Distributed Trust Marketplaces: The Next Step in Marketplace Evolution
27 Nov 2023
Commentary and suggestions on Li Jin and Andrew Chen’s “What’s next for marketplace startups?”
Following Li Jin and Andrew Chen’s 2018 article about the evolution of marketplaces, we would like to suggest a different direction for marketplaces. Instead of regulated services being the new El Dorado for marketplace startups, we’d suggest that distributed trust discovery mechanisms are where the puck is moving. We should be building on the idea of trust that managed marketplaces have added to the ecosystem in the latest era, by assigning that trust to parties better-suited to catalysing demand.
First, let’s characterise the evolution of marketplaces thus far, starting with Jin and Chen’s famous timeline:
The Agora Era
(Our little addition to the OG analysis by Jin and Chen)
Before digital marketplaces were a thing - for all of history, in fact - marketplaces were physical places, town squares, where buyers and sellers of anything and everything would congregate to transact. In ancient Greece, this was called the Agora, translated as ‘meeting place’.
Post the industrial Revolution, this became something like the Yellow Pages and newspaper classifieds, where people would find the goods and services that they needed, reach out, and potentially transact… The Yellow Pages Era, if you will. You could hack this system literally by calling your company “AAA Locksmiths” to be placed at the top of the alphabetically-ordered list. It was unsophisticated, but effective.
Trust between transacting parties in this era was your problem to deal with. Look the person in the eye, speak to them on the phone, make a judgement call.
The Listings Era
The arrival of the internet age quickly brought about platforms like Craigslist, which made discovering service providers significantly easier than flicking through the Yellow Pages. But it was messy - providers in every domain, of every level of sophistication, price, quality and otherwise were shoved into a single, overly-large database.
Trust in this era was still your problem. Trust yourself to know what you need, who you want to deliver it to you, and whether they are deserving of your trust.
The Unbundled Craigslist Era
The next thing to happen was that the listings companies began to separate out horizontally to allow vertical differentiation. You went to Zillow for housing, Indeed for jobs, Tinder for dates, Etsy for craft goods, etcetera. This enabled marketplaces to provide significantly better search and discovery, because they could triage providers according to more nuanced characteristics rather than macro categorisation. They could also add more value-add features like purpose-built booking and communication with sellers.
Trust in this era started to shift, in part, to the platform itself. The platform would do some vetting and have some sort of quality control, so the message became ‘trust yourself to pick the right vendor, from a list of vendors we’ve curated and vetted’.
Source: https://a16z.com/whats-next-for-marketplace-startups/ (original by Andrew Parker, 2010)
The ‘Uber for X’ Era
Perhaps better termed the ‘On-Demand Era’, Uber and its contemporaries inspired a rush to provide all these goods and services immediately. It was the perfect differentiator for automic, commoditized services. Supply became abstracted and fungible - “an Uber driver”.
Liberal application of the on-demand value proposition failed, however, due to ill-suited services… You need massages, IT assistance and babysitters a lot less frequently than a taxi, and the value proposition fades into nothingness when your masseuse ‘will be right with you, in 72 - 96 hours’. You also care a lot more about the nuances of your babysitter than your taxi driver, so ‘a babysitter’ isn’t good enough, as it is with ‘an Uber driver’.
Trust in this era was abstracted away, close to irrelevant. I just need ‘a taxi’, I don’t care who they are as long as they have a driver's licence and anything North of a 3 star rating.
The Managed Marketplace Era
Next up was an attempt to go full-stack, to tackle the value chain end-to-end. Marketplaces now take on vetting and oversight of the service (e.g. elder care), such that you find, communicate with, book, pay, oversee delivery and review on-platform. Companies operating this way must work to build a brand and deliver more to you than other companies in order to gain your trust.
Trust in this era is assigned to platform, the message is ‘we’ve got the best people and the best processes, trust us’.
Jin and Chen proposed that services will be the next thing to be platformed by marketplaces - more specifically regulated services with a legal bar to cross to be a certified supplier (healthcare, law, accounting, etc). Regulated services are easy to triage from a discovery standpoint, because they’re standardised. So although complicated (more so than goods, for that matter), it is possible to ‘map’ the value delivery process of these types of services, and thus cater to their purchase on a platform.
We haven’t seen this play out in any meaningful way in the five years since that article was written, and we’d like to suggest that the reason is that certification is not how purchases in the world of services are made. Trust is how service transactions are catalysed, and having a certification doesn’t move the dial on trust, sufficiently.
Unlike the world of goods, where you can easily compare the spec sheet of the thing you’re purchasing, services are much more complex. Jin and Chen discussed this challenge, as well as three others. Here they are, paraphrased:
Services are much more complex in nature, compared to goods
Quality/success is much more subjective with services vs goods
Fragmentation - services are typically supplied by many small vendors with limited tools and varied offerings, vs the more centralised nature of goods suppliers
Service delivery typically relies on real-world interaction, which limits platformability
The first three above can all be traced back to trust…
Because services are more complex than goods, it takes more to get me to trust and therefore transact with a service provider.
Because the quality of the service is really my subjective opinion, it takes a lot more to get me to trust and therefore transact.
And because services are typically supplied by small, independent vendors rather than standardised monoliths, it takes a lot more for me to trust and transact.
The last challenge is a logistical one, which can only be whittled at with ingenuity and execution.
Because of the reasons above that make services complicated and obfuscated, people buy services based on referral, not merit.
How many marketing agencies have been hired by companies who picked the most highly-rated social media agency on Google vs ‘the one that Jimmy used back at XYZ’? How many consultants have been contracted because some executive ‘knows a guy who knows a guy’? How many business coaches, housekeepers, therapists, attorneys, cat sitters, garden services and architects have been hired through referral over formulaic discovery?
A lot more purchases have been made on referral than on merit, in the world of services.
I don’t need A therapist… I need one that I feel comfortable with, that understands how I do things, engages with ‘my type’ of people.
I don’t need A lawyer… I need one that helped my friend get out of a similar jam, that operates like that attorney I worked with back at that other company, who I know won’t bill me 45 minutes for a 3-liner email.
I don’t need A service provider. I need one that I feel comfortable with. Limbic resonance is so much more prevalent with services as compared to goods because services are personal, impactful, whereas things can easily be replaced.
This is how services work. Limbically. Feelings. Referral.
The Distributed Trust Era
So if referral is the de facto discovery mechanism for services, how might you engender referrals en masse?
At Kintro, we think the answer is community.
Trust is graded by identity and familiarity. We trust a handful of people that are very close to us implicitly. We trust our broader friend and family circles a lot. Friends of friends, and colleagues of colleagues, a bit less. Fellow cross-country skiers from our neighbouring country a bit less. And so on until we get to the person who lives in a completely different country, speaks a different language we’ve never heard before, dislikes all the things we like and does something completely obscure for a living… This person’s recommendation means very little to us, unless perhaps you’re looking for the best local water house while on holiday in their town.
If it were possible to capture this trust, this social capital, on a platform, then a) it would open the floodgates for a 10x better matching mechanism to connect buyers and sellers of services, and b) the whole user experience of discovering and purchasing services would be elevated, because tapping into people you already trust to figure out who to purchase from is far more natural than toggling advanced search filters or intuiting how biassed a platforms LLM-powered chatbot is.
To put it plainly, if we could merge marketplaces with community, we could platform a lot more services.
Trust in this world would be assigned to the individual people you already trust. Trust would be distributed.
You could ask your community of parents who the best babysitter is for Friday’s date night and get a referral to a service provider that the people you trust, trust. Rather than filtering through a database on an app, reading reviews from Anonymous217 and making a judgement call.
You could crowdsource recommendations for a business coach from your startup members club, which you joined precisely because of the calibre of its members, and trust their introductions by virtue of the fact that you trust them.
Moving the burden of trust from the platform to the individuals in your community appears to us to be the next step in the evolution of marketplaces, and we’re excited to participate in the change.