Hacker News new | past | comments | ask | show | jobs | submit login
Lessons from building Plausible Analytics to $1.2M ARR in public (buildinpublichub.substack.com)
210 points by trulykp on March 12, 2023 | hide | past | favorite | 76 comments



If you aspire to run an indie software business like this, you might want to take note of how absolutely boring most of the success in this space is.

Web analytics is boring. Cron job monitoring is boring. Running a backup service is boring. These businesses (the cron one is mine) are all doing great, making millions. But in this space the vast majority of devs are chasing fun and cool projects in web3 and AI that are more tied to cool ideas than actual business problems.

Businesses are, mostly, boring things. They have boring problems. If you’re doing something too interesting there’s a good chance you’re just playing around


How to find these boring problems? Every time I look at a boring problem in tech there is a pretty okayish solution, so I'm discouraged to even try to make something, as I know that okayish solution will be even better when I get mine to work.


> Every time I look at a boring problem in tech there is a pretty okayish solution

Congratulations. It means your ideas are good and there is a proven market for them (assuming the solutions you find are generating revenue).

The fallacy that “there already is a product” out there doing something and as such you shouldnt would imply that we’d all wear identically looking jeans and there would be one car design and one type of phone.

Differentiate a bit. The only questions to answer are: can you execute on your idea, is there a market large enough to achieve your target revenue and is it overcrowded?


Even 'differentiate a bit' might be the same instinct biting.

Really, just market and sell adequately enough to grab the share of the market that you need is all it comes down to. Your product might even be a worse copy of an established competitor and this still holds true.

It's so rarely about more or different features, better UX, or innovative twists on ideas, and so often about just getting in front of the customer in the right place at the right time.


How do you find out if it’s overcrowded or not? Supposedly with a good enough differentiator it doesn’t really matter if it’s overcrowded right?


Talk to the people who have the problem you are solving (or the people making purchasing decisions for those people). Do they know about the existing solutions? Do they use any? Would they be willing to switch? Have they switched in the past?


Riffing..

Take one of those ideas. Make a site that reviews all the current competitors in the market.

Put the one with the best existing seo strategy in first place.

Contact them and let them know and ask that they link to it, offer to write a guest blog post about how you reviewed every tool in the space and this one came out on top, and why.

Share the site all over where you can find people talking about the niche.

if the niche is not too big, this will often get you ranking for that term. Google loves “best ___ in 2023” content.

You are building a traffic source.

Then you can build a product and put your product in 1st place or even above the list entirely. Write a blog post about how you took the best of the other tool you reviewed and created the perfect tool for <certain persona>.

Ok, end riff, i have no idea if that will work but it seems repeatable and not too difficult. It really is all about getting traffic.


Just for clarity, is the product you build in the same category as the products you initially reviewed?


If you want an idea: A SaaS to host product guides with good i18n.

- Gitbook.io doesn't support i18n well. Not SEO friendly and not language detector. - Readme.io is too expensive for just product docs. - Hosting with NextJS etc.. is a pain if you need to support i18n and edit content using markdown.


I’m working on something now that solves just this.

I had to build out a Next app last year with a seamless looking blog and docs site. I struggled to find real solutions, so I ended up getting in touch with Marco from Plausible to get some advice. He helped me to figure out their stack, and I replicated it to my project. I released a dope next app with a markdown powered blog and docs site that looked amazing.

I’m working on releasing a free cli tool for frontend devs based off this work. I’d love some feedback and watches on GitHub if anyone is interested. Link is in my bio.

SHOW HN and IH coming soon :)


For your interest, I clicked through to your website from Github and it gave me a 'site not secure' prompt from my browser (desktop Firefox).


Thank you for checking it out! Much appreciated.

Yeah, I have a few hypothesis that I am going to prove using the Elegant tool that I’m working on.

The first hypothesis is that I’m going to use Elegant itself to scaffold out the documentation. So when the domain is live, Elegant will be ready for beta testing. I hope to get some content up on the domain this week!

Thanks for pointing this out.


How about now? Is it still saying it’s not secure?


Please let me know if you have a beta or link so we can test it as soon is live.


Hey Marco, check out the MVP that I launched! Could I get some feedback?


I would created as a Saas with: - multi language. Default translated using deepL and then a translator can check and review using a WYSIWYG editor. - alternative html SEO tag for Google. - Tailwind syntax template


GitBook founder here :)

We are working on multiple aspects of the products that should cover a lot of this (improved i18n, SEO, faster rendering)!

Would love to get your feedback on what you would like to see improved exactly on our SEO friendliness and i18n support. We have an open GitHub community for feedback; https://github.com/GitbookIO/community/discussions

And for anyone, if you are interested in building a "A SaaS to host product guides ", we are hiring engineers/designers/builders: https://jobs.gitbook.com/ :)


It's been 2 years that I suggest to implement i18n. Or at least let add a JS snipped to use WeglotJS


Mintlify are probably the guys who will close on this one first. Great team, moving very fast.


They do not support multi language.


document360 support i18n


In reality, you just have to get started.

Keep making things, iterating, grinding away.

Some things stick, some things don't.

You network and grow your business along the way.

That's the boring journey, but the reality for most of us.


I agree. As you build you get a better sense for what people want. You also figure out how to build faster as you collect tech stack, or even more than one. When the right idea comes along, you can then move quite quickly.


AI will probably be boring eventually since it's a large umbrella. Even if you scope down to LLMs and NLP, gluing those to work with your stuff in interesting ways will still seem boring compared to publishing papers.

What I never got about businesses doing technically boring things is that only the marketing/sales layer is adding value, so it's interesting people would chose a newer business over an older one. If you have a defining feature that sets you apart, that's not boring, it's innovative and interesting. So maybe web analytics/cron job/backups are still not boring because something is keeping major companies from crushing those specialties.

Web3's always been shit though.


I think the important thing with AI is to start at the customer and work backwards to AI. If you are trying to think of cool things you can build with the new AI APIs you are probably doing it wrong.

One thing to remember about having competition is that most customers are not doing a “bake off” and they probably are not even aware of your competitors. If they found you, and you seem reputable, and you do what they need, and the price seems worth it, you can probably close the sale.

Of course the first one is key. That’s the hardest thing about starting from scratch. They have to find you.


As an old business mentor once told me: if you've found a niche where there is no competition, the odds are very good that there's some huge problem around addressing it that you haven't thought of. If nobody else is fishing in the lake, it's probably because there are no fish, the fish are poisonous, or there's a crazy person around who will shoot you for fishing there.


> AI will probably be boring eventually since it's a large umbrella.

it depends on if advancements will be slowdown. If capabilities will start growing exponentially aka singularity, then we will have very aggressive ride to the new level of humanity development.


Yep, we run an agency doing very boring things with Ruby on Rails. Not chasing the latest and greatest.

We also have side projects/businesses that solve very boring problems in boring old fashioned ways using Ruby on Rails.

And we are making a living in a remotish part of Australia and I don't have to commute and answer to people I dont like very much! :-)


> Web analytics is boring. Cron job monitoring is boring. Running a backup service is boring...

I think you are missing part of finding customers in crowded market. It is very not boring and not trivial.


Web analytics isn't boring to me. I also think it's a really hard area to break into. There's no shortage of competition and none of them seem very differentiated.


I really don't think there is any sort of programming that is inherently boring. There can be programming that doesn't intersect with a given individual's interest, but that's different.

For every sort of programming I can think of, I can think of people who find it boring (even machine learning!) and people who find it exciting (even cron monitoring!).


Cronitor.io is his business in case anyone is interested.


Thank you. Was wondering about that - no link in bio.

Clickable link:

https://cronitor.io

Ed: (business) might be built in the open - but (product) isn't F/OSS?


> the cron one is mine

It did stood out from the list actually.

Web analytics is very interesting from the UI/UX perspective. In fact that's exactly what make Plausible notable - their super-polished wonderful UI.

Backup service is no less interesting, but from the technical perspective, especially around hardware reliability issues.

Cron monitoring - yep, boring as hell.


Not sure if it’s touched on but one cofounder focuses on product while the other focuses on marketing. Cannot stress enough how much of a winning formula this is


This. There's a graph somewhere on their blogs showing traffic before and after certain events including when Marko joined. It was, to a first approximation, flat before that.

It's incredibly powerful if you can have one cofounder (ideally who's good at marketing like Marko obviously is) focus on marketing exclusively.


And one or both focuses on sales.


so true!


I used plausible and it was great. Then I found Beam analytics, which is a copycat with a better free plan. I switched in no time to Beam Analytics now.

I would be more worried about copycats. When you share your success, other developers want to have it too.

Even if you believe you are for the long run you lose money in the short term because of your ego.


There are offsetting effects from building in public. There's the downside that it encourages copycats, but sharing numbers and strategy also wins positive public attention.

I personally pay for Plausible because I like how much they share. They frequently pop up in discussion on HN, and I think their openness drives a lot of that mindshare among the crowd here.

If it's a service I care about, I don't want to be on the free tier because I want a mutually beneficial relationship with my service provider. It's cheaper for me to pay a few hundred dollars per year than to have to scramble and switch platforms when my vendor folds or drops their free tier.


Well said!

I've been a paying customer for over 2 years and their transparency and overall a strong "sharing" culture played a big role. It's just easier to trust.


One huge selling point of using Plausible is that you can self-host it and it's FOSS. If the company disappears, you can continue using what got developed so far, and maintain it yourself. Beam Analytics doesn't seem to feature any of those two things, making it less of an alternative, at least for a sub-section of the users of Plausible.


But in the end, they grew it to $1.2M ARR. That some very cost sensitive people are finding other ways is not much of a loss because it's very hard to get money from those people anyway.

We are building windmill.dev this way and fully open-source, I'm less worried about copy-cats than about the project not taking off because of too much friction.


Nice job with Windmill! Would you be interested in listing Windmill on https://starter.place?


Doesn't seem to be actually a clone though, at least not of current Plausible, because if it is, they are currently breaking the AGPL license.

But Plausible was MIT licensed until 2020, so maybe a clone of early Plausible?


Wow, I thought by copycat you meant that they went for a similar feature set, but from screenshots it looks identical down to font and color choices. Are they just running a white labeled instance of Plausible’s OSS?


It's not white labelled Plausible. The creators of it are on Twitter if you are interested in following their journey (i.e. twitter/TheBuilderJR). It's built on Tinybird, Supabase, and Next.js, while Plausible is built in Elixir using plain Postgres and Clickhouse, and Phoenix.

A lot of relatively new analytics services tend to have similar looking frontends these days I've noticed though.


Just in case, Tinybird is built on top of ClickHouse. If you look deep enough into any analytics service, you'll find ClickHouse inside.


Yeah, I was actually looking at the idea of using it to make a little analytics site for my hobby projects last week :). The reason I mentioned it separately though was that I would assume Plausible has written a lot of code to deal with things that using Tinybird means you don't have to worry about.


…and plausible is a direct ripoff of Fathom analytics (design by Paul Jarvis).

A good rule of thumb: if you think something is original, chances are it’s because you’re just not very familiar with the space.

Ultimately, none of this matters. The goal of a product should be to help customers achieve the outcomes they want, not create something original.


Hi - co-founder of Beam here. Appreciate the discussion about our product in this context. We've tried to differentiate our product from the other GDPR compliant web analytics by also focusing on product analytics proxies which are easy to implement and interpret. We think our funnel analysis and cohort retention tools can be very helpful. To learn more about why we built Beam, check out this blog post - https://beamanalytics.io/blog/why-we-started-beam


Thanks for the context. Looks like a cool product and (as someone who hates cookie banners) I'm excited to see more GDPR compliant providers.


Not possible to self host though.


Actually it definitely is possible to self host plausible analytics! It's an open source project with an AGPL license. You can even find instructions for self hosting here: https://plausible.io/docs/self-hosting

Please notice that although they recommend docker for self hosting, it is very easy to self host an a bare RHEL server (you need to be a little familiar with Elixir though but not so much, you can take a peek at the docker entrypoint scripts to see what's going on) ; I'm doing exactly that for some months now and I'm happy with the results.


GP was referring to the clone (the Beam thing), not PA.


I like what Plausible is doing, and they definitely know how to make content marketing work, but I'm going to be honest - the pricing model they're using is a bit out of this world. On top of everything else you have to pay (subscribe for) on the web these days, paying $500 a year for 500k monthly pageviews is insane!

Now, obviously, my opinion doesn't hold a lot of weight by itself because it seems that they're quite profitable, but I will never understand that kind of pricing for something so basic.


If you've got 500,000 monthly pageviews and $42/month isn't a rounding error to you, I think there's something wrong with your business.

And if you're not running a business, why do you care about analytics at all?

Most medium-large sized companies will pay people $150,000/yr just to translate the data from tools like plausible into ugly powerpoint presentations for management.


Two assumptions there I'd challenge.

1) Pageviews do not equal revenue. The purpose of every website isn't inherently to make money.

2) The purpose of analytics is to understand your visitors and their interaction with content. There doesn't need to be a profit motive to want to do this. I've worked on plenty of nonprofit, informational, and discussion sites where we use analytics to discover what content is resonating with what kind of an audience, how people are discovering the site, what paths people take through the content, where errors are occurring, or just congratulate authors for writing things that got a lot of reads.


Why not free Google analytics? "...but evil big tech and data privacy"

Why not use Plausible self hosted (it's open source and free)? "...but it's not actually free if you have to spend many hours setting it up and maintaining it"

Forgive my snark, but that's usually how these conversations tend to go.

Even if you're a non-profit, if your website analytics on 500k monthly views isn't helping maximize donations by more than $42/month, you're doing something wrong.

Ultimately, the main takeaway is, it doesn't matter how low or high the price is. Some people just believe they're entitled to a world where everything works perfectly for their needs and they never have to pay for it. I suggest ignoring those people--because you'll never make them happy.


I'm not sure who you're arguing with about GA and self-hosted Plausible. I use each, for different projects in different circumstances.

> Even if you're a non-profit, if your website analytics on 500k monthly views isn't helping maximize donations by more than $42/month, you're doing something wrong.

Again, not every website exists to make money. My friend's foodie blog. My wife's not-for-profit farm project homepage. My local community discussion forum. The little page I threw together to support another friend's open source project. Going back to the original comment, these sites 1) could use analytics that respect visitor's freedom and 2) have a budget of fewer than $42 a month for the entire project.

You might be surprised at how many zero-budget hobby sites get to 500k page views a month. Especially discussion forums. I think I may have hit that threshold with a WarCraft II discussion forum I ran when I was a teen in the 90s. But back then, we stuck anything on our sites with abandon, including the "hit counters" I used for "analytics", and all sorts of guestbooks and other nonsense. Not just JavaScript, we'd drop Perl and PHP scripts straight into our CGI-BIN folders that did who-knows-what with user data. I just would have hoped the state of web analytics for the hobbyist website operator wouldn't be quite so depressing 25 years later.


> You might be surprised at how many zero-budget hobby sites get to 500k page views a month.

Similarly, you might be surprised how much it can cost to support a SaaS service scaled for a great many 500K+ views/month sites: infrastructure for recording the data and rendering it back out as reports, monitoring and growin/reshaping that infrastructure as needed, meeting user support expectations, etc.

The SaaS service is operating as a business, not a pubic service.

If there is something I learned releasing simple software and slapping websites together back in the late 90s, is that there are no end of people who are convinced you are missing a trick by not offering what they want under the conditions (price usually being a key one of those conditions) they want it. You can't make all of them happy, you can't serve every market expectation, to try is to guarantee burnout of you business, yourself, and (if more than a one-man-band) your people.

Of course if they've genuinely missed a golden combination of features and pricing strategy, why not capitalise on that untapped market? Or use another product that is trying to tap it already.


I'm not arguing anyone should provide this service for free. I'm suggesting "if your website analytics on 500k monthly views isn't helping maximize donations by more than $42/month, you're doing something wrong" is not helpful, nor is "if you're not running a business, why do you care about analytics at all?"

While there are some corners of the web dedicated to making money, that's not a goal for many (most?) websites, and yes, those websites can also use analytics.

Maybe the only way to get analytics for those webmasters is to subsidize the cost by handing personal data on every single visitor over to Google, because maybe it really does cost $500 a year per site for third party to host analytics software for you. If so, we spent three decades creating what turns out to be a pretty shitty experience for creators and consumers on the web.


> Why not use Plausible self hosted (it's open source and free)? "...but it's not actually free if you have to spend many hours setting it up and maintaining it"

Not really. For example I use Plausible self-hosted through Coolify (open source Heroku alternative) which has a one click add service feature, it has quite a few services including a full blown Nextcloud instance too I believe. It works well because it's all Docker.

I don't have to worry about it after the one click. It even auto updates.


What should a content site relying on contextual ads be expecting to make from 500,000 pageviews per month?


I think it only looks insane because Google is “free”. I’ve been running a SaaS profitably for about a decade (and with employees for half that time), we’re still well below that.

We’re plausible customers for our new thing because we wanted to take a privacy first approach. No need for a cookie banner, no weird permissions or opt out. Just a functional website. I understand GA is crazy powerful, but for what I’m actually using Plausible is so much easier to use.


If they make it affordable enough for high-traffic non-profit websites to use on a whim, they will be leaving a lot of money on the table from web businesses who do actually make money – and those businesses are actually the best customers for their SaaS – willing to pay more money for less traffic.

I'm also not using them for my casual projects because I can't afford their cloud service, and I can't be bothered to self host, but they'll be totally fine without me.


Here's an interview with Plausible (an alternative to Google Analytics) co-founder Marko on the journey of going from 0 to $1.2m in ARR and building as much as possible in public


Recently switched to Plausible from GA for my blog. Analytics are way easier to use compared with GA4. Probably would not have even considered switching if they would have left universal analytics. Privacy stuff and performance we're secondary bonuses. Worth every penny.


Same here.

I went with self hosted version for my personal sites.

Simple features and nothing too complex. Mostly I watch real-time stats, monthly page views and daily 404 pages.


I always feel a bit depressed that so many of these microbusinesses seem focused on sales pipeline stuff. I get it, but at one point it feels like money just kind of sloshing around between various salestech companies.


I am a paying PA customer and I love the service but I am kinda surprised they only just reached those figures. They have great market awareness in my circles but it just goes to show that developers are stingy and conservative and it's a hard market to do well in. If feel they should b making 10x this number by now given they are internationally known and the biggest alt to a gigantic usecase


Inspiring read! I used Plausible back in the early days and it was a great product, I'm not surprised they made it big!


Oh, plausible has goals now ? Time to revisit it to push it again in my org :).


The founders refuse to adopt tracking of user flows through a web site, allegedly for privacy reasons.


Do you have reference for that? https://plausible.io/docs/top-pages#how-it-works seems to suggest that flows are recorded unless you are routing based on query params, unless you are meaning something else by user flows.


https://github.com/plausible/analytics/discussions/149

> @ukutaht, Apr 27, 2021, Maintainer > The vision of this project is and has always been to be aggregate-only with no reporting on individual users or sessions.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: