Hacker News Reader

Comment by sama | Hacker News Reader

Hi Zhuang and Derrick--

I think that 98% of the time, the right answer is to start out and remain focused on one vertical until you take it over. Most startups expand far too early and then fade out everywhere (this is often the fault of bad advice from investors). Facebook started in an extremely small vertical (Harvard undergrads), then expanded to a still-small vertical (US colleges), and then at some point took over the Internet.

Why did you pick this vertical to start with? It seems there are a lot of services helping people get restaurant reservations. The idea of breaking down tasks into small action items and farming them out to people seems good, but this doesn't seem like the most pressing need. What else did you consider as a first vertical?

How many users do you have, how much do they pay you (and how much do you pay the agents per task), and how fast are you growing?

View on HN

We picked a reservation concierge service first, as it is a vertical that is immediately testable in SF with easy to understand mechanics for consumers. With the increasing mass-media fueled foodie culture, our hypothesis being that people don’t want the trouble of deciding where to eat and making the reservation - they just wanted a curated experience. And that made it a good candidate for our platform.

The other 2 verticals we considered were 1) a marketplace for on-the-ground, region specific collection of data and 2) travel planning in countries with low internet penetration, namely bookings and suggestions by locals.

With Loft, we have 20 customers paying us $10 ~ $20 a month, but we haven’t grown beyond that due to limited marketing, as we focus on nailing the operations side of platform. The agents are currently just the both of us, and we can afford $10 - $15 an hour per agent if our economics hold.

All in all, we have our vision of what the future of jobs can be. Loft is an experiment towards that which has seen mild traction, and that’s why we really want your advice!

If the main thing you want to prove out is that you can build a platform where small tasks can be easily farmed out (which again, sounds like a really good idea), I'm very unsure this is the right way to start. It's well established that you can get people to make restaurant reservations for you.

The best companies do something either fundamentally new or 10x better than existing solutions.

As you think about the future of the world, what can't people do today that they should be able to and that your platform could allow?

Thanks Sam, really enlightening points.

I realize we might have actually gone about it the wrong way. From a consumer perspective, that’s where we were struggling to find a niche to start off with. Our long term goal is to enable people to buy services from anyone in the world, no matter how small or large (i.e. reservations, travel), and it’s on us to make it actionable by as broad an workforce as possible.

We felt we had to prove out the concept of farming out discrete tasks in a straightforward niche like the concierge-space, as opposed to perhaps a white labeled platform. That probably wasn’t the right approach, and we’d appreciate your thoughts on this.

I still think picking an initial niche is a good one, but ideally you'd pick one that represents a complex task that is easier to get done in broken-down subsets, and that customers don't have an easy way to get done today.

Gotcha, that's a really good way of putting it. And I agree that reservations as tasks are not the way to go to prove the concept out.

Seems like we have quite a bit of thinking to do. Any further advice for us? Thanks!

i know of a company doing something similar in hong kong. their niche is to get people in various locations aggregate data from brands by going to retail stores and take photos of their products. each user gets 3-10usd for doing this and the brands save on sending expensive people to each location.

Reply on news.ycombinator.com