I’ve written before about how my Stanford entrepreneurship students are often confused about the purpose of a Minimum Viable Product (MVP). I tell them that mediocre entrepreneurs build an MVP to show something, while great entrepreneurs build an MVP to learn something.
Here’s a case study that illustrates this point:
I recently met Suniya Sadullah Khan through my work with the Miller Center for Social Entrepreneurship. Suniya and her co-founder, Muhammad Mustafa, had an idea for a social venture that would help poor illiterate people in Pakistan get jobs as domestic workers (house cleaners, cooks, babysitters, etc). Their thesis was that the “gig economy” had created opportunity for lots of people around the world, but most gig economy jobs are app-based or online-based, and it’s hard to use those things if you can’t read. They believed they could change this by creating an online marketplace that included an app with a user interface that was designed for illiterate workers. Suniya and her co-founder began planning their venture, naming it Mauqa Online; “Mauqa” is the Urdu word for “Opportunity”.
Many founders would assume that the correct first step in developing the startup would be to build a Minimum Viable Product (MVP) of the app itself. In fact, that’s the direction Suniya and her co-founder were headed in as they were looking for a technical person who could help build an MVP of the app. But one day they attended a workshop on Lean Startup Methodology and Suniya says “That’s when it clicked — if we stopped thinking that our MVP had to be ‘tech’, we could start experimenting straight away to see what works and what doesn’t”.
Here’s some important background: The concept of ‘on demand services’ didn’t exist in Pakistan — Uber and its competitors had only launched a couple of years earlier, as had food delivery startups — so Pakistanis were slowly getting used to the whole idea of on-demand services. And the only way this venture would work would be if there were enough clients interested in using it to find on-demand domestic helpers.
Suniya and her co-founder believed that there would be, but one of the key tenets of Lean Startup Methodology is to test your hypotheses. So, Suniya says, “We decided to first test if our hypothesis of clients wanting ‘on demand domestic help’ was true. So we started a Facebook page and just posted in different groups — no advertising spend at all. A day later, we had a customer message saying she wanted someone to come and iron clothes for 3 hours. We didn’t have anyone so my co-founder went and did the job!”
As they continued to run their new “marketplace” as a Facebook page, they continued to gain key insights. They realized their pricing was too low, and so they adjusted that. They found out that some of the helpers didn’t have the soft skills that clients expected in a domestic worker and so they started doing soft skills training for helpers who signed up.
Soon they added a basic website linked to a Google form. Now the flow of new customers started to increase, and they began to experience some scaling pains. For example, in order to make sure that helpers arrived at a client’s house on time, Suniya would often ferry them around — an approach that was clearly not scalable. So they did a partnership with Careem (recently acquired by Uber for $3.1 billion) to handle transportation logistics.
Soon their bottleneck was they couldn’t recruit employees fast enough, so they increased their salary to above market rate, and offered to pay employees every two weeks — something that most employers in Pakistan don’t do.
In the aggregate, these learnings not only proved Suniya’s hypotheses but also provided invaluable learnings about how they needed to modify their business model. Learnings that they never would have gotten if they had started by building an MVP of the app itself.
“If we had started by building an MVP of the app”, Suniya says, “I honestly think our startup would have flopped. By instead launching it with manual processes, doing everything by hand, we started generating some revenue in our 2nd month and were constantly learning every day about operations, supply, demand and customer service points that we needed to address. We documented all of this and used the learning when we finally built our customer app, our backend engine, now for the helper app.”
Most founders with this startup idea — an app connecting domestic workers to people looking for domestic workers — would immediately assume that the MVP need to be a first version of the app. And they would fail. For Suniya and her co-founder they realized the MVP for the app needed to be an off-line version of the service that would provide them learnings and insights before building the app.
Here’s the way I think of it: before choosing what sort of MVP to build, think about what learnings are going to be most valuable in driving your venture forward.
In the case of Mauqa’s business, the question wasn’t “Could we build an app?”. Of course they could build the app. So creating an MVP of the app wouldn’t have proved anything. Instead, the question was “Was there marketplace potential in connecting poor illiterate people in Pakistan with clients wanting on-demand domestic work? And if so, what would the issues be in delivering a marketplace of that kind?”. Choosing to create a non-digital MVP, doing everything by hand, gave them the validated learning they needed to make their venture a success.
Four months ago Suniya landed a big funding round and Mauqa is now scaling up, adding two more big cities and improving the livelihoods of thousands of people across Pakistan.
All because they knew what kind of MVP to build in order to get the learnings required to build and scale a successful venture.
Remember: Mediocre entrepreneurs build an MVP to show something. Great entrepreneurs build an MVP to learn something.