Case study: Zen Educate’s marketplace
How Zen Educate kept supply and demand in balance to build a successful EdTech marketplace.
“It's constantly a challenge,” says Jono Hey, the recently departed Chief Product Officer of Zen Educate, reflecting on the problem of building an online marketplace.
“Think of a seesaw,” he explains. “If you've got more on one side than the other end, it will tilt. But what you want is for it to be perfectly level. You want the supply to equal the demand,” he pauses. “And, of course, at the same time, you want to grow the business. So you want the seesaw to get bigger and grow higher… but without tilting one way or the other too much at any time.”
This is the challenge that many EdTech businesses face. They are often balancing supply and demand. This could be in the form of learners, teachers, schools, employers, parents… depending on what need the product is solving.
Zen Educate’s marketplace
In Zen Educate’s case, they match schools - the demand - with supply teachers - the supply (the clue is in the name).
“I was sold on the concept the first time I spoke to some head teachers,” says Jono, remembering his first experiences of exploring the new opportunity.
“You just saw the pain that people went through every morning. The teachers hated it. The schools hated it. And it costs a lot of money in recruitment fees. It really felt like there was a big opportunity to make things better.”
He also reflects that, over their time at school, on average, a pupil spends a full year in front of a supply teacher. “So there's a really big benefit in not having that time be wasted.”
Seeing the opportunity, Jono left the financial services disruptor Nutmeg, where he had been Head of User Experience for 7-years and joined one of his ex-colleagues, Slava Kremerman.
Slava founded Zen Educate with Oren Cohen in 2017 after witnessing the archaic process a family member had been through to get a supply teaching job and discovering that recruiting agencies typically take 30-35%.
He saw that there was a big opportunity for schools to save money and teachers to earn more by smart use of technology. Since they began, Zen Educate estimates they have saved schools over £14m.
The fledgling team was now nine. “There was me and two developers,” remembers Jono. “We did a lot of things manually for our first customers. And we had a product that didn't really work as well as it needed to yet for the scale we were at. So there were loads of opportunities to make things better.”
So, where to start? How do you solve the classic chicken and egg problem?
Creating initial traction
Jono remembers early on reading a book called Platform Revolution that broke down the challenge of marketplaces into three components:
Liquidity
Matching
Trust
This became his guiding light.
“Liquidity is how well your supply matches your demand at any one time,” Jono explains. “It's fine to have loads of schools and it's fine to have loads of teachers, but if these things are lopsided, then it doesn't work at all. One side of the platform will say this is awful, and they’ll leave.”
He reflects that the challenge is very different from a direct-to-consumer business where you typically have a clear funnel and the challenge is normally about retention.
“It isn’t completely obvious,” he reflects. “You think you're doing a great job in one area. And then suddenly there’s a big shortage in another area. All these things make marketplaces a really interesting product to work on. And one that had me scratching my head many, many times throughout the journey.”
Fairly quickly though, they realised that the key to liquidity was identifying the right constraints.
Creating liquidity by constraining by geography
The first was geographical. Because their product provided a physical experience—a teacher needed to travel to a school—they had to create liquidity in a small geographical area.
“If we suddenly signed up five schools in West London and all of our teachers are in East London and don't want to travel every day to West London, then that's no use. You shoot yourself in the foot because the first experiences of the marketplace for those schools are really poor because you don't have the people to service their demand straight away.”
Quickly they refocused on very specific locations, building supply quickly followed by demand.
“We started neighbourhood by neighbourhood,” he says. “Teachers have a ‘shelf life’. Unless you give them a good experience early on, they will fade from the platform. And so it might look like you've got loads of users… but they're no longer active anymore because you weren't giving them enough work because your liquidity wasn't good enough during that critical period.”
Enabling matching by starting with general needs
The second useful constraint they identified was focusing on customers with broad needs who were able to make decisions.
“We quickly found that our sweet spot was schools where a head teacher or a deputy head can go ahead and make a decision. And so they can say, ‘Okay, brilliant, let's go try this new service’.” he says.
“Secondly, we realised we could get started more easily with schools that didn’t have very specific requirements. We could only service more unusual roles once we built up more liquidity on the platform. Then you have a much bigger pool and are more able to find somebody who's a perfect fit. Your matching can be much more specific.”
Building trust through good first experiences
Alongside tackling liquidity and matching early, Jono was also thinking about trust. He reflects that this is harder and less tangible. But they found some simple things that made a big impact.
“One of the things we did early on was we put teachers who consistently received good feedback into a special category. Then, when new schools joined, the first experience with our product would be with teachers that we knew would do a really good job,” he remembers.
“We found that if you give head teachers a good experience early, they will begin to trust the platform. And then they begin to use the platform more. You build trust in small increments over time by fulfilling your actions and your promises. But you can break it very quickly by sending somebody in who does a poor job or someone who doesn't even turn up.”
Growing the marketplace
I then ask him about growing the marketplace and the challenge of keeping it in balance. How did they approach this? What were the milestones, and when did things change?
“It was a while before we moved out of London because we wanted to get that density,” says Jono. “You don't want to launch new regions until they're self-sustaining. The great experience comes when you have a match of supply and demand and enough liquidity in an area. If you try and do a bit of everything, it'll just be rubbish for everybody.”
Building a flywheel neighbourhood-by-neighbourhood
He reckons that it’s about finding ways to provide the initial push to create network effects for each mini-market so that they gather their own momentum.
“In our case, the more demand you have, the better the matches are for teachers. So you get more teachers signing up, which gives a better experience for schools, which means they provide more demand. And so you create a flywheel. And if you start really local, you can start to build that flywheel in one area quite quickly.”
To start the flywheel, the team knew every teacher's name and started with many manual interventions. “For example, we might think, where are we gonna find a physics teacher in this area? Because we've got a school that needs one. Right, let’s go and do that.”
But gradually, the flywheel starts to spin... “And then, you find that you're not thinking about that at all, and it's just working. And you see these signs of the marketplace getting better by itself.”
He says that a bit like product-market fit, there is no one simple metric that told them it was working in an area; it was more seeing a range of signals. But once it was turning, it was time to move on to the next region. They started in Manchester, then Birmingham. And then gradually they moved into more of the centres of action around the UK.
Start by building supply
In general, they found that they should start by building supply in a new neighbourhood.
“The playbook for us was to build up teachers before you sign up schools,” says Jono. “Mostly because the relationships with the schools are so valuable and long-term. It was easier to keep the supply side warm while building up the demand side by adding schools to the platform.”
But he reflects that each situation was different: “Every time we did it, we thought we'd nailed it. And then we went to a new situation and found that we needed a different way.”
Local teams for big new markets
The most recent step on the journey has been moving into the US. “We were somewhat dismayed but also delighted to find the need was even greater and the system was even worse out there,” he says. “The US has grown significantly, and we're operating in five or six states now.”
Fortunately, the product didn't need to be as different in the US as they had originally thought. But it was still important to have local knowledge. “We built a new team to do that. In different countries, you have very different requirements for the kind of checks you might need to do on teachers. And, of course, in the States, every state is different.”
Balancing markets with different maturity
Despite retaining the same underlying product, adding more and more regions added complexity. To begin with, regions need to cope with low liquidity and provide general matches. As time goes on, you want them to be able to take advantage of volume and cater for more precise requirements.
“At any one point, you've got all of these different nodes which are in different places in the journey,” says Jono. “The product needs to be different when you launch in Wales or when you launch in California than it does in London, which is at a more mature stage.”
This can also happen in a region but with a new type of customer.
“You've got a lot of density in London, but now you're working with nurseries, and you don't have many nurseries on the platform. So you have a small side set, which needs to operate differently from the rest.”
Adjacent opportunities
As well as growing the core offering, the team spotted adjacent opportunities and had to work out how to respond.
Long term cover
“What happened very quickly was schools said, ‘We like the people that you send in’,” remembers Jono. “‘We've got somebody going on maternity leave from Christmas, and we need somebody for the next two terms. Do you have anyone?’ So, we started doing long-term placements as well. If we didn't do this, they'd have gone somewhere else. So it was a no-brainer.”
However, initially, they supported it manually to understand the opportunity. “We ran it with spreadsheets and CRM systems for as long as we could,” says Jono. “Until we had the capacity as a team to focus on supporting it within the product.”
Pandemic and a new market
Then the pandemic happened. Schools closed, and the team looked for new opportunities. “We spun out a sister product for working with opticians and pharmacists. It's called Zen Locum and does a very similar thing. They have a similar need but in a very different market.”
They built that product out separately with a small team. They learned a lot that they could apply back to the education product. “But there were a lot of things that are just different about that marketplace,” says Jono. “I’m really glad we did it as a separate product rather than trying to bolt it on to the education one.”
Marketplace metrics
Given how much more complex it is than a direct-to-consumer business, I ask about what metrics they found helpful.
“It's very easy to spawn hundreds of metrics and drown in analysis,” he warns. “But our guide for the business was quite simple: the standard metric in the industry as a whole is Days Booked.”
He says that, unlike a digital product like YouTube, it’s not all about engagement. “Ideally, a head teacher spends 10 seconds on the product booking cover and then rates the person who came to the school with five stars.”
The other core metrics were related to measuring liquidity: “Fill rate and fill speed. What percentage of the demand do we fill? And how quickly can we fill those?”
But even with those, they realised it was a bit more subtle. “Fill speed is a really interesting one,” he says. “If it's 6.30 am in the morning, and you need someone in school at 8.30 am, you want it filled within 10 minutes. But if you're looking for somebody next week, you probably don't mind if it takes a day. You’d prefer to spend longer if it might mean a better match.”
“And so we then subdivided these into fill rate and speed for requests made within eight hours of the role start, versus roles starting a day out, a week out, two months out…”
Organising teams around a marketplace
Finally, we talk about how to design and grow a team building a marketplace. “We experimented as we went,” he says.
“In the beginning, we were a very small product team. We had one team focusing on liquidity, which was about bringing on teachers quickly. And one team working on improving our matching.”
He says that this worked well in the early days because to make an improvement for one group of users, you often have to make changes for other groups.
“A really simple example might be that a teacher wants to know what the normal start time of the school is. Is it 8.30am or 9am? But to do that, you have to build something so that schools can add and set their standard hours for the school. And then those schools have to go in and edit that information. Or maybe your account managers have to go do it? So most features we found had two or three sides to it.”
Later, as they grew, they switched to teams focused on specific users.
“If you're a product manager, to know what are the right things to do in your product area, you need to speak with not only schools and teachers but also the team leads in different areas on the teacher side, on the school side… and suddenly you find that everybody's talking with everybody.”
So now they organise around teams focused on teachers, schools or serving internal teams. “It's great from a product manager and a knowledge perspective, because you can really immerse yourself in the needs of the school or what are the needs of our teachers. But if you need to make changes, you might need the other team to make those changes.”
He reflects. “My guess is that each marketplace business does it slightly differently, but generally, people probably settle on a user-focused approach. I think that’s because otherwise, as a product manager, you're splitting your focus and aren’t really on top of the needs of each of the types of users that your area of the product is serving.”
Finally, for the adjacent opportunities like long term cover, they created dedicated new teams.
Key takeaways
We recap the learnings that could be applied to other EdTech marketplaces.
Define your supply and demand and keep them in balance. For Zen Educate, teachers are the supply, and schools create the demand.
Begin by constraining the marketplace to create liquidity. Because Zen Educate provides a physical experience, this means geography. It might be different for you.
Start with customers with general needs that you can easily service with a limited supply. Only try to match specific demands once you have built up liquidity.
Don’t remove the constraints or move to new markets until you have created network effects and got the flywheel to turn on its own. This might require manual interventions to begin with.
When moving to a new market, start by building supply but then quickly create initial demand. Your supply is likely to have a ‘shelf life’.
Your North Star metric is likely to be the transaction between the two sides of the marketplace. For Zen Educate, it was Days Booked. You may also want to look at metrics that reflect liquidity.
To begin with, organise around solving problems like liquidity and matching. Later, you may prefer to focus teams on understanding different user needs.
I ask if he has any final advice for others in a similar situation. “Don't be precious about building the perfect system. Just get started,” he says without hesitation.
“You can start having an impact immediately with just spreadsheets and a landing page. And that will help you figure out what your product needs to be. It's really easy to say, our product needs to do this and it’s going to be beautiful. But we were providing value to the market just by getting started. The technology was secondary. Ultimately transformational. But originally, secondary. So, don't be precious about it. Just get going. That'll help you more than trying to figure out and build the perfect product straight away.”
Jono is now taking some time out to promote his book Big Ideas Little Pictures, a collection of his Sketchplanations. These are amazing little drawings explaining the kinds of concepts we’ve explored above. There are a lot of product sketches—you should check them out!
Further reading recommendations
Jono also shared four things he would give to all his new product managers that he believes helps people understand the problem of marketplaces:
Platform Revolution, a book all about building marketplaces, which explains liquidity, matching and trust.
An article by Bill Gurley. All Markets Are Not Created Equal: ten factors to consider when evaluating digital marketplaces. “I came to realise that a lot of our success was due to these little quirks of our marketplace being really well suited for a product like ours. And I saw many other marketplaces start up in different areas that struggled for various reasons. I think his list is really a really good one.”
The first Masters of Scale podcast episode with Brian Chesky. He talks about how they got Airbnb started. “There's a few really good lessons. One of them is to do things that scale, you have to do things that don't scale. Like an early story of them hiring a photographer to go take nice pictures of everybody's apartments. Do things until it hurts and then automate it away.”
An article called The AI Agency by Tom Thomas Tongas. “He suggests starting by doing what the traditional players are doing and then keep applying technology to get that extra advantage. And before long, you're not having to do all the manual stuff, and it's the technology doing it. I think that's kinda what we followed at Zen.”