27
42 Comments

SaaS founders 👋 - what is your biggest technical painpoint?

I'm currently trying to understand what the biggest technical painpoints SaaS founders have? Some examples:

  1. I am not familiar with all of the technology stack that I need to use
  2. Deployment infrastructure is difficult
  3. I am alone and it is difficult to find the right people to have technical discussions with
  4. My painpoints are exclusively non-technical

Or anything else that you might find hard.

  1. 16

    Offering Custom Domain was a struggle. I made me switch couple of infrastructure. Moving to Render.com tomorrow.

    1. 1

      I can imagine. What infrastructure did you end up using to accomplish this?

      1. 2

        Render.com now. They have wildcard domain support natively.

        1. 1

          I was thinking more on the technical side rather than provider, i.e. docker images, static sites, Kubernetes, docker-compose, standard VMs? Something else?

          1. 2

            Got you. I went the server-less way so far. Was using Firebase, Vercel and now Render.

            Technical pain points I have come across are recently are setting up reverse proxy, server side rendering.

    2. 2

      This comment was deleted 3 years ago.

  2. 8

    If you are trying to bootstrap your company, you want to reduce the infrastructure cost as much as possible. Choosing a cloud infrastructure provider is such a big decision. Once you get locked in and have customers, then migration to another cloud infrastructure vendor is a huge pain. At the same time, you don't know which cloud infrastructure gives you the cheapest option.

    For my previous company, I used GCP because that was the platform I was most familiar with, but not sure it was the best choice.

    1. 3

      I agree. The hosting costs aren't a big issue when the website isn't doing much but one of my biggest concerns is getting a huge bill due to a sudden burst of traffic after posting it to HN or product hunt.

      Honestly, I don't know if this is real concern because I've never had it happen to me but I've heard stories and I've had large bursts of traffic to blog posts from HN.

      1. 4

        I just wanted to add two small notes here that might be helpful -

        I can now say that the traffic from being #1 on Product Hunt and Hacker News are nowhere comparable.

        PH maxed out at like 16-20 simultaneous users. Hacker News was over 600 at a time, and crashed my static website running on Nginx (without a CDN). I had to set up an "emergency CDN" with cloudflare.

        That said, for the current SaaS, we're using GCP and have invested a lot to make our serverless setup rather linear in terms of our costs vs. user traffic.

        We run an analytics tool that people install on their websites, and had someone install it on a website that had over 8 million monthly visits. The spike was so huge, we thought they installed it several times. We also theoretically have to handle scenarios where our users receive viral traffic.

        Thankfully, everything scaled perfectly and no other users were affected, and it didn't explode our bill.

        I think the biggest thing is just regularly analyzing your cost reports and finding stuff that's really expensive and optimizing it. Could be through caching, compression, performance optimization, etc. GCP has some pretty nice tools for this, I'm sure AWS does too.

      2. 2

        Tremendous price hike you are talking about, is of concern only when one chooses BaaS(like Firebase or AWS). Instead one can also choose VPS(PaaS) from a server host like DOMAINRACER(www.domainracer.com) or HOSTINGER(www.hostinger.in).
        In that case the plans are fixed very similar to your broadband bills. You can upgrade them anytime though!!!
        The advantages BaaS platforms offer are:
        # Test your MVP free of cost.
        # No backend code to be written by you. All prerequisites are available readymade.

        Since most of us are very likely to have an outburst of traffic ;) , their(BaaS's) freemium model is deceiving developers like us, to be ended by a hefty bill!!!
        Using PaaS(Platform as a Service) is a good option if you are concerned with your money(initially). But an important catch is you will have to code the backend yourself...
        Kindly correct me if their is a better option available...

        1. 1

          I agree: start with VPS, when you generate revenue or getting performance issues introduce HA functionality to SaaS and move (partly) to the cloud

      3. 1

        This issue can be faced by considering failover strategies according to your software. Most commenly you're getting to the point, that relying on the hosting provider doesn't give you any advantage according to high availibilty of your SaaS platfrom. You will recognize quite quickly, that the best way to introduce HA for your SaaS platform to introduce a non vendor lock in on the infrastructure side.

        E.g.: Deplyoing automatically a cointainer enables you to do that on any infrastructure given by any hosting provider

    2. 2

      Yes, I agree. This is definitely a huge pain. With my latest project, I've decided to deploy using Kubernetes in the hope that it will be easier to migrate, but the learning curve has been steep.

  3. 5

    I’m currently finishing my new web app Keepsy.app and my biggest technical pain point is setting up the payment system. I want to make sure the basics are there but also other areas like refunds, expired cards, etc. For some reason this topic gives me a lot of respect and I don’t want any bugs or unhappy customers.

    Regards

    1. 2

      Yes, I know exactly how that feels. I've just implemented the payment system for https://indiemy.com a few weeks back and it was exhausting to 1) to understand the documentation (as it is quite large for Stripe) and 2) make myself feel confident that I have tested enough and covered the corner cases.

      What tech stack are you using for your app?

    2. 1

      Check https://memberstack.com They'll help you streamline all these. Since they lean on Stripe, so you want to make sure Stripe works in your country.

    3. 1

      Same for me. I know stripe makes it a lot easier and with the stripe customer portal it gets even easier, but if you need to customize the flow, think about every step of the subscribe/upgrade/downgrade/cancel process makes it a hard part for me too.

    4. 1

      Did you consider something like chargebee.com to do this? I'm using it to process subscription payments for my web agency and it does a great job with refunds, dunning, ect.

    5. 1

      Your developer is struggling with that? Or are you a non developer and you're setting that up? Nice idea btw

      1. 1

        I’m a non developer, but finally I have it setup. I thought it was easier than I expected. Im using stripe.js with some API connections and it’s working nicely. It has only one subscription. Users can update the payment details or cancel their subscription. I’ve even setup users to get notifications when card is close to expiry date. I’m pretty happy with the result not being a developer. I’ve also had to setup the company and company Bank account. Hoping to go live anytime soon.

    6. 1

      I found using stripe makes it as easy as embedding an iframe.

      1. 1

        Stripe is definitely easy to use once you understand its details, but I found documentation a bit overwhelming + it took me a bit of time to understand how to set a subscription up. It is not super intuitive with Products and Prices at least not until you realise that they have merged Plans with Products and Prices.

  4. 4

    As a non coder, I have been struggleling to find how to setup a backend for no coders tools such as UIbakery, to integrate user management and payment.

    1. 1

      Interesting. Did you find a solution in the end, or are you still working on that? And just make sure I understand this correctly, the key problem is the combination of user and payments, i.e. some like subscription or product management?

      1. 1

        Uibakery lets you make an Ui and add logic to it, but you need a backend with rest API, like firebase but I haven't been able to work with it. As I understand i need to code.
        I just found https://www.buildwithrebase.com/ which should suit my case if I understood well the product, but it's in an extremely early development phase.

        1. 1

          That sort of defies the purpose of a no-code platform. I can see why this could be quite frustrating - you get 80% of the way real quick just to realise that the last 20% is going to be hard.

          1. 2

            Is not a easy/hard issue, but it's playing in a different field issue. In the latest years amazing tools that let you make a beautiful frontend without coding has been founded (webflow, dorik, etc) but there is not a backend version yet, at least that i know of, but seems something is coming.

            1. 2

              Thanks this information is quite useful to have. I might want to reach out to you at later point in the year if that is OK?

  5. 2

    Not sure which category this falls into, as it's a little bit of both.

    Current pain point is figuring out which modules to outsource (can't afford to hire), and how to spin up the development team on a platform where I can control what cloud resources they use and their deployment pipelines, for when we offboard them.

    We use Azure (it's what I used working a 9-5) and other teams are more used to AWS, GCP, etc.

    Essentially, there's friction when it comes to the deliverable. Right now if you outsource, the dev team goes crazy in god knows which cloud platform, and then zip it up and send it to you when they're paid. This presents two problems:

    1. Non-technical founders have no clue what to do with this zip.
    2. CTO's will have to spend an inordinate amount of time building the deployment pipelines to get that code online, live, and in working order. Additionally, since the prevailing sentiment is "Clients pay for working code, not tests", we have to build all the tests for the deliverable too, and often find some specs were left out, or don't even work.

    It would be amazing if there was a platform for people looking to outsource development, who could connect with a dev or dev team, list the feature specs/requirements, provide a cloud resource subscription on the platform of their choosing, and when the project was finished, the dev team could go back to the original scope and link their E2E tests (many of which offer video recordings of the automated test) to the appropriate features.

    This would:

    1. Allow the customer to immediately pick up where the dev team left off
    2. Give the customer direct visibility into which features were delivered, and which were not
    3. Allow the client to avoid using their own cloud subscription for client development costs
    4. Connect clients with customers and finish projects faster (since all development and delivery happen within a single cloud subscription).
  6. 2

    If you're talking about SaaS for businesses. For me it's two things:

    1. Integration with an existing technology stack.
    2. Managing the ego of the person who currently manages whatever process that your software fixes. If they feel threatened, they will be defensive/talk behind your back to their coworkers/completely ignore you.

    How I get around this? Find the decision-maker that is fed-up by whatever your process fixes.

    I found a business that was being charged ~$850 / month for a tech stack that cost the vendor $50 (I think -- they weren't very sophisticated). The client kept saying "I don't know what I'm getting for this."

    I offered to build a similar stack and be available for questions anytime (be careful with this) for $1,500 up front and $200 / month in recurring charges.

    I now email them 1x a week saying "Hi, is everything still going well?" for the most part they just pay me and ignore my emails.

    Biggest regret? Should have priced higher, but I'm always learning.

    1. 2

      You can always change that, give them a few months (6-12 months) and then come up with upgrades and more value; if you add more value and suggest a better stack/service, they will pay you more if they feel good about you, which sounds like they do

  7. 1

    Front vs. back-end trade-offs - by nature start-ups are cash constrained and there is a constant battle between shipping front-end customer facing features vs. building amazing back-end and data architecture that will enable scalability long-term

    We have found that mapping out potential customer growth against features and forecasting when this would have a drastic impact on scalability a useful proxy for immediate tech spend.

    May not be the perfect solution, as is always in a start-up it's a massive learning journey!

    An example we have faced recently to bring it to life:
    There were a set of features customers were super excited to have added (identified through a design council of 5-10 customers) but we also wanted to introduce v2 of our data architecture to make analytics less impactful on cost and customer. The question was where to spend the $$? We basically mapped out how these features would a) impact conversion of free customers, b) impact onboarding new customers and compared with the point in time where we could not scale due to our data architecture. The answer came back that feature development was right for us but this could have been a different answer if we were 3-4 months ahead.

  8. 1

    Handling several third party code according to:

    1. Dependencies (Code A needs Dependancy X in Version >5 and Code B in Version > 5)
    2. Security (Solution ist here: https://snyk.io/what-is-snyk/)
    3. Maintenance (not longer supported open soruce software (parts))
  9. 1

    I want to learn the basics about that, but also other areas like refunds, expired cards, transactions, updates, etc. Fortunately some tools were added and it let you make a beautiful frontend without coding, but it takes time to fully understand how it works.

  10. 1

    My biggest pain point was to just get going with a huge variety of topics that were new to me. Here's a little story:

    Last week I was invited to an interview at a nearby IT-Security company. They were pretty unprepared for the interview and suddenly came up with the idea to test my development skills through a little coding challenge, which was to retrieve data from a web site (no API).

    Until then I had a huge list of topics in development I wanted to get into, but just hadn't had the time or motivation for, including web scraping and setting up a kubernetes cluster for my projects.

    For the Challenge I chose python - a language I have barely worked with - to build a web scraper - a topic I have barely even heard of. And to make things worse they mentioned that the best applicant took 30min to complete. "Seems impossible", I thought.
    But pressure produces diamonds.

    It took me roughly an hour to setup my environment and work out some installation issues, and another 90min to engineer a quality piece of software, all while working out the basics of a library in a new language, with new IDE and new build tools.

    Finally I was invited for another interview and spent the week setting up my long-awaited kubernetes (k3s) and ceph cluster from scratch, for my personal web- and app-dev projects, including a consulting business I'm building with a friend of mine.

    So what did I (or can you) learn from that?
    Whenever you think there is some topic or tool you want to work on, but think it's out of reach or too complicated, just try it. Iterate fast, test stuff, and do not wait for more information or a better idea. There will be no better YT tutorial, no better documentation, no better book anytime soon.
    And at the end of the day software development is a non-linear process. You will move in the wrong direction from time to time, but you will never ever lose. Because while you move, you will learn things and find solutions that are actually important.

    ----

    TL;DR:
    During an interview I was challenged with writing a tool I didn't know how to write, in a language I didn't know, using a library I haven't even heard of. Completing the task in 2h instead of 2 month taught me that it wasn't that difficult to overcome my imposter syndrome, my anxiety of not knowing the concepts well enough, and my procrastination in hope for some tutorial to premasticate the topic for me.
    Afterwards I took a week to setup a kubernetes cluster, after waiting 6 month for reasons I cannot retrace.
    So stop suffering, start solving. Iterate faster, fail sooner, learn more. The solution might be closer than you think.

    ----

    So this is what caused me the most pain in recent years. It made every other problem I encountered a mildly inconvenient learning experience.

    At last some random quote I found on the internet:
    "A life spent making mistakes is not only more honorable, but more useful than a life spent doing nothing." - George Bernard Shaw

  11. 1

    UX / UI - I am not a specialist, so I often give in to the desire to create a tool "for myself", not for a larger group.

    A CTO who is involved in product development, and not just a subcontractor waiting for orders.

    Integrations - it's so hard to decide if integrations should run through a 3rd party provider or be direct and what in-house team will be needed to keep it going.

    Marketing - despite the fact that I am in the marketing industry myself (or maybe that's why), I think that choosing good tactics to promote SaaS is not an easy matter. You can put a lot of money into the promotion, especially if you promote the application, and this will not guarantee a committed community and social proof. That's why there always should be technical marriage between your marketing and your product, eg. good marketing automation tactic. It's hard to convince your developer it's an important, sprint-worth matter though.

  12. 1

    The biggest technical pain point was building an intuitive user interface that provides a seamless experience to the end-user. It has been somewhat achieved through robust back-end development and design.

    Another one was about automation technologies like chatbots which will help streamline interactions between users and your company (that's still in progress). https://aliliaquat.com/

  13. 1

    The troublesome pain-point according to me, would be
    getting right customers for your software. It includes __marketing, convincing them to subscribe to our product and of course retaining them __.
    Reaching the right customers is very critical..
    Any ideas on how to do it in a smart way(other than Google Adsense and Facebook Ads).
    Waiting to dive deep into the same!!!

  14. 1

    I was alone and couldn't do the selling part. No issues with the technical.

    1. 1

      True, that is something absolutely worth contemplating on!!!

  15. 1

    This comment was deleted 2 years ago.

Trending on Indie Hackers
Yayy! Made my 2nd sale in one month 43 comments Help me positioning my SaaS product 24 comments Need feedback about the landing page 22 comments 🤯Blown Away, Everyday. 20 comments Productized service: Got my 1st client (€2500/m) with 100% upfront payment 17 comments Need Feedback About My Landing Page 14 comments