Hacker News Reader

Comment by cultofmetatron | Hacker News Reader

on year 7 of my startup and profitable. I'd like to add a few things.

1. look for a boring problem that needs solving. related to the product market fit. a lot of founders are trying to build a sexy solution to a problem that no one has.

2. Sales is half the equation!. you can have the best engineering team in the world but it means nothing if you dont have good sales. and good sales people don't just get sales. they are the first point of contact for your prospective customers. Empower them to get feedback and ideas for how to evolve your product. half of our features at my startup are things we never would have thought of until our sales guys kept telling us how many restaurants were requesting the feature.

View on HN

> half of our features at my startup are things we never would have thought of until our sales guys kept telling us how many restaurants were requesting the feature

As a developer in a smaller company I like to chat with both support and sales due to this.

Could be at lunch or by the water cooler, but just gently probe what they're struggling with or what our customers are telling them.

Often there are low hanging fruits that have significant customer impact for fairly minimal development effort.

1. look for a boring problem that needs solving.

Even that is not boring enough I reckon. It is totally fine to find a problem that already has a solution and just execute better on delivering the solution than your competitors. Many (most?) businesses start out this way yet I've heard plenty of times that somebody will not start a business until they find a problem that needs a solution.

Add to that, lots of tech is "over-priced" these days. There are a lot of PMF validated out there and you can compete on price. I don't understand the reluctance to do that.

Hard to make that work when all your competitors have VC funding. More generally, this will normally make your unit economics look worse, which can limit your valuation.

Obviously, in the real (non-VC) world, price wars are a great strategy if you can provide the service more cheaply.

competing on price is generally a bad idea. its an inevitable race to the bottom. Plus now you're running on less profit which limits yoru ability to grow. you're much better off adding features that your competitors don't have or optimizing for a subset of the userbase that would prefer a simplified workflow.

cultofmetatron is hiding an grace/ability ratio of an olympic ice skater. My friend worked for a POS company where potential customers had features in mind and engineers were expected to place the next block on the barely standing Jenga tower. They were bootstrapped, had excellent moonlighters and years of headstart in Indian market. They were dead by the time most shops got Android POSs.

one of the best decisions I made as CTO was to focus purely on IOS devices. That massive moving target of android devices is a nightmare to develop for if you're going for bug free turkey operation. Second was moving to hire someone with "olympic ice skater" level experience with ios development. We tell our customers what devices to get and in B2b, its a very small part of their overall budget.

The second best decision was to build our server infrastructure on top of elixir/phoenix. I don't think I could have gotten this far if we used node (and I had 10 years of nodejs experience)

Interesting. Do you do anything “special” on the server thats hard / impossible to do with node js?

well for one, in elixir, you can get away with having a monolith architecture. elixir has the concept of a "genserver" which is a dedicated thread. the beam has a message passing system baked in. So lets say I want to create a service to which I want to send jobs, I can create a genserver which is just a file taht inherits genserver behavior and mount it at the top level's application.ex in the supervision tree. phoenix will take care of keeping it up and monitoring it in place of crashes. You can also have your api nodes connect to each other and maintain persistent data across them. this means a thread in one machine can send data to a thread on a completely different machine.

If I were to do this in nodejs, I would need to setup a supervisor. I'd need a message passing broker like rediis or rabbitmq and I'd need an orchestrator to manage the various services. in elixir, we get all this baked in. I can add a microservice as easily as you can add a controller in rails.

Also, the websocket system in elixir is so good makes nodejs look like a toy.

Hey would love to learn things on two on the traits you look in a sales person. I have been selling myself as a founder last 8 years, now want to hire someone.

What does your company do?

cloud based restaurant POS

Isn't that super saturated and commoditized? How do you compete against the likes of toast?

I listed the two main ones in my original post. everything else is a competitive advantage we prefer to keep secret.

I can tell you, from the other side of the house (the eater), that there is still a giant yawning gap in restaurant software. Toast must have great sales people, or some compelling meal ticket manager or something, because the experience given to the diner is absolutely trash.

you can always build a solution to a problem only one billionaire has though as long as you can convince them to part with their money.

Reply on news.ycombinator.com