Tips for Developers Who Want to Build a SaaS Startup
Having built Proxies API and made it to 100 paid users for the first time in my startup life, I tried to compile a list of tips for other developers like me who might be thinking of starting something on their own.
Developers are a different beast, and the advice here may not apply to anyone who is not a developer and is a normal human being.
Pick an idea you have used. When you start a startup one of the persistent problems you face is that you don't relate to the problem. If you are not a parent, its almost impossible to relate to any conversations that parents have about their troubles. You just don’t. This lack of empathy in a startup setting adds up to a huge disadvantage and will never allow you to be fully confident in what you are doing. 2. Pick an idea you have already paid for: If you have ever paid for a service, think very carefully about that. Can you code it yourself if you give yourself a few months? It is because developers don’t pay for shit. If an app forced you to do that, then there must a real need for that app.
3. Don’t pick something that needs a pretty UI: My product Proxies API is an API. I can get away with almost no UI. It is such a relief that I don’t have to work with a designer. I find that I am 4 to 5 times faster when I don’t have to deal with UI stuff.
4. Don’t pick anything that you need in-person sales for. Marketing is a developer’s friend. I realized this only later in my life. In-person sales are very weird for the developer personality. My advice. Don't do it. Learn how to market instead.
5. Pick something that is a self-serve model: People signup for a trial and decide to pay or not pay based on their trial experience. This is a beast that you can conquer. You don’t want to be going around talking to actual humans. It is not for us.
6. Try writing: If you can code and write, you will win the world. The whole success of Proxies API is based on constant improvements to code and a little bit of consistent writing over time.
7. Keep it real: Be clear on your motivations. I know that developers are not motivated by the prospect of making millions. But they are excited by the promise of financial freedom. So I was clear on my motivations when starting Proxies API — I didn’t want it to make me millions. I wanted it to give me financial freedom by earning what I earned in my day job. I didn’t need a penny more. By being clear about it, you are not living someone else’s life but your own, and it will give you the necessary fuel to keep going.
Have fun.
The author is the founder of Proxies API the rotating proxies service.
This article originally appeared here: https://www.proxiesapi.com/blog/Tips-For-Developers-Who-Want-To-Build-A-SAAS-Startup.php This Proxies API service seems to be using AI-generated "This Person Does Not Exist"[1] style profile pictures for its positive customer reviews near the bottom of the page: https://www.proxiesapi.com/assets/img/steve.jpg https://www.proxiesapi.com/assets/img/customer1.jpg https://www.proxiesapi.com/assets/img/customer3.jpg https://www.proxiesapi.com/assets/img/customer4.jpg Makes me question if the customer reviews or even the customers themselves are real. I can't find any evidence any of those people, or the companies they purport to represent, actually exist. One thing to slap faces where you don't have any, but straight up making up reviews, people AND the company is actually hilarious Almost every restaurant website uses stock photos as well - "oh this place looks nice" and tineye reveals that indeed the stock photographer made it look nice as well. Logos, customer reviews, etc. are just tricks. I read the "How to recognize an image of a fake person" section on the site you linked and I'm curious how did you spot that the pictures are fake? If you see enough of these AI-generated portraits on the Internet you'll start to notice the similarities, especially since This Person Does Not Exist is usually the source for the vast majority of them that I've seen. They're very often used as profile pics by Twitter bots. The way the person's head is framed in the shot is pretty much identical most of the time and kind of a tell on its own. Other tells: - steve.jpg: The skin above the top rim of his glasses has some repetition/distortion. - customer1.jpg: Distortion in the hair, most noticeable at top left. Mouth looks fake too. - customer3.jpg: I didn't notice anything significant, but it just looks like an image taken straight off TPDNE. Do "vibes" count? - customer4.jpg: The teeth merge into each other. Another tell is that they often look cropped from a larger image, specifically not from the center. Lenses produce specific distortions towards the edges. It looks like to increase the corpus of images many headshots have been cropped from larger images. This causes the model to learn and replicate these subtle lens effects. > - customer3.jpg: I didn't notice anything significant, but it just looks like an image taken straight off TPDNE. Do "vibes" count? Look at the background, does it make sense? Tell me what the background to the left of hear head is.
Look at her earbobs, how often do people wear different ones in each ear? Does the left earbob look like anything? Hmm, thanks. To me those look like random compression artefacts or poor photo retouching but I get what you mean and they do all look similar. I guess you can get used to the distortions and the general vibe of the model. For me a big giveaway on customer3.jpg is earrings - they don't have a real shape and don't match at all. Good catch. Go to https://thispersondoesnotexist.com/ and spend a few minutes refreshing the page. There's something distinctive about all these fake faces - I can't explain what it is, but once you've seen enough of them you can recognise the "look" quite easily. Probably the biggest giveaway is that the faces are all positioned identically in each photo, but there's something else that juuuust quite doesn't look real. Maybe some of the faces just seem a little bit more androgynous than anyone I ever see in real life. There's something about them that sits just on the edge of the uncanny valley. Outside from the odd file names, minor details like ears, glasses, earrings. Our friend Steve has something going on with the left hinge of his glasses. Customer1 has an interesting right ear. Customer3 has very asymmetrical earrings. Customer4 has a bulge where the left earring would normally be, and the right hinge of their glasses is crooked. I think the image names gives it away more. "customer1.jpg" and "steve.jpg" lol Usually eyes are not completely symmetrical. Yeah the reviews on the landing page seem fake. It's the kind of fake it till you make it that doesn't bother me. It's not like he's faking blood tests. Problem is that it makes you question the whole post. Ie does he actually have 100 customers or is it just something he writes because it is well known that people are less likely to buy something no one else is buying. Yeah, you never worry about doing business with an out-and-out liar, of course? It’s one thing if Reddit founders post news articles themselves on their free site. It’s another if a saas site has fake reviews. I think this is actually quite common now - where you get permission from a person to use their review of your product but they don't want you to use their likeness. Maybe sometimes it feels a bit much to ask to use someone's review and also ask for a photo of them. It is somewhat disingenuous but I don't think I mind too much if they are legit customer reviews. Try googling the names, job titles and companies of the people who "wrote" these reviews. I can't see any evidence that these people (and in some cases their companies) actually exist. I don't believe that these testimonials are real. When compared to a similar site hosted on the same server[1] the reused pictures and reviews are definitely suspicious[2]. [2] see the Our customers feel it section Actually teracrawler.io and proxiesapi.com are build by the same person, same page design, same footer design (left column of topics, no privacy and terms policy pages), even is the same chat software with the same customer support operator picture. Sure, that's how I found it, there's links to each other in the menu. I didn't mean to suggest the services themselves are fake, I didn't word it well - just that the pictures and reviews are reused on similar but different services. The pricing is very different, could be some kind of test to see how much people are willing to pay I dunno. I call bullshit on almost all of these tips. Not all developers are awkward, socially-inept, and incapable of designing their way out of a paper bag. I find that to be a really tired stereotype these days. I’d say: if you want to be successful, push yourself. All of these things are skills that can be learned. Any tips on how to push oneself? I frequently find that I reach the “I give up” ceiling very quickly when I try to push myself on most things. I reckon it’s because I can be quite impatient and intolerant towards frustrations. And what are your thoughts on talent? Sure, most skills can be learned, but the rate of learning can vary and time is limited. How worthwhile is it to learn something if you know you might be slow at it and the skill plateau is probably not too far off? Do you normally aim for good enough for those skills you did learn but don’t think you can be an expert on them? I'm impatient and easily frustrated as well. But I've learned to kind of lean into my frustration as a motivator when learning new things. I try to spend time learning things that interest me personally, and when that happens to overlap with things I deal with at my day job, great. If not, oh well. I don't really consider how long it might take to learn something if it's interesting to me, and I tend to dive pretty deep, even if I don't get to "expert" level. I figure if I need to become an expert in a thing, that'll come with time. I see. I like the idea of leaning into my frustrations; maybe I should use it as a signal to persist rather than give up. Strongly disagree about talking to people. You've got to talk to your customers and non-customers. It's a skill, you can learn it. I'm a ~30K MRR indie hacker, and 10y backend developer. I 100% confirm everything he said, but I too agree the most important features all came from talking with users/prospects/customers. I'm shit at it, but the mantra I tell to myself is: "remember when you were a kid? These are other kids that play with the same toys as you. Just talk about the toys and how to make toys more fun. Same enthusiasm!". Just stay away from the mindset of "convincing people". Nobody likes being convinced to do something. > I'm shit at it, but the mantra I tell to myself is: "remember when you were a kid? These are other kids that play with the same toys as you. Just talk about the toys and how to make toys more fun. Same enthusiasm!". The problem is: this kind of "nerd talk" that you are enthused about often comes across as awkward in particular to business-minded people. In other words: I can perfectly understand why many developers have bad experiences with talking to people: too different interests and thus quite a lot of bad experiences in the past. You can make more money in a five minute face to face conversation than you can in 6 months of internet marketing in some industries. edit: example: one such industry is selling big expensive machines and/or software to run them. But big expensive machines and software sales usually involves a lot of customer research and pre sales intelligence and rocking your sweet lone self (allowing you might be able to fake the roles and personalities necessary to mimic a remote team) to a enterprise closing meeting, isn't going forwards. Possibly you could look for interim executives from among the growing pool of experienced management exiled by industry ageism. But this(and my other suggestions) do require a broader professional ability. Enterprise software sales seems to have suffered something akin to job title inflation in IBs and start ups. And I can buy a lot of enterprise software directly online even in single digit seats. If you can do any of what I just pondered, almost certainly you're better off using your energies to hone your customer product fit, Steve Blank style or any which way. The only thing I'll differentiate in my thinking from the OP advice is consider thoughtfully how well you are served by diarizing your development process and how much more can come from thinking in reverse from the customer perspective. Absolutely, this is probably the most important thing a founder can do Sale skill is different animal though. I'm confident I will never be able to sell insurance or car. Not sure why this is not being flagged. It seems like spam and the footer "The author is the founder of ... the rotating proxies service. This article originally appeared here..." just feels like someone copy-pasted their marketing blog post to HN. There isn't a great deal of content here... this is purely marketing spam? Yeah, it’s very low on actual insights, and appears to be an excuse to just promote their own product, as they mention it a bazillion times. The post would have had more merit if they left their startup’s name out of it, e.g. only on their profile page. Yep. The first tip is to do some spamming to get customers. I would like to hear more about how you approach the issue of taxes. The way I understand Stripe Tax is that it will give you a huge list of sales like: Is that true? Do you pass that giant list to your tax accountant and they fill all the correct paperwork? What if some of those sales mean you have to pay taxes directly to the country in which the buyer is? You deal with that countries government directly then? To be honest I think most small saas are not compliant. When you are doing low revenue numbers like this guy under 100-200K per year its hardly ever relevant. A lot of countries and states don't claim sales tax to international countries under a certain treshold: 30 - 100K per year. (if you are selling to consumers would actually be a waste to charge sales tax) The EU is different, but if you are in the EU relatively easy to comply with. it's not easy, it's a nightmare for both you and your accountant (which is happy to charge you for that) Use paddle or revin instead of stripe or use stripe with quaderno or with stripe tax. Stripe tax doesn't work too well yet imho but it may end up getting there In EU it's easy if you have their VAT numbers, otherwise you just have to charge it. If you have EU VAT nr it's actually work for your EU customers, they have to pay for it without it being on the invoice. You can use a MoR payment provider such as paddle or spring which handle tax compliance for you globally. They will pay tax and take liability on behalf of you. They charge 5% per transaction. Do you use Paddle or Spring? I would expect that MoR's have their own set of issues. As they become the seller. So they are liable for the product quality, safety etc. So I don't think signing up and working with them will be very easy. Especially if your product is not as standardized as an ebook or something. So if you use those, I would love to hear about your experiences. What you sell, how the signup was and how the relationship with them is going. Last time I looked at the Paddle signup flow, I already had two issues I could not resolve: - What product category my product falls under. None of their categories seemed to fit. - How to sign up before the site is live. Paddle wants to look at the site. So do you have to build a dummy that works up until the point where the user buys something and then it says "Sorry, we are still looking for a payment solution"? Paddle only replied with nonsensical semi-automated messages to my questions. And when I answered and asked for a real answer by a human, I never got a reply. I have recently used Paddle for a small saas. I am selling a Pro plan that give access to a restricted feature of the web app. As the Pro plan was introduced as a later stage, Paddle verification team accepted to review a "staging" version of the web app before I post it to production. I definitely talk to real humans, but I took time to express my situation. Paddle integration not too painful (good but not perfect documentation) considered it was my first saas. I might write about it if some of you are interested. How did you give them access to the staging version? Via a secret url? Or with login and password? A secret url who wasn't to secret, it was simply a subdomain : preview.mydomain.com. Even if it was a subdomain, Paddle team was accommodating and approved the main domain. Just talk to them. https://www.paddle.com/help/start/account-verification/what-... That would be extremely useful to know about. I never understand how SAAS products deal with tax. FastSpring is another alternative for SaaS & international taxes. I'm sure there's more besides Paddle & FastSpring. The idea is to look for a company that will act as the "merchant of record", not just a payment processor: https://fastspring.com/tax-management/ You end up paying more to them than Stripe would charge, but it might be worth it if you just don't want to deal with that administrative headache. (I'm not running a SaaS, but shareware / downloadable software companies have been dealing with international taxes for 15 years or more now, since the digital EU tax laws came in. The idea of global consumption taxes wasn't really a thing until then, so you only had to follow your own local tax rules in the early-mid 2000s.) Thank you. That's really helpful. It's nice to see governments working together to create whole new industries dedicated to solving administrative nightmares :) I use Paddle on https://webtoapp.design and went through their and my TOS with a german lawyer. Basically they're only responsible for payments & customer support (however most customers come to me directly or get directed to me, as Paddle can only really resolve billing issues). Paddle checked my site & product 3 times to make sure it complies with their requirements (an important aspect in my case was that my app creation process is automatic and not done by hand). I always got in touch with a human quickly, both during the account setup and later when I needed help (through chat or sellers@paddle.com). In fact I'd say it's probably the best support experience I've ever had. Regarding your sign-up questions:
- Back when I signed up there were no such options but now I'm in "Standard Digital Goods" which seems to be the catch-all for everything that doesnt fit into one of the more specific categories.
- I had integrated Stripe and Paypal (never again) already, so I didn't have that problem. You could use a "Contact Us" placeholder to manually process payments I guess. If you immediately apply to Paddle then, you shouldn't have to keep that up for long anyways. And it allows more payment options like Paypal. I've tried to make stripe recurring payments once many years ago, but during that time the backend payment api did just not give out any info about the possible tax region. In eu it's a nightmare because of reverse vat... If you're posting an article, you should post the link instead of the full text, FYI. If you want to do a Tell HN that works too, but then there would be no need to include the link at the bottom. I actually appreciated to read this article in full on HN. People should do that more often… I disagree - HN already gets enough self-promotion, the last thing we need is everybody copy/pasting their blog-based marketing here. This specific article was decent, but it absolutely is marketing and I don't want to encourage that. Yes, no pictures, no ads, I liked this format as well. When there's no domain listed on the front page listing, it is harder to tell what is worth checking out. FWIW, for this sort of thing I prefer the fulltext on HN, rather than clicking through to some ad-filled, slow and bloated destination that fingerprints my browser and tracks me forever. Congratulations on getting your first 100 paying users. Most startups never make a single dollar. How long did it take you to hit that number? Many of us devs want to do SAAS but don't understand what's a reasonable amount of time to expect until we can get "ramen profitable". It took me around 1.5 years (at first part-time and then full-time) to get to 1.5k-2k€/month with https://webtoapp.design Probably quite long but I had a lot of stuff to set up to get the product automated (and I did barely any marketing, basically just wrote a couple random blog posts for SEO). Why does your site say: > It's not allowed for us to publish your app in our own developer accounts I've published apps for others in my developer account I can't find it in the guidelines right now, but it's disallowed by both Apple and Google. I've personally seen Apple reject apps because of that before (not an app that I created though, since I follow the guidelines) I found it, it's a restriction on "app builders" or "app generators". If I manually build an app for a client, it's fine for me to publish it in my account. Glad I didn't do anything wrong. Ah you're right it's section 4.2.6. Interesting point that it's not a restriction for all app creators but just more commercialized services. Nice work! Judging be your price that is at lease $2300/m which is impressive. It is also a great idea and good site. I like your guidelines alot. Definitely I tried cold calling once and it was a horrid experience. Mostly because any who wanted to talk wanted freebies only. :-(. Sales is it's own profession, it is hard to do as a hobby, and especially if not that way inclined. I suggest adding your site link to your HN bio, as it might be of interest to some HN people. probably an overestimate: first users are usually free or cheap, the price is raised later. but still impressive. Stickiness is more valuable early on than that $10/mo or whatever you charge. Make sure you measure how often a user returns, uses, consumes, etc to see if what you are building (or built) is actually used like you want. Often people forget this and it leads to hostile user behavior and payment traps. You - being the future SaaS startup developer. > Don’t pick something that needs a pretty UI For frontend engineers, should the advice be the opposite: "Pick something that needs good UI/UX"? I think generally you should pick something where you have that unfair advantage and you like to do. If you are good at sales of course you should pick something where you need lots of sales, same to UI if you are god like designer of course you should pick something where nice UI is advantage. I agree with this. The harder the thing is you're doing, the bigger your moat. Ideally it shouldn't be too hard for you though of course, e.g. because you have prior knowledge about it. If you're a designer, pick something that needs to look good If you're a UX person and have access to a pool of people to do user testing with: pick something that needs good UX Those two are somewhat at odds with each other, and neither is really the job description of a typical frontend engineer (though of course some frontend engineers are good at one or both) I think the author is making some huge generalisations here. "Don’t pick something that needs a pretty UI": What if I'm really skilled in UI dev and have a good visual eye? Or a friend happy to do some free mockups? "In-person sales are very weird for the developer personality": there is no one developer personality, I wouldn't rule this out at all. "You don’t want to be going around talking to actual humans.": this is almost never going to work for a startup. Actually talking to people (users, clients, integration partners) is probably the most important thing you can do as a founder. Very solid. I've experienced all of these as well. > 1.. you don't relate to the problem Yeah, it's like being blind. "Why do people buy that thing?" is a question I barely get answer confidently > 2.. because developers don’t pay for shit It's getting worse. These folks always demand open source alternatives, self-host. Sometimes, the moral of story is "I want your free labor no matter how your family is doing" > 3.. Don’t pick something that needs a pretty UI: Super related here. When I realized my product value is literally UX/UI, it's like the end of the word because there's no many HCI solutions at all. It's always hard to use for any target, techie, non-techie, it's hard still. > 4.. Learn how to market instead. Developers try to do marketing is like trying to do self-surgery, ones just can't > 5.. You don’t want to be going around talking to actual humans I ended up shitting on a customer because he thought he knew problems (my app trying to solve) more than I do, but I'm confident he didn't know shit. > 6.. Try writing Yeah, this is part of content marketing. It kinda sucks to me though, lots of content marking is just for sake of marketing, doesn't bring much knowledge. > 7.. Keep it real. The most real sad thing is going back to working for corp. I would also add that you shouldn't worry too much about whether or not people want what you're building. Just solve problems. Also, be prepared to fail a lot before you succeed...your first solution might not go anywhere...your tenth one might not either...but your eleventh product...who knows? >Try writing: If you can code and write, you will win the world. The whole success of Proxies API is based on constant improvements to code and a little bit of consistent writing over time. Not sure what you're referring to by "writing" here. Blogging? Writing marketing copy? I find it interesting that you have a completely different URL/Brand for a same/similar product with different positioning. https://teracrawler.io/ How has that experiment been working out? Outsource the most as you can. Focus only on the logic and all will be fine. But don't outsource talking to users! It's the literal compass for a successful roadmap. This is why I think it is critical for a founder to do customer support for as long as possible. It should also be the last thing that is outsourced (or better: never outsource customer service!). Customer feedback is the most valuable resource for a starting company. Understand what brings value to your customer, and understand what the pain points are. Where do they get stuck trying to use your product or service? Apply what you learn from customer feedback. This is also a feedback loop on the improvement cycle of your product, as you should observe a drop in customer service requests once an improvement has been applied. With every improvement cycle, the product should bring more value and less friction to the customer. This will make marketing so much easier as you now understand the customer's pain points, and you have a product that solves just that. Do it right, and the product will basically sell itself (you still need marketing though). Good tip, any recommendations on outsourcing? Would you pay folks on fiverr or upwork to do stuff for your startup? Working on a SaaS startup my self, realized marketing is very important, i recommend you read the book “the mom test” it will give you tools to speak with users and get to product market fit This is awesome! Instead of pitching your ideas to VC-s who will then demand unrealistic growth, developers are building their own small internet startups. Yes, in most cases, they're not making millions, but who needs millions anyway? If you can buy food, pay rent, and support your family with a product that you made with your own hands, wouldn't that be nice? This is what my site https://microfounder.com is exactly about:
"It's possible for a solo developer to build a profitable microstartup to pay the bills and live life on their own terms." Haha I like your little pitch there ;) AND you got me interested! Definitely noted, thanks for sharing! Gratz man. I think most important hint is to just start building. I disagree, as an engineer building is too tempting and will get you in trouble. Instead just start selling. Keep the building to an absolute minimum until you've sold something. yes, but you need to or you SHOULD have MVP to sell just to not get into position where you sold something that isnt there :) Nope I still disagree. Having sold something that doesn't exist is a good position to be in. Then you just make it real fast (we're good at making stuff!) In the worst case you can't make it fast enough and you lose the customer, but you've at least proven there's a market for the idea. There will be other customers. that sounds good in vacuum but in real life if you fail one customer it can have ripple effect for future. This is Why i would suggest always to build MVP and start selling from that point. IT is easier to sell something new to existing client rather than sell blind and fail. as a dev and made few software/services as product, but don't know how to market them with good results. How do you market your software-product? with small budget..? bro you don't even have a valid cert on your website Cert loads just fine on my end. bro you don't even have an up to date browser / trusted cert store.
And then you are supposed to do the right thing with that list. $210 paid from a company in the USA
$123 paid from an individual in France
$300 paid from a company in Japan
...
