Ask HN: How to find profitable side project idea?
Hi there HNers,
Last summer I worked on a startup as lead developer. I learnt about running a startup and talking to users while working there. I have now quit working there to focus on my studies (second year university in London).
I would like to work on a side project that eventually would lead to some revenues this summer. My question is, how did you find problem(s) to solve ?
I have read books and blogs suggesting that the best problem is the one that I have faced before. I find it difficult to do this when almost every problem that I found, there has always been existing solution or the solution can be solved with some quick searches.
Any idea or thought is really appreciated. Thank you.
213 comments
[ 3.2 ms ] story [ 229 ms ] threadThe lowest effort/highest reward ones I have seen are curated subscription email groups e.g. a vetted list of high quality freelancer jobs for freelancers.
The most common thing I've seen is that freelancers will (as the situation calls for it) strike a deal with their client to carve out a feature as an independent SAAS, bill the customer a much reduced rate for that "set" of work and bring them on as the first "customer" of the SAAS so there's some social proof.
When you're exposed to a bigger set of problems, you're more likely to see something that you can provide a solution to. And when you see similar challenges happening again and again, you start to get a sense of what problems need a software solution. If only one client has a specific problem, it might not be amenable to a SaaS solution. But if you see the same problem occurring over and over across multiple clients, you might just have a good SaaS idea on your hands.
https://i.imgur.com/waRlW5h.mp4
https://i.imgur.com/rms768q.mp4
If you haven't had such thoughts yet, I'm sure others around you could make suggestions. You'll make even more money by building for a niche market that has been overlooked or underserved by technology companies. The more specific the market, the better, as these tend to be quite sticky.
Not sure you can go that far. The first part of your statement sounds reasonable though.
You don’t have to find a novel problem. Find a product that already exists and out-do them. If you can criticize a product, you can make a better version of it and create value for your customers.
Finding a novel problem also works but less often than going after an existing known problem with some unoptimized solutions out there.
The secret is to go out and make friends who work in different industries that are totally unrelated to your experience. The world is full of profitable niche industries that need custom solutions that software developers have no idea exist. The business and industrial world largely still runs on excel spreadsheets, emails and word docs. Those are all areas ripe for new products and many of the top companies in those niche areas have no idea how to apply technology to solve their problems more efficiently.
So why hasn’t supply met demand? Why isn’t the market working more efficiently?
There can be various explanations for that, obviously, including the lack of information you hint at (“secret”, “have no idea”), but Occam's Razor would suggest we examine the assumptions first.
I’m not so sure that there is such a strong demand side in this market.
Even though many software companies and startups are trying to push SaaS software "rental" terms, many of the market participants are looking for a one-time purchase more similar to "ownership"
I grew up in the mid-south and know of literally one or two other people that code.
I know live in a major metro area and still only know maybe 3 people that code. It is actually a rare skill.
At the same time, I can't even count how many contractors, real estate agents, sales and marketing, and doctors and lawyers that I know.
For ex, in multi-stakeholder industries like healthcare where the people who pay for a product often have incentives that conflict with the incentives of the user of a product, following the North Star of "making something people want" can lead you astray. So many digital health companies make products no one will pay for, or that end up in a narrow consumer lifestyle niche instead of actually improving healthcare in a meaningful way. A team of engineers / data scientists who "follow the user" isn't enough. It isn't even enough to focus on physicians. You really need to talk to administrators and see how the sausage is made
> The business and industrial world largely still runs on excel spreadsheets, emails and word docs. Those are all areas ripe for new products and many of the top companies in those niche areas have no idea how to apply technology to solve their problems more efficiently.
Spreadsheets are incredibly useful tools! They're general-purpose, easy to use and powerful. Millions of people use them for programming without even knowing it.
Email is still great for communicating! It's everywhere, it's easy to use and it's totally interoperable. How many startups are building decentralized communication tools right now? Email has been federated for decades.
The notion that use of these tools is an opportunity for some more efficient proprietary solution is a bad assumption. People use them because they're good and they work. The parent is right: there's less demand than you think.
I strongly disagree with the implied defeatism here. There have been many attempts and many failures with many popular ideas for non-obvious reasons, and it is worth being #11 if you can A) identify why the previous attempts failed, and B) find a different angle that avoids the issues.
Google wasn't the first search engine. But they were effectively the last. First to market is only important insofar as it's correlated with making sticky products, building a moat, and not dying. That's what you actually want.
An aside, but funny enough, one thing I like about DDG is how much better it is at finding YouTube videos. I assume Google is encumbered by antitrust fears or something when it comes to video search.
The product has to be good enough to solve the problems of the market. #2 is also a downside of this - you can't dump marketing money to get around PMF.
Facebook is an interesting company which had a poor market (people are fine without it) but they managed to improve the product, the economics, the scale, just enough, and ended up one of the biggest companies in the world.
There's a large demand for, say, a more privacy based social networking website. Or a gaming based social networking site, like what FB-Zynga used to be.
Google has a large blind spot too. Yelp is filling one, Pinterest covers their image search weakness, tons of travel apps/sites are filling another. But there's many, many things that can be more localized or made into a niche.
I'm not sure if it is harder, as technology has significantly improved. I think we're at the stage where less of the effort is building/designing an app, and more of it is in better product development.
Of course if they have a specific insight into a business that has failed ten times that they think is a major differentiator, then of course they should investigate that first. But then they wouldn't be posting on a forum for business ideas :)
Great advice, my SO works outdoors doing field studies and the amount of shit they talk about doing manually is astounding. They pay people (or sometimes unpaid interns) to go through hours of videos documenting everything they see! Almost all cataloging and measurements are done with little to no technology. Another example; I volunteer for a program that does some human intake and processing, before I got there everyone used a series of paper forms that got updated by work of mouth. Now? Google Spreadsheets.
There's hundreds of industries, organizations, and groups that do extremely manual labor intensive work because it hasn't even occurred to them it can be automated or turned in to a service, but you'll never meet them if you hang around tech people all the time because the solution is always on their mind.
> I work at NASA creating procedures for spacewalks (EVAs).
> Currently we write these procedures in Microsoft Word.
https://news.ycombinator.com/item?id=5585535 (2013)
Hm, I wonder if they changed the system by now?
edit: Seems like there was still work to be done 3 years ago at least..
https://news.ycombinator.com/item?id=13080941 (2016)
[1] https://github.com/jamesmontalvo3/spacewalk
Or in other, more specific words (and I have no affiliation with the product): Showing them how to use something like Airtable and setting up a simple template for them could change their world.
There are niche apps that do it but they charge a high fee to the college. Case you are wondering:
https://observideo.com
You need the insight in to that industry — by definition, you don’t have it. So find that person, and don’t just take their idea and leave. Work with them. Have them be the champion in the industry for this thing you make. Who knows best where the thing can be sold? They do!
I'd slightly rephrase this: make friends who are business owners in different industries.
Having a good idea is only the first step. You need to be able to market it, and ultimately sell it.
Business owners face business problems, will have an idea on how much they're willing to spend on it, and will understand the budgets and where they can be applied. Rank and file employees often do not make purchasing decisions on behalf of their companies. Some may, but most do not.
My personal story:
I became friends with the owner of a nearby gym, and I was immediately impressed with how he ran his business, and more importantly his ability to stay ahead of the curve, learning and adopting new ways to market his business.
For a free membership, I would do one off small projects, that typically string together API's offered by the different services he used (Scheduling, waivers, forms). None was particularly interesting, until he came up with a hodgepodge system for contacting new members.
I helped him build some flexibility, and we both came to the conclusion that if we built a platform that specifically did this, we could market it and sell it.
We boot strapped it and made about $13k in revenue the first year.
Now that we have a viable business, we've shifted to growth: get customers.
I find it difficult to put in place a process to consistently meet business owners.
Some ideas: - going to meetups or better organizing one - going to conferences or better give a talk - be part of mastermind groups (but dont know where to find good one) - hang out on entrepreneur communities like indiehackers.
Other ideas?
> The business and industrial world largely still runs on excel spreadsheets, emails and word docs.
Why assume that changing this hasn't been "attempted 10 times"? Perhaps the reason for that is simply opaque to you and as a software developer you think in terms of software solutions. However, getting these solutions sold is the real problem.
I agree with this, but I'd add one more point. In addition to face-to-face time with these people (which can be hard to come by, either on your end or on theirs), you can try to learn about other industries by reading their books, whitepapers, journals, etc., watching their videos, documentaries, etc., and following industry trade magazines, analyst reports, and the like. This stuff you can at least do on your own time, late at night or on the weekends, whenever you have time to spare.
Given this kind of research, you can put together a tentative "value hypothesis" regarding a potential solution in some other space. Then you can maximize the value of your face-to-face time with other people by talking through a specific hypothesis, instead of just an open-ended fishing expedition.
Of course, if somebody is willing to give you a half day of their time, and you have that half day free, then go for it. But generally, people are more open to sharing their time if you can make it clear that you respect their time and that the interaction will be somewhat directed / focused.
1. Find a cool machine learning project with preferably pre-trained models so you don't have to do much cleaning/moving data.
2. Get those models doing inference on a server and expose it as an API.
3. ??? (marketing/sales/business stuff)
4. Profit
For example https://remove.bg got a lot of upvotes - it probably uses this on the backend:
https://github.com/tensorflow/models/tree/master/research/de...
I'm not the person to sell you it. That's just an example where someone might make money.
Building your own deep learning model is expensive and resource intensive, if it’s a solved problem it’s a great thing to outsource.
There are forms of ML that have memory/stack, and I would think you have to use something more complex for the stock market than "recognizing patterns". For short term trading there are definitely useful pattern recognition systems that can effectively trade.
Examples: - Newswhip crawls news articles and uses the FB api to get share counts. Newsrooms use their data to find out what is trending
- SuperData crawls info from Twitch and YouTube to provide insights into the games that are engaging.
- Similarweb provides data into the traffic that websites are getting
- AppAnnie scrapes App rankings to provide insights into the growth and trends of apps.
- Ahrefs built a huge database of backlinks and provides insights into who is linking to your site or your competitors.
Sometimes you just have to compete against existing ideas and try to do better than them.
If you already have money there are other strategies that you can use to generate extra revenue.
I agree with this advice, but my problem is that most problems I have faced are software development problems. I have tons of great ideas for software development tools that would be super cool. However, there is no market for selling software development tools, because everyone is used to getting them for free from open source projects subsidized by large companies.
Nobody wants to pay for software development tools. Even worse, software development tools tend to be local programs instead of cloud software, and licensing/DRM for local programs is a nightmare for everybody involved. Either you just accept that the majority of people will pirate your software, or you go crazy implementing invasive DRM schemes that hurt your paying customers and still don't work.
I think this is actually a big problem in our industry. Our tools suck because there is no strong profit motive to produce improved tools. We have to muddle through with whatever big companies see fit to subsidize in their open source programs.
IntelliJ and Docker give most of their products away for free. Why? Probably mostly because there is no good solution to licensing local software. They rely on the patronage of a relative few corporations, who have legal departments that will enforce the licensing rules on their behalf to avoid legal risk. Unfortunately that means that their business is an enterprise licensing business which is difficult to break into. Success in enterprise licensing is less about product quality and more about sales. That's not the kind of business I'd like to be in.
I know I have an active subscription still for PyCharm, even though I've not used it personally since moving to vim full-time a couple of years ago. It's a great product, and I find it extremely useful when teaching people new to Python - the autocomplete is excellent, and there are visualization tools that I've not found a good replacement for.
First, they charge different prices for an "individual" licenses vs. an "organizational" licenses. Why? Because businesses have larger checkbooks.
PyCharm: there's a free Community version and a paid Professional version. Python is the most common beginner teaching language, so give the basic stuff away for free. Once you get into "pro" stuff -- sci stuff, web dev, DB stuff -- that's when there's a budget and so people are willing to pay.
WebStorm: Javascript is a mess of an ecosystem and there's tons of free tools out there so no point in JetBrains giving away a free version. However, once you value your productivity and want something that Just Works out of the box with reasonable defaults, that's when a webstorm license becomes something you pay for. Besides, every startup does some form of web dev these days, so larger market.
IntelliJ: Java used to be the go-to language for learning object-oriented coding, so like python, offer a free community edition. Once you dive into Spring stuff, well, you're going there because you have a boss and the boss has budget.
For an extreme case, it worked for Linus
I think your point still stands, but instead of Linus I'd use the more relevant Mike Perham and his Sidekiq tool: https://news.ycombinator.com/item?id=12925449
Oh, quite the opposite I'd say. Because we are software developers, and we can fix our problems with software, we accumulated quite a rich toolbox to solve our problems better. We're quick to dismiss our ecosystem (everything sucks/is broken), I tend to enjoy such conversations as well, and sure, there are plenty of broken/unfinished things out there, but in all honesty: what other subculture has accomplished such complex projects as the Linux kernel or the GNU toolchain, and gave it away for free? Because a lot of people know that we all are better off sharing our tools, and copying is essentially free? Isn't the brokenness often there just because we went too far, too fast?
I think looking somewhere else for a side project is a good idea, exactly because our own problem space is saturated with good (enough) tools, and deeply explored. And even if something is missing, someone's itch is bad enough so they just scratch it themselves and make it open source. But oh boy, other domain's tools and workflows suck. And often, they don't have the skills to fix this themselves, so that's where we can be helpful with software.
Wouldn't it actually be a good example of "whatever big companies see fit to subsidize in their open source programs"?
Is it? From my experienced with tools from other industries where there is no subsidization of FOSS alternatives, I wouldn't say they suck any less, would you?
Ok not literally. But the analogy is suitable. I'm suggesting you try learning about an established business model, on a small scale, so that you learn some things about business in general. How to do accounts, pay your VAT and tax, buy supplies, advertise, ship, and so on.
So something like a webshop that sells your favourite widget might be suitable since you're a software guy. You can easily look at how the pages are set up, how payments are hooked in, how to do a landing page, and so on. The advertising side is a deep dive as well, but well supported by your background in coding.
Follow up reading: Peter Thiel’s “Zero to One”
https://www.amazon.com/Zero-One-Notes-Startups-Future/dp/080...
I did a consulting gig for an art gallery a couple years ago and I came up with a dozen ideas by watching how they operated. I'm sure the same applies for a gazillion of markets out there, ones where technology isn't that dominant.
Side note. You may want to take a look at the famous "passive income" threads in here. There are hundreds of ideas and projects in them that could inspire you to build something new. Just make a search for "passive" on the search box at the bottom and good reading.
It's not what I'd (personally) start a startup on, but the idea is there. I was also super green and coding in VB so I'm glad I let that miss. If you can get in with a small business owner and see where they're spending repeated human cycles for data entry, you can potentially have an idea.
My project that comes to mind is when I helped a NAPA parts store retrofit a new cash register into their inventory system. Their inventory system was running on AS/400, and their old register that failed was a beast that let them enter a part number for each item, then communicated with the AS/400 over a proprietary protocol to get the price. When the transaction was completed it decremented the stock. It took me about sixty hours to reverse engineer the protocol from the failed register, then another forty or so to write a Python web service that communicated with it and exposed an API and ran their dot-matrix receipt printer. I farmed out creating a web UI to someone on Upwork and left them with a documented solution that let them use off-the-shelf hardware to talk to a small server that then spoke AS/400. I've since replaced the server with a Raspberry Pi, but otherwise it's been in place since 2003. The front-end was originally a touch-screen Windows CE client I snagged at a freight auction, and the last time I stopped by (a couple of years ago) they'd replaced that with an Android tablet on a pedestal and another tablet that they could carry around.
That sort of thing obviously has value to a lot of businesses, but each one is going to have a unique setup. There probably aren't a lot of parts houses out there that bought that particular vendor's inventory system in the 80s/90s, haven't upgraded, want to keep using the same backend, and are willing to upgrade their registers to something modern. I'd go so far as to say it's probably a unique situation.
Alternatively, think about trade-offs that you or people in general feel forced to make, and try to build something that offers a compromise, a third alternative that previously didn't exist. That's innovation.
at its core it is a matter of can you save people money? time? or can you make them more money? those are the reasons people pay for things. Also competing on price alone is a fools errand.
feel free to email me for more help, i literally do this for a living (helping students generate and grow their businesses at a university).
Sometimes it's just the simplest, least technical and gimmicky things that make money.
- was the domain related to your expertise, or in a niche that was new to you?
- what made this project stand out from others / appear to have potential?
- was codebase quality a factor in purchase?
- do you inherit things other than software from the seller? EG customers, marketing channels, analytics, etc?
- what was the value-add that made it work under your stewardship but not the seller's Sales / marketing / dev / featureset / etc
- Are you interested in doing this repeatedly?
It is not?!
They did post about it three years ago, so it seems linked to them. But I'd wait for OP to confirm: the person posting is not OP and may just be guessing based on OP's post history and having searched 1k projects.
edit: and since it's from another HN'r, good work! It's a good idea with a really tasteful/simple execution.
- Yes, the domain area was something that I was interested in already, but not necessarily practised. - Quite honest, I chose it purely on price and the design of the items bought. The dev put in effort to get everything designed nicely. - No, not at the time. It was a rails project so quality was not considered since its too easy to add on/rewrite where needed - Got customer lists, source code + design source, social media pages, analytics, payment profiles, domain. - I had a network locally that I knew I could sell to. I sold hard and 3 big agencies are now using it at a decent monthly fee. - Definitely! I'm scouring 1k daily now and have found a few other gems.
I suspect this is the key to success in most smaller startups. Instead of focusing on getting 0.01% of a global market go for 10% of very local/focused market.
A prepopulated sales channel goes a long way.
Wishing you the best – really great to have an income stream like that, buying you room to make even more successful bets.
I'm a bit surprised you haven't heard about this one: https://flippa.com
1. List 3-5 things that you enjoy doing routinely 2. Now imagine how you can improve some of its aspects 3. What would that ideal world look like? 4. Everything can be improved!
Mine - Meditation - When I was starting guided meditation options were limited, still are to an extent. Some are expensive and some are limited with options. It does not have to be a solo journey like mine and many others' were. If I did not have anything to work on I'd create a database of guided meditation tools resources, there are 1000's on youtube. I'd categorise them by mood, skills-level etc... Have a basic social aspect to it by adding a community that share the passion and I then can introduce paid features such as Journaling etc. By leveraging the free database of contents out there one would have a decent competitive advantage. Just a thought...
I hit on a solution to a problem while engaged on web forums related to my hobby. Now I sell a little gadget on the side. The people on the forum, with whom I was already on good terms, provided my initial sales and word-of-mouth.
Here is archive of sorts - http://www.oppslist.com
The ideas are of a mediocre quality, but I find they help stimulate thought.
Another idea is to go to a place like Fiverr and look at which services are most widely bought. Then build a saas solving one of those problems.
And the opportunities listed all seem at least a year old. That's not very "new", is it?
How can someone follow this list? I'd be interested in following and possibly knowing which are new when I come back.
I personally tried out an outsourced sales venture based on one request. I made around 1k in revenue the first month doing that. I found that my overhead was high and it was a competitive landscape to find clients.
At this point in time, I'm just playing the messenger. I'm not trying to get people rich. Although, I think in the future that may be a metric I start to measure myself by. Hope this helps :)
I'm avoiding mentioning a product or category not just to avoid throwing someone under the bus, but because I can think of many examples of this. None of the software (or any other kind of product) that I use today was the first of their kind. They learned the lessons from the earlier versions, and did it better.
Even if every problem you can think of has been solved (and I'm skeptical of that), I'm sure you don't think that all existing software is already perfect. Pick something that's terrible, and do it better.
Ideally, you would then become more responsive to problems and have a wider solution set that perhaps doesn't just involve software.
Don't listen to any other advice :)
Most of the time failure nets a great learning experience regardless of other outcomes.