Serious question: what is the value of web analytics for people?
I run a SaaS business and I dropped Google Analytics a long, long time ago. Primarily because of the tracking, but also because I really couldn't see the value of the data.
In the old days, you could at least use the "Referer" (sic) header to know where people came from and what they searched for. But that is long gone, and the only source of that data is Google/Bing search console.
Page visits are a vanity metric: they tell me nothing about my business. The only thing that actually matters for a SaaS are signups and MRR. Measuring your business by page views is like measuring the business performance of a Walmart by counting cars on the freeway nearby. Yes, the numbers are somewhat related, but you can't draw any conclusions.
I made it a point not to include any third-party JavaScript on my site, but even if I were to make an exception for these analytics, I can't really see the point, unless you are running an ad-driven site where pageviews are king.
This seems contrarian just for contrarian sake, given how much literature there is about this, and the fact that it's almost self evident. Tracking impact of your changes, seeing if your users are getting lost after changing something, understanding where they spend the most time, etc.
Say for example, if all your users start spending 30% more time in your reset password page after you pushed out some changes. How would you know? What could be causes of that? Could something be broken with the login? Apply this to everything.
Not having analytics is literally not caring about what they do in your product, so you're either never changing the product and 100% confident it'll always work, or you're probably giving them a worse experience than you could.
How you do this tracking is another story, but there's ethical ways to do it.
> Tracking impact of your changes, seeing if your users are getting lost after changing something [...]
The change of adding obnoxious tracking of course accounts for some user loss itself, which it cannot measure.
On some of those "modern" websites, that show me a whitescreen without JS, I check my uBlockOrigin and see the domain of that website and some Google shit? Tab closed. No thank you, I will go elsewhere.
Sales is driven through traffic. No traffic == no sales.
Understanding what drives traffic to your SaaS website is such an important piece of information. For instance, if you write two articles, one describing how to use your product to achieve a certain thing which customers want to do, and another article which compares your product to a competitor product and one of the two articles creates 50x more traffic than the other then you'd certainly want to know this, because then you know what articles give you the biggest return on your time writing them.
Just one of so many examples how web analytics is such an important tool to being a good sales person.
That sounds like a non-example. Why do you need invasive, personalized surveillance for that? Traffic and aggregate data are an entirely different question.
Do you not at least want to know page views ratio to sign ups so you can see conversion rate? Or do you have a different / better way to do things like testing a new design / price.
Only page view? That's not really useful and you already got that with backend logs.
With true analytics, understanding typical session helps you optimising users workflow, making sure relevant features are easily discovered at the right place.
It really helps when you want to work on user experience. You may need metrics such as LCP, INP and CLS with details per type of page, ability to drill down data and get that in real time.
ROI of such script depends on what you do with the data. If that's vanity or not even looked at, you are emitting CO2 for nothing.
>optimising users workflow, making sure relevant features are easily discovered
>work on user experience
These are qualitative improvements which are extremely unlikely to stem from quantitative metrics, especially when the sample size is not significant (which it is for the vast majority of pages in existence).
In around 8 years of web development, mostly focused on consulting and focusing on ecommerce, I've never seen a net gain from using analytics on a site. If the end goal is to produce data for the sake of data, well sure that will work. Rarely does anyone analyze the data though, and I've never seen anyone dig into the validity of the data and ensure that Google Analytics is in fact accurate and reliable for them.
One of the most disappointing client experiences I had was after building a custom shop for a company that was heavily focused on graphic art. We optimized the hell out of their site, getting performance scores of 97+ when every page was image heavy and included a product grid designed for a masonry grid look similar to Pinterest.
A few days before launch they asked us to add their Google Pixel script. The next day they had included 7 or 8 different third party scripts and blown performance scores into the mid 50s. Its their site and they can do what they want with it, but I sure could have saved a lot of dev time if performance didn't matter at all.
Page visits tell you have many people you get. If you then use how many sign ups you get then you have a conversion rate. That’s an important figure. Page visits can also tell you if your marketing efforts have worked. Imagine doing all the marketing work and not knowing if it did anything.
OP clearly stated that signups and MRR are the really important figures for SaaS. Not incidentally, those two metrics also tell you if your marketing efforts are working.
The value of web analytics for our organization lies in the same realm as the value of Plausible over any third party analytics: The Funnel.
We're a membership driven organization, and by "membership" I mean we rely on donations to fund our content creation (Though whether you're a member or not you have the same level of access to our content). We care about raw traffic numbers, because it relates directly to our mission of informing people. It tells us how many people we inform day to day.
So yeah we care about those raw numbers, and those numbers are difficult to get w/out javaScript r/n because caching and the terrible log retention of our hosting providers.
Raw traffic numbers only tell part of the story though. We want to know the path people take from first landing on the site to becoming a donating member so that (in theory) we can do more of the things that promote that behavior in more people. That's The Funnel, and that's where orgs like Plausible are best. They're first party tracking, so the data stays with us. Also since they're first party tracking we can track a person's overall relationship with our site, from the first news story they read to the moment they first hit our donation page 3 years into the relationship or whatever.
We should be able to do that with our GA set up, but one of the reasons I want us to shift to Plausible is for its simplicity.
You got quite a few seo garbage-level nonsense replies. In my experience, you are right, and most tracking metrics have long since become the (vanity) goal to justify the existence of these 'digital marketers'.
It's funny that they spout nonsense about better UX or how you wouldn't be able to do CRO when you'd just laid out two metrics that are actually important and don't require any website analytics to track.
Cool. Perhaps companies and governments (gov.uk I'm looking at you) could consider using this stuff instead of forwarding all their public interactions to an unaccountable US corp.
Let me also plug my free, open-source and self-hosted event-based analytics solution: Fugu (https://github.com/shafy/fugu). Fugu does not track unique visitors (not even daily like Plausible does) and is made for event-based tracking. Comes with included Docker config to make it easy breezy to self-host.
I'm planning on running a small niche WordPress blog that I would like to monetize with adsense & possibly an affiliate program. I see there's a lot of choices for analytics available listed by users in this post. Does Adsense require Google Analytics or could I use one of these more privacy friendly ones?
AdSense and Google Analytics are two separate products, you can use either, both or neither. If you have AdSense though, you've already allowing Google to track your users, so I don't think ditching Analytics would make the blog any more private.
It's also cookieless, the hosted version is free to use within reason, and it's extremely lightweight if you choose to self-host it. It doesn't even need a separate database, it can run self-contained with SQLite (or Postgres if you prefer). A good fit for small sites where the big industrial-grade solutions are overkill.
This service claims to not track personal data, yet their docs admit to storing hash(siteID + User-Agent + IP) + seen_paths on their backend for session tracking.[1]
Sites can track sessions without tracking personal data.
right below that the docs also say that this hash is not persisted, only cached in memory and mapped to a UUIDv4. The UUIDv4 is what persists between sessions.
> The IP address and User-Agent are never stored to the database or disk, and there is no conceivable way to trace the random UUID back to this.
>
> It’s only stored in memory, which is needed anyway for basic networking to work.
I can't say whether that is GPDR compliant but it's definitely not storing the hash
„ In comparison, in the context of the European GDPR, the Article 29 Working Party[6] considered hashing to be a technique for pseudonymization that “reduces the linkability of a dataset with the original identity of a data subject” and thus “is a useful security measure,” but is “not a method of anonymisation.”[7] In other words, from the perspective of the Article 29 Working Party, while hashing might be a useful security technique, it is not sufficient to convert personal data into deidentified data.“
Currently using Umami, but I've considered switching to Plausible due to Umami's less-than-stellar development performance (e.g. breaking the site details page for a few days recently).
Also happily using hosted GoatCounter. Last year I noticed some occasional operational hiccups, like service brief downtime, but this year it's been completely stable as far as I can tell.
How are you self hosting it? I find its requirements extremely heavy for a simple analytics solution. It requires a PostgreSQL and Clickhouse database. I don't find self hosting Clickhouse particularly easy. Wish they had an option to just use SQLite as an alternative.
I completely agree that the self-hosting story for Plausible is overkill for most websites.
So much so that I made my own that focuses on self-hostability using SQLite and DuckDB (no external dependencies, can run on a 256MB VM): https://github.com/medama-io/medama
I've not tried their hosted version no. I doubt that there would be a seamless way to switch between them since all the data lives in Clickhouse, but I could be wrong.
Can I use plausible in a desktop application? I would like to have an idea of exactly which versions of an open source desktop app I maintain are being actively used so I know what to pay attention and invest efforts as I would like my users to be constantly migrating forward - we do have like 20 years of backwards compatibility so we push things forward very slowly.
I was able to do it pretty easily with a mobile app, should be just as easy on desktop. You could even register custom “pages” for various parts of the desktop app.
Readit News
I run a SaaS business and I dropped Google Analytics a long, long time ago. Primarily because of the tracking, but also because I really couldn't see the value of the data.
In the old days, you could at least use the "Referer" (sic) header to know where people came from and what they searched for. But that is long gone, and the only source of that data is Google/Bing search console.
Page visits are a vanity metric: they tell me nothing about my business. The only thing that actually matters for a SaaS are signups and MRR. Measuring your business by page views is like measuring the business performance of a Walmart by counting cars on the freeway nearby. Yes, the numbers are somewhat related, but you can't draw any conclusions.
I made it a point not to include any third-party JavaScript on my site, but even if I were to make an exception for these analytics, I can't really see the point, unless you are running an ad-driven site where pageviews are king.
Say for example, if all your users start spending 30% more time in your reset password page after you pushed out some changes. How would you know? What could be causes of that? Could something be broken with the login? Apply this to everything.
Not having analytics is literally not caring about what they do in your product, so you're either never changing the product and 100% confident it'll always work, or you're probably giving them a worse experience than you could.
How you do this tracking is another story, but there's ethical ways to do it.
The change of adding obnoxious tracking of course accounts for some user loss itself, which it cannot measure. On some of those "modern" websites, that show me a whitescreen without JS, I check my uBlockOrigin and see the domain of that website and some Google shit? Tab closed. No thank you, I will go elsewhere.
Understanding what drives traffic to your SaaS website is such an important piece of information. For instance, if you write two articles, one describing how to use your product to achieve a certain thing which customers want to do, and another article which compares your product to a competitor product and one of the two articles creates 50x more traffic than the other then you'd certainly want to know this, because then you know what articles give you the biggest return on your time writing them.
Just one of so many examples how web analytics is such an important tool to being a good sales person.
With true analytics, understanding typical session helps you optimising users workflow, making sure relevant features are easily discovered at the right place.
It really helps when you want to work on user experience. You may need metrics such as LCP, INP and CLS with details per type of page, ability to drill down data and get that in real time.
ROI of such script depends on what you do with the data. If that's vanity or not even looked at, you are emitting CO2 for nothing.
These are qualitative improvements which are extremely unlikely to stem from quantitative metrics, especially when the sample size is not significant (which it is for the vast majority of pages in existence).
One of the most disappointing client experiences I had was after building a custom shop for a company that was heavily focused on graphic art. We optimized the hell out of their site, getting performance scores of 97+ when every page was image heavy and included a product grid designed for a masonry grid look similar to Pinterest.
A few days before launch they asked us to add their Google Pixel script. The next day they had included 7 or 8 different third party scripts and blown performance scores into the mid 50s. Its their site and they can do what they want with it, but I sure could have saved a lot of dev time if performance didn't matter at all.
Page visits tell you have many people you get. If you then use how many sign ups you get then you have a conversion rate. That’s an important figure. Page visits can also tell you if your marketing efforts have worked. Imagine doing all the marketing work and not knowing if it did anything.
We're a membership driven organization, and by "membership" I mean we rely on donations to fund our content creation (Though whether you're a member or not you have the same level of access to our content). We care about raw traffic numbers, because it relates directly to our mission of informing people. It tells us how many people we inform day to day.
So yeah we care about those raw numbers, and those numbers are difficult to get w/out javaScript r/n because caching and the terrible log retention of our hosting providers.
Raw traffic numbers only tell part of the story though. We want to know the path people take from first landing on the site to becoming a donating member so that (in theory) we can do more of the things that promote that behavior in more people. That's The Funnel, and that's where orgs like Plausible are best. They're first party tracking, so the data stays with us. Also since they're first party tracking we can track a person's overall relationship with our site, from the first news story they read to the moment they first hit our donation page 3 years into the relationship or whatever.
We should be able to do that with our GA set up, but one of the reasons I want us to shift to Plausible is for its simplicity.
It's funny that they spout nonsense about better UX or how you wouldn't be able to do CRO when you'd just laid out two metrics that are actually important and don't require any website analytics to track.
Their most recent blog post:
Things I hate about GA4
https://plausible.io/blog/things-i-hate-about-GA4
(https://news.ycombinator.com/item?id=40904139)
It's also cookieless, the hosted version is free to use within reason, and it's extremely lightweight if you choose to self-host it. It doesn't even need a separate database, it can run self-contained with SQLite (or Postgres if you prefer). A good fit for small sites where the big industrial-grade solutions are overkill.
Sites can track sessions without tracking personal data.
1. https://www.goatcounter.com/help/sessions
> The IP address and User-Agent are never stored to the database or disk, and there is no conceivable way to trace the random UUID back to this. > > It’s only stored in memory, which is needed anyway for basic networking to work.
I can't say whether that is GPDR compliant but it's definitely not storing the hash
Could you detail how that would work?
https://www.gtlaw-dataprivacydish.com/2021/03/what-is-hashin...
Deleted Comment
So much so that I made my own that focuses on self-hostability using SQLite and DuckDB (no external dependencies, can run on a 256MB VM): https://github.com/medama-io/medama
Stuck it behind a NGINX frontend and it works just fine.
https://github.com/Glimesh/glimesh_app/blob/main/lib/track.d...