13
45 Comments

Is it a good idea to build an Open Source Simple Analytics alternative?

I am thinking offering a SaaS solution for a web analytics alternative.

  • Analytics for the web app
  • String privacy policy which does not sell the tracking data, similar to Simple Analytics, but different to Google Analytics
  • Free option (which simple analytics does not offer)
  • Offer self hosted solution

Thanks for everyone's feedback and sharing. :)

To make it clear, I built this small table, this is basically a comparison between my ideas and other solutions. 😀

I know there are many challenges, i.e.

  • maintain a free option
  • high concurrency because of free option
  • how to get better or same performance than others

not only a coding challenge, but more infrastructure/architecture and marketing.

on September 9, 2020
  1. 18

    When I started Simple Analytics I bookmarked new privacy analytics products that came out. Most of them where very similar to my product. I still have that list and when I visit them again, most of them don't even own the domain name anymore.

    I would ask yourself: "Would I really have the motivation and perseverance to make this into real working product?" If the answer is yes, go for it!

    When you say you want to offer a free version. Do note that it's very difficult to make money with it. In hindsight I'm very happy I don't offer a free version and now I can pay my rent and I don't have to find other ways of making money (like selling customers data).

    One more thing on building a privacy business. Do you really care about privacy? How do you prevent yourself from making more money when you shouldn't? You will get customers asking you to build this nice feature for them and they will offer you money for it. Can you turn down the money and stick to your privacy ideals?

    I'm glad there are more people making privacy friendly products out there. But please be genuine. There are many other tools and parts not having a privacy friendly counterpart. Maybe focus on those if you care about privacy.

    1. 1

      I suppose if you wanted to build a privacy-focused business or open-source tool that wasn't analytics, you could go to any popular website, open your browser's cookie panel, visit the domains to see what they do.

      You'll find stuff like hubspot, hotjar, intercom, google analytics and sometimes hundreds more. There might be some untapped hidden gems lurking!

      1. 2

        That's a great way of finding privacy invasive tools Adam. I like your way of thinking.

    2. 1

      thanks. for your long comment and your sharing.

      I totally agree it is hard to be privacy friendly and offer free option. 👍
      and Free option is truly a big challenge.

    3. 1

      I really liked your point that ability to persevere is a major factor.

      Having built some data analytics tools for internal use, I'd add that starting an analytics co is on the challenging end of the spectrum technically - even for our internal tools we needed decent FE, BE, data scientists and devops to cope with the raw events... would be interested to hear about your experience @adriaanvanrossum

      1. 1

        Thanks Harvey! And you are right. It seems simple when you start (and it is simple when you start) but later you will run into complex issues. With 10 customers you don't have to think about all those things (and you shouldn't) but when your base grows you need to.

        One big thing that took a lot of time in the past few months was adding the ability of filtering data. For example, when you click on one referral it would show all countries corresponding to that referral. That feature seems simple at first, but with million of records we needed to do a major upgrade of our database system.

        But my advise for when you start: just use the tools you know and you are familiar with. Then get your first customers. Then optimize your stack if it's required by new features.

  2. 5

    you should check out plausible.io. If you have something unique to bring to the space go for it though

    1. 6

      Agreed. I see so many clones that have nothing unique to add.

      1. 0

        Thanks for your comment.

        How about these two?

        • language, Plausible Exlixir, Alternative Golang or Node, for native better performance
        • pricing plan, support a free plan for 10 k request or another number
        1. 5

          Rewriting it in Go or Node won't necessarily give it a better performance, and Plausible is actually the most performant one I've seen.

        2. 1

          Are you sure that rewriting it into another language will boost performance in this case?

          1. 1

            no I am not sure, it really depends.

            go does have a high performance nature, but it really depends.

            it is basically part of my expect

            @egoist
            @ostap

    2. 1

      thanks, I did not know this website before. it seems good, they does not offer a free pricing plan, I assumed this can be unique, I need to solve the problem is to keep the infrastructure as low as possible to support a free plan.

      1. 2

        They have an open source project.

        Most indie hackers never get 10k requests.

        1. 1

          true,

          but no free plan. :( 😂

          https://plausible.io/#pricing

          so using plausible, it will basically cost you $6/month, if your request amount from 1 to 10k.

          self host is another option, but self host is not free, it cost time and infrastructure fee

          1. 3

            Exactly. That's why it's difficult for us to offer a free cloud hosting plan while trying to run a sustainable startup at the same time. It costs money to host even a site with a few thousands visitors. Free as in beer self-hosting we do offer and put effort into but in many cases doing all that for a site owner costs more money (and time and effort) than getting on a cloud plan.

            But we have tried to make the plans as affordable as we can to the website owners. For instance other products pay for their marketing and affiliates etc which adds a lot of cost to the customer plans while we have decided to grow organically only which means we can reduce the cost of our plans. Plus if you sign up for the full year, the cost is reduced by 33%.

            In the future, it would be great to offer a free plan too but that would mean that we would need more customers on higher plans in order to help us fund the free plan.

            1. 1

              I personally dislike the idea to let customers on higher plans to help maintaining a free plan.

              but I would say plausible is great.

              I just wanted to validate this idea first, thanks a lot for your sharing.

              Free option is truly a big challenge.

              1. 1

                Thanks! Yeah, so either you charge directly for it like we do, or you "charge indirectly" such as what Google does by offering free services in exchange for your data or you fund it with other means (such as having investors or having income from other sources such as paid plans that allow you to take some of that revenue and put it into having a free plan).

          2. 2

            I only pay 2.98 to host umami.is superb...

            1. 1

              What umami means?

            2. 1

              But what if your user base increse still your hosting price is similar or you increase the plan for hosting?

            3. 1

              Very curious, how can you get it hosted in 3.dollars. from I know a reasonable digital ocean vps is $5

                1. 1

                  Nice. How good is hetzner ? All their plans are really competitive compared to DO or AWS.

                  1. 1

                    Well, I guess is fine. I have only used it for 2 weeks now...

    1. 2

      Been using Matomo analytics for couple a weeks now, so far s good - highly recommend!

      1. 1

        we just spun up an instance at the company I work at. I use it for personal projects and the more I use it the more i like it :)

  3. 2

    I am self hosting umami.is

    Love it..

  4. 1

    @here @all

    Thanks for everyone's feedback and sharing. :)

    To make it clear, I built this small table, this is basically a comparison between my ideas and other solutions. 😀

    I know there are many challenges, i.e.

    • maintain a free option
    • high concurrency because of free option
    • how to get better or same performance than others

    not only a coding challenge, but more infrastructure/architecture and marketing.

  5. 1

    I would say it's not a bad project to write if you're trying to learn a particular stack or language but it's not really a missing project.

    1. 1

      I built this to make things clear, text is not that direct.

      not very different, but still it is unique

  6. 1

    There seems to be lots of alternatives to Google Analytics popping up. What would make yours different?

    1. 1

      thanks for your question.

      this is it

  7. 1

    It's pretty intesting watching this privacy space grow.

    Here is where my thoughts are as a product owner, with a B2B SaaS.

    • Tracking conversions is really important.
    • Tracking key events and goals is really important.
    • When you run PPC and do content marketing, tracking your sources, conversions, funnels are important.

    As you grow as a business you care about the above things. These privacy focused tools just can't match these features. But maybe that's by design too?

    It's not that I don't care about privacy, but I also don't want to run my business blind. Which is what it feels like sometimes when using these tools like fathom, etc.

    1. 1

      thanks for your sharing.

      yeah, the case from your business is the case non privacy friendly solution can help.

  8. 1

    Speaking about open-source solutions:

    I was looking for open-source analytics, but I didn't found some with API.

    API can be a competitive feature.

    1. 1

      thanks, a really good point! 👍

    2. 1

      Matomo has API.

  9. 1

    It's not such a good idea anymore because the list of alternatives is growing.

    I currently list 6 alternatives that are Open Source in my Google Analytics alternatives list.

    As a user, I would say go for it! If someone brings something better to the table I would give it a go.

    However, I believe there are many options now and starting today is a disadvantage.

    I would rather recommend you to join efforts with some existing Open Source solution and think of a niche that you can serve by improving it in certain way and offering your own hosting.

  10. 1

    I think the main thing is, how small and tiny it would affect your website. When it comes to adding extra JS files , it's very sensitive. Is that extra weight and connection worth it? How many % does it eat up ? Especially when it's overloaded, you won't want a file trying to connect for 5 seconds.

    @indie321 do you have a telegram handler?

    1. 1

      thanks.

      basically, for an analytics solution, there will be two parts.

      1. analytics JS script
      2. server receive the ajax request

      CDN (cloudflare or cloudfront can help the first one), it won't slow your app/web, and it is basically an optional script (after footer not before head), you basically won't feel it.

      The second problem, but a failing request will not slow your app, what the lose is one request of user's analytics data, it is bit of more challenging.

      so 5 seconds waiting, if CDN infrastructure not down, it will never happen, not even 1 second.

  11. 1

    Did you have a look at https://umami.is/ ?
    It is free, simple, open source and privacy oriented. I really like it.

    1. 1

      Yes, and I did some small contribution to that project too.

      I really like it too.

    2. 1

      Me too. Hosting it..

  12. 1

    This comment was deleted 3 years ago