it’s often a future based scenario, buyers/consumers are in part always guided by system 2 thinking, (a.k.a impulse), and losing immediacy in a sale process drastically reduces the chances of making it happen.
Here are a few ways to subsidise supply.
Use your own inventory
For a proof of concept, one can use the assets at his own disposal. For example, if I wanted to set up a platform of DVD swapping, I could start by opening up my personal collection. Similarly, AirBnB founders famously started by putting three air mattresses in their spare bedroom.
When Uber launched in San Francisco it asked a number of drivers to work for them full time so that those drivers would be guaranteed certain earnings and Uber could be confident to have enough liquidity in the market for passengers to have a positive experience.
When we launched Treasurehouse a platform for the swapping and recycling of kids clothing, we started with a stock of 5,000 used items we purchased — so that initial customers could get something and try out the service before they decided to send in their own kids clothes.
Here the lines can become blurry but a series of tactics can be deployed.
It can be achieved by offering non-monetary incentives like status, (which could mean having a special status/profile in the platform), or by offering anequity stake.
Crowdfunding can be an extremely powerful tool: it allows founders to validate the market, get funds and create the first set of customers all in one go. A number of multiplayer video games have done this, so from day one after launch they had 1,000s of people playing the game.
A joint-venture like approach be particularly suited in a scenario where supply is concentrated, which is not the typical scenario, but it can happen. For example, a neutral third party could’ve engaged all the major film production companies and given them equity to produce an exclusive video streaming platform a’ la Netflix. (The Hollywood studios did actually attempt to do this, but it was too late and execution wasn’t good enough).
This tactic has the potential to be turbocharged by blockchain technology, where platforms participants can become also platforms owners, creating the ultimate viral loop.
One option is to build an aggregator: this works particularly well for goods that are already sold online. For example Moneysupermarket in the UK compares deals for financial services products, for each click they send to a partner site they receive a commission which is agreed beforehand; this is also known as an affiliate model.
This type of information arbitrage opportunity is thought to be found very seldom nowadays, but nonetheless, it is still present. In Milan there are four different pay-as-you-go car services: an aggregator app tells you which one has a car closest to you. This platform is less than two years old. Long term however we don’t believe this alone can lead to success, but it can indeed be a starting point.
Start with community
We often think about starting with a transactional element and then adding on a community on top. This can work, but starting the community first has significant advantages:
You validate an interest in the topic
You create a very engaged audience from day oneIt’s a very capital-light solution
Generally, a solid and easy to implement strategy to do this is by starting a blog and allowing people to comment.
Leverage remnant inventory
Unfortunately, this is not a scenario that presents itself often, but when possible it’s always worth considering. It’s about leveraging third parties unused assets. You could argue AirBnB works this way as it allows people to monetise on the spare unused bedrooms in their homes. Another good example is Groupon that allowed producers to quickly gain a high volume of consumers, albeit at a discount, and consumers who took advantage of that discount. In this case imagine a theatre that two days before a big show has only sold 30% of tickets: getting more people in for a low fee will be better than having empty seats.
Partnerships (for distribution)
Previously I wrote about getting major players on board as equity partners: this makes sense if they are also producers. In many other cases, it is possible to approach larger enterprises in different sectors, with no conflicting interests and ask to offer a service to their significant user base in exchange for a success fee on each transaction.
When working at Jobinasecond.com, (the last minute job app I mentioned earlier), we made a deal with Just-Eat: at the time the largest aggregator of food delivery in the UK. As their partners, (the restaurants), sometimes struggled to have delivery drivers, the overall customer experience suffered. We integrated into the Just-Eat platform a button that allowed partner restaurants to contact us when they needed drivers. In this particular case as the value creation happened on all sides we were able to get leads at no cost.
Sometimes whom you decide to target for your platform can be the critical factor between the success and the failure of a nascent start-up.
Uber didn’t start by offering taxi rides anywhere in the world, but simply in San Francisco as focusing on one city made it easier to reach liquidity and have a strong proof of concept to raise more funds.
Similarly in 2013 when I worked on the launch of Jobinasecond.com, a service that allowed businesses to hire workers for relatively low skilled tasks within the day, we focused on one category of workers in one area of London e.g. food delivery staff in NW London. The required number of people needed to cover all jobs from security, to delivery, to manicuring in all corners of London would’ve made it extremely difficult to reach liquidity. (In the end, we failed due to disintermediation).
There are two takeaways here:
Understand if you are dealing with local or global network effects (…and try to start as local as possible)
Decide whether to start narrow or broad in scope
Keep in mind that starting narrow is always an advantage but in some categories there may be economies of scope e.g. Thumbstack famously offers a contact for all house related jobs required.
Once you have narrowed down who to market, you will be able to be more targeted with your marketing. The large advertising platforms offer some great tools for this like in market audiences (Google) and interest-based targeting (Facebook): both can be overlayed to a geo-targeting strategy.
While the above may sound a tad obvious, keep in mind the reverse process: running a relatively small test on these platforms can inform you which customers you can acquire more easily, which in turn can inform your initial platform strategy.
Paid acquisition is rarely a favorite for new businesses, but ultimately it is all about profitability and long term gains. It may also be a necessary pain for very innovative concepts e.g. AirBnB did a fair amount of activity on Google Adwords when it started.
Early adopters are a certain breed of people: they like to try new things and are constantly attracted to innovative solutions. Testing new stuff and see if/how it works is almost a hobby for them and often they are happy to deal with buggy solutions is they like the core value proposition. Broadly speaking these may be younger audiences, technologists, urban citizens, but in reality, if we leave broad categorizations aside, early adopters are quite a rare breed.
Having them on board can be very valuable as they are often happy to give feedback and once happy with the solution they will likely tell others with enthusiasm.
Finding early adopters isn’t straight forward, here are a few ideas:
list on Product hunt website: this is now hard to get on, but if it works it’s a great place to start
nascent social networks: Telegram a year ago, maybe TikTok now?
using other innovative products: they are using blockchain applications outside of bitcoin, have a VR headset or are experimenting with a 3D printer.
Design to facilitate launch
Earlier on I wrote how the boundaries between design and growth are blurring: in the cases below the two overlap completely. Design should bemade to facilitate the launch.
Single player mode
This tactic entails building a SAAS solution for one set of people, (generally the producers), get to critical mass, monetise and then add the demand side of the equation.
This solution is particularly popular as it’s easier to implement and it can generally be a viable business even without the demand side, thus decreasing risk and making it easier for investors to fund it. Interestingly in a recent interview James Currier, founder of NFX and guru on launching new platforms said:
“….not all our new portfolio ventures are two sided, but they are all network effect oriented”
The likelihood is that a number of the companies they are following are adopting a single player mode strategy. The most classic example here is Amazon: initially, they were just an online bookstore with a huge catalog and a huge warehouse to store that catalog; once they reached critical mass they became a marketplace, allowing third-party bookstores to sell through their site.
On a smaller scale, I recently met with a start-up that built a tool to help office managers manage the office supplies: once they reach critical mass it will be easy for them to add supplier on the platform.
When building a single player solution, the key is to always start with an MVP, or a pretotype (many tips on this on the book “The right it”).
Often times the solution built is distributed for free to increase penetration and speed, but it doesn’t need to be the case. Similarly, the solution can be designed for multi-player collaboration (see above section: Build for virality).
Not every idea is a new idea and similarly not every platform is unique. In the podcast “Masters of Scale” Airbnb founder Brian Chesky describes how they used to think about what would be a 5 star experience for their guests….and went on imagine how to exceed expectations to a 13 stars experience where the guest would be welcomed by their favorite rock star and was picked up at the airport with a Ferrari, (or something of the like).
Exceeding customer expectations is one way to get word of mouth goingand create loyal customers.
This strategy is increasingly used for competing against large platforms and there are many case studies. Instagram did it to Facebook, appealing to a younger audience, and Twitch and Vimeo did it to Youtube, specializing on music and gaming video viewers. Having a vertical and narrow focus can beat breadth offering. I am a big fan of the below diagram that shows how challengers are going after Craigslist in many areas:
It’s also interesting to see how the Craigslist website does not offer a great experience from a design point of view, leaving them vulnerable.
Curation is also a strong differentiator. A smaller but well-curated/higher quality supply can be very well received by your audience, even if competing platforms are already out there. In particular, this can be interesting for attracting connoisseurs, which often tend to be early adopters and influencers…so not a bad place to start.
One sided markets
One-sided markets are different from single-player mode in that there are transactions going on between different players, but those players may at different times play different roles. A good example may be Whatsapp where all users are at times the producers of content and other times the consumers of content.
Ebay started by targeting niche collectors as they buy and sell their collectibles of choice. This is a perfect case study as it’s the quintessential example of how digital platforms can connect niches: on eBay people who collected stoned to sharpen knives can easily trade those stones (that group includes a friend of mine who has over 200 stones).
At a smaller scale, when launching Treasurehouse.com, (a platform to trade kids clothing), we decided to give it a strong sustainability angle in order to attract people who would be interested in buying and selling clothes through us, rather than going after sellers and buyers separately.
This strategy when well executed can half the acquisition costs for new customers.
Often we tend to think of platforms as start-up businesses made by small teams or solo entrepreneurs. This doesn’t have to be the case. While agility favors smaller teams, enterprises have other key assets they can leverage: their large customer base.
There is little doubt that a solution like Waze, the traffic prediction app that helps driver pick the quickest route, would’ve been easier to launch for a firm like Volkswagen rather than having to start from scratch. It could’ve been implemented in all cars adding value to its drivers and then potentially sold as a mobile app to drivers of other cars.
The competitive landscape today is so hard to read because the big 4, or GAFA, (Google, Apple, Facebook, Amazon), have learned this extremely well. They move from market to market leveraging their large consumer base and deep pockets. Other enterprises, however, could follow this strategy.
For example a large bank today will tend to offer insurance products, like home insurance. It is not hard to imagine how a company like that could offer a discounted policy in exchange for the application of a smart device in the home. Such a smart device will then collect the data that will inform future policy prices. Through such a device and the data it collects it would be able to offer a brokerage service for home utilities…and that could just be the starting point. (Of course, this requires incorporating some new capabilities).
This entails getting on board some high visibility players and it suits a fragmented but heterogeneous market, like the music industry for example. While there are plenty of musicians out there and the music industry is certainly made up of a very long tail (Spotify adds one million new songs a month), there are still some blockbusters that everybody knows. In this case artists like Lady Gaga, Bruno Mars and Coldplay.
Having these big hitters on board will create substantial buzz on the demand side, (as many people want to listen to their music), and it validates the offering on the supply side: e.g. if Madonna is producing on that platform it can’t be such a bad idea.
This tactic creates visibility and validation simultaneously.
This is arguably the oldest trick in the book: it’s about artificially creatingscarcity. Demand will inevitably go up. Any club that is exclusive is where people want to be and just because there is a sense of exclusivity we feel that much better about being part of it. That means we will probably talk about it.
When planning the launch of business banking app Tide, (which operates as a platform for numerous B2B players), the team decided to have a purposely limited referral scheme, just to create a sense of exclusivity. Ideally, such exclusivity is fictional so that everyone feels special when joining, but growth is not limited. On certain cases however real filtering for quality can be a successful strategy e.g. a platform for doctors or lawyers.
There are two remaining tactics that are worth mentioning.
Most major platforms today cautiously scan their ecosystem to find a pattern of value creation and then decide how to cater from them. If such value is generated at scale by a third party, the platform may decide to own that space and offer that solution themselves cannibalizing the existing player. This, however, doesn’t always happen and it doesn’t always succeed.
Facebook decided to stop the growth of Zynga within its platform, but eBay ended up buying Paypal (only after trying and failing to copy the solution). Booking.com is an interesting example: their whole website was built to advertise on Google, so while their nesting had a cost, they became the most efficient player in what is certainly the biggest distribution channel in the world for hotel rooms and that allowed them to grow to huge scale.
While this strategy can be risky, it can also deliver meteoritic growth. When possible it’s one of the strongest launch strategies possible and that is why so many people today are looking with interest at nascent platforms like Blockchain, VR, wearables and home devices.
Focus on comms
A new product is never great, especially if it’s a new platform. If you can keep the expectations of your customers low, potentially by engaging early adopters, you may be able to start with a relatively weak P/C ratio, (poor liquidity). As always, if you can do one thing exceptionally well, that single thing may offset a poor experience in all other areas.
Whatever the case, it’s always wise to manage customer expectations carefully: ultimately they make the platform.
Launching a successful platform is never easy, but don’t be discouraged and follow the three following steps:
Carefully plan the design of the platform
Set a P/C ratio target that leads to the delivery of positive experiences
Work on achieving that ratio using one or more of the tactics above (and for each one of those there are many ways of implementation).
Finally, and this is where experience helps, the key challenge is managing the relationship between speed and quality. While speed always helps, none of the tactics above will work if implemented poorly, so take your time for the correct implementation, (or you’ll get false negatives) — simultaneously keep an open mind. If you are not getting traction perhaps it’s time to go back to the drawing board and find other ways to generate value within the ecosystem. Building a platform solution isn’t about imposing a service to an ecosystem but facilitating an existing one: this shift in mind frame is critical in creating a valuable solution.