Ask HN: What problem in your industry is a potential startup?

44 points by anthropodie ↗ HN
This has been asked previously here [1] but it has been over a year. The thread actually has suggestions to do this monthly to get insights into other industries but I guess that did not happen. Anyway I think it's been over a year and I think we should have a discussion about this.

[1]: https://news.ycombinator.com/item?id=13139638

34 comments

[ 0.26 ms ] story [ 74.8 ms ] thread
The question is valid, but there is something missing from the equation from a startup perspective: experience.

It is the person who has experienced the problem that will fully understand the nuances of the problem and the potential users of the solution.

An alternative question might be: “how does one engage and learn about problems in your industry?”

I understand that this is difficult, but this is what investors want to see.

There are many billing solutions geared toward SaaS providers, but not many for service businesses that bill monthly (at least in the API sense).
Could you elaborate please?
Our product is performance-based, rather than unit-based, and there's no exposed API. We would like to have a detailed itemized invoice (think a phone bill). Our options are build from scratch or shoe-horn existing solutions. It'd be nice to not have to reinvent the wheel and also keep our customers within our UI.
Could you provide some examples of performance-based billable services?
Our company audits shipping invoices and bills the customer based on a percentage of what we save them. It's a common business model outside of the tech/SaaS landscape (utility bills, property taxes, etc)
I guess there's a lot of complexity I'm missing. Is it something like this:

1. Customer original bill (variable - user provided) e.g. 100

2. Saving (variable - user provided) e.g. 9

3. Saving percentage (calculated) e.g. 9%

4. Billed Percentage (variable based on savings percentage?) e.g. 25% x 9

So you would have a billed percentage "rate card" that depends on variable factors like the customer bill plan, the amount/percentage saved etc? Is that the tricky bit?

Do you mean a billing service that connects to some internal project/time tracking software and automatically builds and sends an invoice?
No. I'm talking about service that is performance-based, rather than time or cost based.
metronome.com
It's closer than I've seen elsewhere, but not quite there. There isn't a good way to predefine what they consider "billable metrics" for the use case I'm considering.
Something like Kable perhaps? https://www.kable.io/
No, this looks like a product for companies who sell API access.

Our product is performance-based, rather than unit-based, and there's no exposed API. We would like to have a detailed itemized invoice (think a phone bill). Our options are build from scratch or shoe-horn existing solutions. It'd be nice to not have to reinvent the wheel and also keep our customers within our UI.

How many items are we talking about on an average invoice?
Technical debt. How do you quantify it? Which tech debt to work on first. Can static analysis tools identify it?
Test Environments and Test Data in big enterprises. Scaling from small to huge is as challenging as building the product itself. Spinning new environment on the fly that includes 100's of subsystems, creating an anonymized subset copy of production on demand or generating enough synthetic data on the fly are huge challenges even though some of the tooling exists.

A related problem related to data creation and subsetting is that one person (and practically nobody) knows the entire data relationships between subsystems, but you still need them to create your data.

Designing a kitchen requires a human designer. If you could feed a cabinet and benchtop range into a design AI and allow people to use the AI online to design their ideal kitchen, you wouldn't need a designer.

This would enable DIY kitchens to be sold to a much larger market. Alternatively small kitchen installers could handle the sales and installation. Manufacturing can already be done by many 3rd parties. The expensive part is in design.

Isn’t that what IKEA already does, without requiring AI? All you need is a design tool that lets you plop the cabinets and appliances where you want them.

Maybe you can add a simulator that estimates the efficiency of a particular layout. For example I felt I needed more floor space, but in retrospect I have to move around a lot more to reach things, making my layout less efficient. A simulator could have told me to put the counters closer to each other.

Yeah Ikea has cornered that low end market. There is a big market for that and lots that are happy to learn how to design. There is also a big market that is not willing. They want to tweak but need a lot of hand holding.

Thanks to Ikea there is not much point trying to compete in that space.

Mid range cabinetry still needs more customer support in the design time.

Also Ikea cabinet construction is not great. I'd recommend not using them in your house. Maybe in a cheap rental.

I get that, I just don’t get why you can’t use the same tool but with the upmarket ploppables.

Adding the aforementioned usage simulator and maybe using it to suggest starting layouts doesn’t seem like it would depend on the market positioning of the cabinet company, surely it would be just as applicable for low end as for midrange.

Have you not been to Home Depot or Lowe's recently? Or IKEA?

Let alone standalone kitchen joints ... this seems like a pretty much solved problem

Working in the industry I can assure you it's not. There is a huge market segment that will not use the existing tools and solutions as the learning curve is more than they want to put in.

These people will always use a kitchen designer and often get bespoke cabinets. The cost of doing business that way is orders of magnitude higher than IKEA or Home Depot. Those offerings also suffer from very low product quality.

Seems like it's not a tool problem, then - it's an attitude problem :)
Customers are lazy. You can make money by making their lives easier. Uber eats is a prime example. It's not really that hard to drive out and pick up your food but they made it so easy.
>It's not really that hard to drive out and pick up your food

But it is time-consuming - why spend 30 minutes going, grabbing, and coming when you can spend 30m at home waiting for it to arrive?

An anti-bloat tool for software development.

There are plenty of tools that lets you plan, organise and bikeshed over all the things you want to do, like Jira, but I think a much bigger efficiency booster would be one that helps you decide what not to bother doing at all.

along those lines ... prioritizing what you don't need to keep (in terms of libraries, unused code, unused features, etc)
One problem is that the most important metric of a feature, its ROI, is also the hardest one to measure. Measuring how hot the code paths are may not be much better than measuring programmer productivity in lines of code.

How can we quantify the investment without invasive surveillance and micromanagement of developers, and how can we quantify the return without reading the customer’s mind?

It shouldn't be that hard to excise unused code blocks, at the very minimum, from a finished product before it goes to prod :)
No, excising unused code blocks is a solved problem. The trick is to identify code that is used, but shouldn’t be.

I mean identifying features with a negative return on investment, features that you should avoid building. What makes the problem even harder is to consider not only the developer hours, but the opportunity cost of not building a different, more valuable feature.

Complex scheduling. I'm yet to see a product that would allow me to enter multiple constraints/availability or related resources and it would automatically calculate the best options for everyone involved.