Please learn from our mistakes

No-bullshit lessons in business and careers. One mail every day. 15k+ readers love it. Join in?

Oops! Something went wrong while submitting the form.
29 Jun

The leaky marketplace hypothesis

Recently, I noticed a pattern emerge from all my experiences of shopping from online marketplaces. I’ll share a few of my experiences and observations, and attempt to try to make sense of what exactly is going on.

A good friend recently renovated his house. I was surprised to discover that the contractor who designed the house actually worked part-time with Urban Company (UC). My friend told me that soon after the first service via UC, the carpenter suggested I reach out to him instead of using UC.

Coincidentally, my wife found her favorite beautician also because of UC and reached out to her directly for the subsequent visits.

Turns out, this is a common practice. A consumer, at first, avails of a service from a UC technician, and then if their work is good, reaches out to the technician directly instead of using the UC app.

Another instance of the same kind took place when I and a friend booked a villa on Airbnb. He urged me to book the villa for just a day, get the host’s contact details, and directly reach out to him to book for the remaining days of the trip. He reassured me that it worked.

But this time, we were at the receiving end of the ordeal. As soon as we reached the villa we discovered that the host forgot to update the availability on his Airbnb dashboard, and took up another booking for the same dates. We eventually had to shell out a lot more in spite of all the street smartness. Because the host was obligated to Airbnb, he had to cancel our bookings and serve them instead.

Well, Airbnb does give a direct warning against moving the transaction or the communication off the platform:

Had I paid a little more attention, I could’ve avoided the fiasco that ensued after booking directly.

From both the instances above, do you notice what’s common between Urban Company and Airbnb?

Both UC and Airbnb are service-centric marketplaces. Unlike in the case of Amazon or Flipkart — who sell commodities — scouting technicians for odd jobs or connecting travellers to hosts involves facilitating interactions between buyers and suppliers of the service.

But facilitating these interactions exposes the business to leakages where both buyer and seller collude to bypass the platform and transact outside of it, in order to save on platform commission fees.

In fact, I would go so far as to say that these interactions between buyer and seller make leakages inevitable.

If you’re wondering why it is inevitable, let’s quickly go over how buyers and sellers interact in other marketplaces,

E-commerce marketplaces

On Amazon, I search for the product, add a filter or two, and then scroll through four or five products before I finally add one of them to my cart. Often, there is no need for me to interact with the seller.

On Amazon, even if you try to reach the seller, an Amazon-mediated servicing agent responds on behalf of the seller. After delivery, if there is some issue, the return order is also picked up by the Amazon pick-up executives. I don’t recollect any instances where I spent time trying to reach the seller directly.

Additionally, the number of people positioned between the main buyer and seller to ensure the product goes from point A to B also makes it difficult to trace who a buyer can reach out to.

There is a thick and seemingly impenetrable layer that keeps the buyer from directly engaging with the seller.

And this friction is fully intentional: by shielding customers from sellers and acting as an intermediary in any step, Amazon or Flipkart can ensure they're not in the blind about any buyer-selller correspondence happening off the platform. In this way, not only do they manage to capture have of the data around a customer's shopping experience, they can also avoid marketplace leakages.

Food delivery apps

Can you recollect what happens when you order food from Swiggy or Zomato? As buyers, we are fully aware of the significant price differences between Swiggy and the restaurant. Oddly though, it still isn’t enough of an incentive for us to order directly from the restaurant.

I've seen many restaurants printing QR codes on their packaging, incentivizing the customer to directly buy from the restaurant. It creates a win-win situation for both the customer and the restaurant: the restaurant can save on platform commissions and the customer enjoys a good discount on the normal Swiggy listing price.

But I still don't find myself availing this benefit.


I think this could be because the platform abstracts away all the logistical details and provides a super-convenient interface. You see a product and it arrives at your doorstep. Everything in between is conveniently hidden away from you. Both Amazon and Swiggy take care of all the processes in between. There is barely any need for you to interact with the seller.

Additionally, it's also about the frequency of the use case.

If I'm consistently ordering from a single restaurant every day, I'm more incentivized to completely bypass Swiggy and deal with the restaurant directly. But this isn't usually the case. Generally, I'm on Swiggy for the discovery it offers. I do not order from the same restaurant every day. So, going outside the platform to place the order directly with the restaurant seems frictionful — even if that option is available.

The ease of buying creates lock-in behaviour. The lock-in further disincentivises both the buyer and seller against taking the transaction offline. Doing so would both increase the hassle to finish a simple transaction, and the logistical effort involved.

Also, Amazon or Swiggy know that a transaction will probably not take place without the services they enable as intermediaries. Not every restaurant has its own logistics. And not many buyers are sufficiently incentivized to go out of their way and have 10 different apps for 10 different restaurants. They are habituated to the umbrella offering and convenience these platforms provide.

Does that mean we are not habituated enough to the use of Urban Company or Airbnb?

No, I don’t think that’s the conclusion here. Urban Company and Airbnb serve specific needs which only they can. When planning a trip, I will look for stays on Airbnb or when looking for someone to quickly fix my washing machine I will use Urban Company.

The habit to use the app exists but there are two factors which I sense create enough incentive for buyers and sellers to exploit loopholes in a service-based marketplace.

1. The business model of a service marketplace

If you’re a service marketplace, where your product is the service provided by involving another person (Urban Company technician or house host), your business cannot afford to create too much friction between the buyer and seller.

If there is too much friction in the interaction, people will either not use the app or service providers won’t have enough incentive to stay because you’ve made their life complicated.

2. The frequency of usage

I order food almost every other week, and keep buying something or the other from Amazon a few times every month. Anytime I’m at home and want something delivered to my place, I don’t have to put a thought before I click on these apps.

In the case of UC, my wife’s frequency of availing beauty services isn’t that much, although periodic. She will avail these services maybe once a month or once every two months. That frequency when booking that holiday beach villa is even lesser.

Even though I am habituated to using the app, the number of times I do so is conditional. A washing machine breakdown, a pipe leakage, and a vacation don’t happen as regularly.

This becomes a barrier to creating the kind of dependency one usually has on an e-commerce marketplace.

And here's the double-whammy:

Unlike Amazon, where the seller doesn't really matter as long as I'm getting the product — or Swiggy — where discovery and exploration is one of my primary reasons to use the app, if I like a plumber, a barber or a beautician's service, I want to build a relationship with them. I don't want to be dealing with a new barber every month.

Especially with service personnel who are visiting your house, trust is an important factor. This further incentivizes the customer to ask for the seller's personal contact details if they love their service and leak out of the marketplace.

Having said that, there are ways in which UC and Airbnb make an attempt to keep these leakages in check.

For instance, on UC, before I made the payment, this screen popped up. Now carefully read the second point in the disclaimer.

Urban Company also ensures that the contact details of the professional are shared only 60 minutes prior to the service. But this doesn’t safeguard the platform in any way because, ultimately, the technician has to go to the buyer's house to provide the service booked via the platform. And UC cannot gate-keep conversations at that stage.

Airbnb also takes certain steps to ensure that contact between buyers and sellers is only established after part of the payment or full payment is made via the platform. But regular users of the platform are quick to find loopholes even with the stringent checks put in place by the platform.

For example, I tried to send a message to the host, whose villa I wanted to book.

The first notification showed me the option to send a message directly. But read the following copy:

Slight nudge by Airbnb to prevent me from potentially leaking out of the marketplace? Perhaps.

But this isn't the only thing the platform does.

The platform also auto-detects phone numbers and email addresses in the chat and blurs them out.But of course, customers have found easy loopholes to bypass these checks. Some break down the phone number in words, others use vernacular languages. Here's an example hack I found on a random website.

Ultimately, the only strong nudge Airbnb can implement in order to disincentivize the customer from leaking out is by scaring them.

But is all this enough to stop a customer from taking the conversation with the seller offline?

I am not sure to be honest. My own experience says no. In spite of all the direct and indirect disclaimers provided all along the booking process, what happens once the contact details of the service provider are shared is beyond the control of a marketplace.

Many a time, travellers choose to book the first few days of their stay via Airbnb and then extend it for a few more days, but off the platform. Their subsequent stays at the same place also get booked off the platform.

So then, what makes leaky marketplaces sticky enough to thrive?

Discovery and credibility is the real proposition that a service-based marketplace provides.

For the carpenter to actually land a full-fledged interior project, their finesse, quality of work, and communication have to be of a certain level to actually land a bigger project. Being vetted by Urban Company provides that edge to a service provider. It also serves as an umbrella of trust for the consumer.

Secondly, even if the service provider is successful in forming a personal relationship with you, a buyer’s need for their service won’t suddenly increase. They will still need regular business which only UC can help them with. UC still acts like a distribution platform and continues to serve as the top-of-funnel lead generator.

Even though leakages happen, they may not be significant enough to become a threat to the marketplace. Especially in an age where one extra click on a form can drop submissions by >50%, expecting the customer to go out of their way to contact sellers directly to save a few bucks is perhaps expecting too much. Of course, this is much more valid for the younger generation than older folks, who will still hunt for coupons and rejoice in a rare ₹5 cashback.

Maybe service marketplaces make everything convenient enough so that even if you can do without them, you’d rather not.

Feeling Lucky?
Subscribe to get new posts emailed to you, daily. No spam.
Oops! Something went wrong while submitting the form.
15k+ business professionals act on our advice every day. You should too.
Subscribe to get new posts emailed to you, daily. No spam.
Oops! Something went wrong while submitting the form.
15k+ business professionals act on our advice every day. You should too.