Which decentralized marketplace token models work?

Originally published on June 8, 2018

I am always fascinated when crypto teams take on large established, but still fast growing and innovative centralized businesses. For this example, I decided to use a fairly well known service: online bookings like AirBnB and booking.com

The platform I will analyze is called Locktrip. It aims to offer a decentralized 0% commission booking alternative with hotels and other properties. An international team is developing it, they have a working alpha version of the platform and are working towards a full launch in Q3 2018. Their token is called LOC.

DISCLAIMER: I have neither participated in the ICO nor have any relation to the company or the team. This is not investment advice. Invest at your own risk and do your own research.

The TL;DR of why looking at this:

• A decentralized marketplace and many of the points below are valid for other marketplaces
• A utility token, which is used for internal payments, which, in my opinion, does not really provide any meaningful incentives to own and its long-term perspective in this form is not convincing (also questionable if you need your own token for payments only)
• Main bottlenecks: # of listings, brand development, business development, insurance, verifiability, UX
• To drive network growth, #listings need to increase dramatically
• Other token models could be much better suited to set these incentives: work token, TCR, discount token etc
• The current price is pricing in quite a bit of development and network value already — plus, quite sensitive to a few key assumptions

I start by valuing LOC using its current token design —internal payments and a lock up for the duration of the traveler’s stay.

Locktrip at a glance:

• LOC is an ERC20 token
• ICO in November 2017 at an avg price of 885 LOC for 1 ETH, raised 10.500 ETH ($4.5m)
• Effective ICO price in USD of $0.49, today [8 June] LOC is at $1.06 (118% increase relative to its ICO vs. an increase of 37% if you did not participate and held ETH for the same period)
• Distribution: 50% public, 25% team/advisors/contributors (1 year vesting), 25% locked in reserve for future fund raising/bus dev etc (1 year vesting)
• LOC is listed on 6 exchanges, however volume only on HitBTC and Fatbtc: $50,000–$100,000 per day
• 100,000 hotels currently on the alpha platform
• Marketplace expected to officially launch on 30 September 2018

Blue sky assumptions:

• Become as big as AirBnB today in 10 years: ca. 28.5mn bookings worth ca. $20bn annually — this equates to a global market share of 1.5% in 2029
• Global digital travel market grows at 8% pa reaching $1.4trn in 2028
• Have 4.5–5mn online listings
• Ratio of bookings/listings >6x, inline with AirBnB
• $127.5 average price per night ($150 less the 15% commission that LOC aims to eliminate) for an average of 4.5 nights per booking

Doing a back of the envelope calculation based on supply/demand for the token given the above numbers would produce the following conclusions:

• At some point in a few years, demand for the token should eat up all available tokens at face value and from then on the token value will increase inline with network growth
• The LOC team has the token reserve to support liquidity in the network until demand increases sufficiently. I wrote about the working capital pool issue in more details here
• Once the network is approaching full token demand, velocity should start to decrease and approach its bottom limit which, according to LOC’s team, should be driven by the lock up period for tokens — 4.5 days average stay + 2 weeks reservation period = 18.5 days. When we divide 365 days / 18.5 days, we get a velocity factor of ca. 20x. All this means is that when the network gets large enough, all tokens will be locked for 18.5 days, then released and re-locked, 20 times per year. When this happens, tokens start to “split” and increase in value
• The fact that we can compare the same service in USD and in LOC (people don’t care how they pay for a hotel stay as long as it costs the same), makes it easy to derive a relative price between LOC and USD. This is commonly referred to as a Purchasing Parity Price
• The highest discounted utility value (i.e. value today) comes around year 10 — $1.52, which, compared to the ICO price of $0.49 is a good upside as long as the team executes and the network achieves its target
• Current LOC price is $1.06, still below the $1.52 “fair value”, meaning that, in theory, an investor can achieve more than her target annual return of 40% by buying today and holding the token for 10 years when it should be worth ca. $44

Given the above, it makes sense to calculate how much of the anticipated future network development is already priced in the current price. This is a useful exercise because it provides a sanity check of the market valuation and expectation vs reality. Let’s stress some of the assumptions in order to see under what conditions the discounted utility value today will be equal or less than the market price:

• Velocity: if we remove the 2 weeks lock-up period before a stay, minimum velocity jumps to 81x, which decimates the current utility value — it falls to $0.37. This is big — right now, if you book on booking.com, no funds are locked, payment is made instantly. On AirBnB, the funds are locked from the time of reservation till the day after checkout. Depending on what model is used, the reservation time is a huge driver . To be fair, any increase in the 2 week period, would have a very positive effect on the utility value.
• Utilization of listings: I have assumed that Locktrip reaches the bookings/listings ratio of AirBnB (ca. 6.3x) in order to achieve the same number of bookings that AirBnB has today in 10 years. Changing the factor to 4.4x for example, would bring the maximum value to $1.05. Increasing it by 2 to 8.4x would increase the value to $2.00
• Market share: decreasing the long-term target market share of 3% (current AirBnB levels) to 2% results in a value of $1.02
• Discount rate: massive effect — dropping it to 25% would increase the maximum utility value to $5.24, whereas setting it at 50% would decrease it to $0.79
• Tokens issued: 25% of the issued tokens are held by the Founders. If we assume that they will not release them for 10 years (highly unlikely), the value increases to $1.35

Among the main value drivers is business development. Centralized marketplaces focus on the product and growth on their own. Funding and valuation come as a result of this. Decentralized ones need to nail incentives via the token. However, in doing so, they have the opportunity to do business development in a decentralized way and scale just as fast.

Locktrip needs to get from its current 100,000 listed properties to 4.5m in 10 years. Currently, they do not provide any meaningful incentive for listing aggregators to join the network and put their inventory on it. In my opinion, this is a huge opportunity missed. The key is to incentivize travelers to use the platform and suppliers to add listings via token economics. And, at the same time, ensure that token value increases with network growth.

But how?

Alternative token design models:

The token models below have been described at length on the net. Most if not all evolve around some sort of staking. The question is who is staking and why, what they get in return and what their incentive is. Here are a few ideas:

• Work tokens (e.g. like REP): the main idea is to stake tokens in order to do work — e.g. verify listings, provide insurance etc. For example, a user can stake LOC to verify the features of the home he/she is staying at and get a small discount for that. The host can provide part of his saved fee to the first X customers in order to rate him. The host has an incentive to enhance his listing and users get monetary value for reviews. Economics must be defined upfront to avoid cheating. Similarly, an insurer can stake LOC for the right to provide travel insurance, which enhances the value proposition of the whole platform. There are two factors to consider here: the providers of a service will receive some cashflow for their service and this can be valued with a standard NPV calculation. The platform would benefit from staking — if, for example, 10% of all tokens are permanently locked for staking purposes, the current utility value would increase to $1.69, with 25%, it jumps to $2.03. In the end, incentives are aligned for everyone to co-create the value
• Marketplace as a Token Curated Registry: both users and hosts want to be on a list. There could be independent lists (like the AirBnB double rating system, but with incentives) — users/ aggregators (e.g. agents) to curate hotels/hosts via ratings and hosts/hotels to curate users who behave nice, e.g. arrive and leave on time, don’t ruin the premises, don’t steal from the minibar, don’t smoke inside, buy stuff from the minibar and the hotel bar etc. Tokens are staked by the parties to be allowed on the list and remain staked while they are on the list. Cheaters lose their tokens and they are distributed to everyone else. The incentive for the travelers is that they may get better deals if they are on the list. Incentive for the hosts is that they get access to better customers by being on the list. Valuation aspects to consider are 1/ staking demand as % of issued tokens, 2/ NPV of tokens received from cheaters, 3/ NPV of discounts travelers receive and 4/ NPV of increased marginal spending by travelers
• Stake to govern: network participants stake LOC to receive governance rights: vote on how the platform is being developed, what token features are introduced, partnerships etc. This can be explored more as the token governance models evolve and voting economics is understood better. More relevant if a marketplace is built on a blockchain which already has such incentive structure (Decred for example) and the underlying token is used for that. Valuation effects would be in terms of staking demand
• Discount token: even if the whole point of using Locktrip is to not pay commissions, sometimes charging a small fee (1–3%) may unlock very interesting incentives. Similar to a loyalty program, travelers accumulate points when paying with LOC. Hosts pay the fee (in the form of a slightly increased price) and those fees are accumulated in a discount pool. Here is how it would work: a traveler buys 1,000 LOC, but he also uses Locktrip and contributes to the formation of the discount pool. The important points is that the incentives are to both own the token and use the service so that the network is grown. Let’s also assume that the network facilitates ca. $90m in bookings (our projection in year 1). Having 3% of that set into a public pool ($2.7m) allows the traveler to consume a discount. His 1,000 LOC out of the total 18,6m represent 0.0054% and equate to a share of ca. $145 in the total discount pool. By locking his tokens for say 30 days, the traveler can now use this amount for his next trip or other services which are sold for LOC on the platform. This way, travelers are incentivized to grow the network. Here comes the fun part — after X days, when the traveler has paid for her trip, her tokens get unlocked, not used up. If the discount pool is similar in size, the user can use the discount again by locking up LOC. To ensure that customers of the service who grow the network are properly incentivized by the token, the Foundation can use its 25% reserve to distribute tokens based on purchases. There can be a formula, similar to the example provided by Julien — trip value * network growth/2, to make sure that these 25% tokens are spent to incentivize X users and hosts, which would bring the network $Ym of annual value etc. The example above concerns travelers as an incentive to use the network and grow it. There is a similar exercise to be done for hosts/aggregators in order to add listings. Maybe the fee pools can be split equally and hosts can earn more money, earn free services, be able to grant better prices when they want etc. Everybody has an incentive to stake and use the pool money to grow and use the network. From a valuation standpoint, holders of the token will always find LOC more valuable when they use the service vs. only holding the token. The discounts can be quantified and valued with an NPV. Token staking effects can also be quantified as I showed in earlier examples.
• Premium discount tokens: a simpler model can be to ask hosts to stake LOC in order to receive premium features: e.g. get better listing ads on the platform (stake limitation for X days for the ad or something). If users stake LOC, they can get better deals for hotel/car/flight combos — e.g. these offers are there, but a discount is unlocked when the user stakes for some time (stay, flight time, car rental duration etc). The incentives here would be to receive/provide a better service by staking. A slight disadvantage of this model is that the staking is not strictly linked to growing the network by the parties.

The real value of an open source database of listings — AI?

I know we are still early in crypto life for this, but once the public chains get populated with structured data, many more applications will emerge that utilize these, obviously AI being on top of the list. One can imagine so many cool things being done that the value of these platforms may exceed the current centralized ones by a good factor.

But, before we get there, token incentives need to drive adoption and usage.

So good luck to all the teams.