Seven Lessons on Marketplaces

^{Authors: Laura Constantini, Lucas Abreu, Cassio Azevedo, Luiza Ulysséa e Guilherme Lima}

Marketplace is one of the most frequent business models used by startups, and is central to many of the investment theses we deal with. Today marketplaces are much more than just platforms that connect buyers and sellers with products and services – they are platforms that aggregate networks of markets, and generate value to users by using their superior experience and by offering greater convenience than other business models and spaces.

As value investors, we look for patterns, bases for comparison, that can help us value investment opportunities, and help scale-ups to build their businesses with scalability and efficiency. Learning from the experience of other companies is fundamental for achieving these two goals, and is how we have created a knowledge base and processes for exchanging ideas and discussing new business models with entrepreneurs. With this in mind we interviewed the founders of the principal marketplaces of Latina America:  Paulo Veras (99), Tiê Lima (Enjoei), Eduardo L’Hottellier (GetNinjas),  Helisson Lemos (Mercado Livre and Via Varejo), Pedro Carvalho (Cayena), Guilherme Bonifácio (Mercê do Bairro), Diego Libanio (Zé Delivery and Mercê do Bairro), Gustavo Fino (Zé Delivery and Nana Delivery), and Fernando Montera (Doji).

We did this survey in the first quarter of this year, to deepen our knowledge of how marketplaces are initially built and then developed. We interviewed 9 founders of marketplaces of different sizes, at different stages and in various business segments, from unicorns and companies already listed to startups at late and even early stage. Our aim was to map the main challenges that these guys have faced in a marketplace’s journey, and get a view of the way they discovered solutions along the route.

Below is a list of the main insights we gained from this research: 

Lesson #1: Solve the most difficult part first

Construction of a marketplace inevitably raises a chicken-and-egg question – which side to deal with first: Supply, or demand? Balancing supply and demand within a marketplace is an important step – so that both sides adopt the platform as recurring users and pay for the convenience and services that the marketplace proposes to offer. Recurrence is the first step for achieving the ‘network effect’ that will give the marketplace faster growth and ensure success for all its participants. 

And the first strategic step is to carefully define the focus of attack: whether it will be on the buyers or the sellers. Looking at the histories of successful startups, there is always a clear decision on which side to attack first. This will of course depend to a great extent on the context: In the case of Cayena, for example – a marketplace that connects food distributors with restaurants – the initial focus was to generate value for the buyers (restaurants). On the other hand, for Enjoei – a marketplace for used clothes – the initial focus was on people who were already selling used clothes on the web.  

Andrew Chen, who invests in consumer technology including social, marketplace, entertainment, and gaming experiences at a16z, recommends focusing on what he calls the “hard side” of the network – the side that is more difficult to motivate and attract – who will usually be a small number of people who generate disproportionate value. They contrast with what he calls the “easy side’” – the users who flow more easily to the platform, because they see the value that they can capture more easily than those on the ‘hard’ side do. 

Examples of the “hard side” – who would be more difficult to attract – might be: 

(i) for Uber, the initial core of drivers (who will mainly become “super-drivers”, providing 60% of all rides), or 

(ii) in the case of Wikipedia, the initial core of volunteers – now some 4,000 – who edit the vast majority of the pages – it was much easier for Wikipedia to attract readers than to find these writers.  

To make the product an important solution for the ‘difficult’ side is the critical point to be addressed right at the beginning. Our survey gave us an understanding of how each founder created their growth strategy, including an understanding of which side of the network was the more ‘difficult’ in each case: 

– Enjoei started as a blog helping people sell used clothes. 

– Zé Delivery started by focusing on generating value for distributors. 

– 99 focused on improving the experience of the driver, opting not to charge the take rate at the beginning. 

– GetNinjas provided the demand, and was able to find the supply by the team looking in classified ad pages to contract suppliers to find matches for the first clients.

Summing up: Success is in first winning over the “hard side”.

Lesson #2: PMF takes longer in marketplaces than in other business models 

When we talked about Product Market Fit with founders of marketplaces, they unanimously told us they were certain that they reached PMF around the stage of Series A – whereas in our experience PMF is normally reached by SaaS platforms at around the Seed stage.

Why later?

Usually technological innovations improve the user’s experience: they make existing solutions 10 times better, or 10 times easier to use, or 10 times less expensive. And usually, what marketplaces are offering is a solution of access. What they are giving consumers is a place to find what they are looking for, and at the same time reduce the selling effort of those who have a supply of goods or services to offer. PMF happens when people form the opinion that the platform provides an access solution that is much better than the existing ones – freeing up repressed demand, and reducing cost of sales, or cost of delivery, for the products or services.

For a marketplace, proof that PMF has been reached is when the platform succeeds in proving that it adds value for both sides – both buyers and sellers. On the other hand, most frequently founders perceive that it is much easier for them to add value on one of the sides than the other. This discovery is very important, because winning the challenge on the more difficult side only happens when the team finds a strategic solution to meet the needs of one of the sides, so that it can then show the other side the existence of an access solution – whether it is access to their product or service, or to a base of clients. 

How can this be measured?

The answer we heard is: by monitoring metrics of growth in activity on the platform, and the size of the retention cohorts of buyers and sellers. In other words, in marketplaces the need to create a minimum density of buyers and sellers – able to guarantee liquidity of access to what both sides are seeking – takes more time than SaaS platforms (for example) normally take to create their sales machines. 

Lesson #3: Liquidity is the key orienting metric 

There is nothing more important for a marketplace than “liquidity” – which translates as “the probability of finding, on the platform, a buyer or seller for what one is selling, or buying. On the buyer’s side, this may be measured by ‘share of wallet’ – i.e. by how much of a user’s budget or need for a product/type of product the platform supplies, or by the time it takes for a buyer to have their needs met. On the seller’s side, it can be measured by the rate of engagement – for example how important a role the platform plays in the sale of its inventory. 

In all the companies we talked to, adequate liquidity for the opportunity in question was the most important indicator for assessing the success of the network: 

– In the case of Enjoei, the inventory turnover of stores – how long it was taking to sell (e.g. 7, 30 or 90 days) – was an essential component of the liquidity metric. 

– For 99, the useful metric was: what percentage of drivers’ income was provided by the platform.

• Cayena calculates (i) the ‘share of wallet’ of restaurants’ purchasing budgets, and (ii) what percentage of the distributors’ sales are made through the platform. 
• Doji, a second-hand consumer electronics marketplace, calculates what percentage of bids are converted into purchases.
  

Lesson #4: Pricing is an empirical science (even subsidies have been used)  

Even in cases of success, pricing is hardly a very scientific science – 87.5% of the founders we spoke to reported that it was very challenging to arrive at the most efficient way of monetizing the value that the platforms delivered.

The trend is usually for marketplaces to monetize through ‘take-rates’ – a percentage of the price of the transaction. There are two assumptions underlying pricing for marketplaces, including those that are successful: 

1. The need to continuously carry out tests.
2. Analysis of the pricing points applied by competitors. 

Mercê do Bairro is a good example. They are a B2B marketplace for small/medium retailers, whom they aim to help find the most competitive prices for products by comparing wholesalers and distributors. As soon as they get a significant base of customers, they intend to grow their monetization through new lines of revenue, such as verticalized products. Their strategy of subsidizing some products came from the idea of offsetting over the long term, through the sale of new solutions to their target clients.

By comparison, Cayena always gave priority to charging a take rate from the first moment, beginning with a very small percentage, charged to buyers, which it then increased over time, as the team began to aggregate other products and services and proved their importance to the sellers, to the point of convincing them, too, to pay take rates.

The entrepreneur Lenny Rachitsky – who is an angel investor, and publishes a marketplace-focused newsletter – has created a way of rationalizing pricing of marketplaces, and helping founders to arrive at their take rates: 

 “ Take Rate = Convenience + Demand – Competition. ” 

– where:  

1. Convenience refers to how much easier it is for the client to use the services of your marketplace to carry out its activities (the greater, and easier, the access is for buyers, the better).
2. Demand: How much valuable demand the platform will generate for suppliers/sellers (reduction of cost of sales, and ‘time to revenue’): the greater the demand, the better. 
3. Competition: How competitive the market is (the fewer the alternatives to your marketplace, the better).

Lesson #5: Macro metrics and funding rounds 

In our quest for metrics indicating traction for each round, the interviews gave us an understanding of what were the expectations and challenges for each round, where each one happened (‘milestones’), and what was the size of the rounds raised. We already had some interesting statistics – such as those in the Marketplace “Napkin” published by Point Nine, but we are interested in bringing this analysis to Brazilian marketplaces, and having an updated local base for comparison with SaaS platforms.

When we consult our own data and history library – which we call “The Plot” – and compare it with the information from founders we interviewed, we get the following points:

Size of rounds: 

• We did not find significant differences in the amounts in early stage rounds (up to Series A, inclusive) between Marketplaces and other business models, but we did find them for late stage rounds.
• When we said – in Lesson #2 above – that marketplaces need more funds to reach PMF, and that the rounds have practically the same size, we are saying that it takes more time, more rounds, and thus more funds, to reach: (i) PMF, and (ii) the same amount of recurring revenue per month, than in SaaS platforms.  

Speed of growth:

• On the other hand, growth of GMV is faster, and especially after PMF the network effect created in marketplaces produces important benefits such as reduction of CAC, maintenance of accelerated growth, and the possibility of increasing the take rate over time.

Pre-money valuation:

• We have seen that the size of rounds in marketplace startups is not much different from those in SaaS, but if we compare generation of revenue, SaaS businesses are out in front. 
• However, the dynamic of marketplaces often facilitates periods of fast growth, especially after PMF, and at this stage marketplaces can then command premiums in relation to the rounds of an SaaS project. 

Read more about this lessson at: Comparing macro metrics for Marketplaces across investment rounds , by Cassio Azevedo

Lesson #6: 4 major phases 

Going deeper into the numbers and information we collected, we are able to identify 4 phases of development and growth of marketplaces. Here we list them, and their main challenges

Phase 1 (Pre-Seed) – Discovery: 

The aim is to understand how the planned solution can generate convenience, better access to products and services, efficiency, and added value, for the value chain as a whole. 

Components: 

• Proving that the solution makes sense (i.e. that PMF is possible) – monitoring the day-to-day activity of clients, mapping any areas of attrition or friction imposed by the traditional solutions, and understanding how the planned marketplace can bring solutions for any of these points that are identified.
• Understanding of the margins on both sides – what financial benefits the marketplace can bring, and the possibilities for monetization. Reflecting the fact that marketplaces give buyers ease of access, and liquidity, and give sellers a lower cost of sales, it is based on these analyses that teams understand the potential for monetization of platforms.   
• Creation of possibilities for the ‘Network Effect’ (NfX): Convenience, liquidity and scale gains are important components for construction of NfX, and a good NfX can also be an entry barrier. 
• Consideration of the possibilities for positioning.
• Consideration of the possibilities for scaling up the sales machines of both sides.
• As with all scaleups at the pre-PMF phase it is very important that a lot of the activities and tasks should be done by the founders themselves, and ‘by hand’ – not only to make it possible to map them accurately but also to get optimum understanding of what needs to be built.

Phase 2 (Seed) – the ‘Atomic Network’: Liquidity

• With this mapping and these alternative hypotheses in their hands, the founders begin to build the tool and automate the activities, to add value to both sides of the marketplace.
• When the MVP of the platform makes it possible to increase the volume of transactions, the team begins to experiment with the first signals of liquidity.
• Phase 2 has an important aim: Arrive at the “Atomic Network”. This term was coined by Andrew Chen in his book The Cold Start Problem, where he defines it in these words:

“ An atomic network is the smallest network that can stand on its own. 

It needs to have enough connection density and stability to break through early anti-network effects and ultimately grow on its own. ” 

A small network able to stay alive on its own is a very important element of scale and network effect for marketplaces, and for this reason may be considered as the equivalent of PMF for SaaS platforms. While SaaS start-ups usually achieve PMF at the end of the Pre-seed phase, marketplace startups usually achieve an Atomic Network between the middle and the end of the Seed round. 

Phase 3 (Series A) – Efficiency and scale

• Phase 3 is when the funds of the Series A round enter the start-up to multiply the Atomic Network. In this phase the team already understands the effort and cost required to build an atomic network, and therefore can build a playbook to boost the growth rate. This playbook is comprised comprises of repeatable and predictable processes, meaning it can be taught and replicated internally with predictable results. 
• Growth enables the team to identify other demands from both buyers and sellers that have not been met, and give priority to construction of functionalities both for monetizing these opportunities and also for reducing friction and avoiding frustrations for users – intensifying the Flywheel effect, and recurrent use.
• The founders can create a growth strategy involving other forms of engagement and monetization, with the aim of increasing frequency of use, and share of wallet.

Phase 4 (Series B >) – New adjacency strategies entry barriers

• From Series B onward, the teams make efforts to discover and build new adjacency strategies and effects – one of the interesting aspects of marketplaces is the capacity to improve participants’ experiences by adding new services: Addition of new services and creation of a more comprehensive experience allows the platform to increase take rates. Other services can be created and offered on top of the original transaction, which can help the marketplace to bite off increasingly large slices of the addressable market.   
• At the same time, a higher volume of transactions also makes it possible to create a database that can be analyzed for constant improvement of performance and unit economics. 

Lesson #7: Unit Economics 

Marketplaces are network-effect businesses, and as such they need a minimum density of users to enable the value proposal to be delivered.

Until the platform reaches a minimum transaction rate and volume, it is natural that the founders should have difficulty in getting a clear view of what their unit economics are in practice. As described above, the first notions of monetization and unit economics begin to appear with the creation of the Atomic Network. 

One point that successful founders have in common is an understanding about the value chain, and the available profit pool, in which it will be possible to anchor proposed monetization structures. 

Another interesting point is that there is a significant difference between older and newer business models in the way unit economics evolve and develop. With today’s more complex structures – more evolved technology and greater capacity for integration with various solutions, including offer of financial services – operation becomes more efficient for the user, and this in turn allows greater visibility for possible new adjacency strategies, and for the next strategic steps that could provide a significant improvement in unit economics. 

‍