8 ms·

I see one risk for Firefox: if they block all trackers, e.g. Google Analytics, then it would appear to people viewing browser statistics in Google Analytics tha

I see one risk for Firefox: if they block all trackers, e.g. Google Analytics, then it would appear to people viewing browser statistics in Google Analytics that no-one is using Firefox, and thus the developers will be told they don't need to test anything on Firefox.

bn7t 7y ago
Website owners can just use the user agent in weblogs to determine this.
- dajonker 7y ago
  Yes they can. But in my experience, at big organizations, the people using Google Analytics are not developers and will have no idea what weblogs are or how they could use them to determine the real browser statistics.
  - NotASithLord 7y ago
    I'm a dev and I have no idea what weblogs have to do with this discussion. Quick search didn't do anything to clue me in either.
    - 0xffff2 7y ago
      Which part is confusing? The browser sends lots of data to the server with each request. The server usually logs at least some of that data. Among other things, that data can be used to generate a subset of the data that is typically gathered by GA and other client-side analytics tools.
zaarn 7y ago
Nobody looks at logs anymore? Strange, those are very useful. Google Analytics isn't the only analytics tool and as blockers have become more prevalent relying on GA as your sole source of truth is becoming increasingly desperate to worship Google. Ideally you run a simple analytics tool alongside that is lightweight and doesn't get blocked (ie, privacy friendly) as well as running a log analyzer on your apache/webserver logs.
- geraltofrivia 7y ago
  I host my personal site (https://priyansh.page https://priyansh.page) on Netlify and they give me analytics computed from server side logs. [Link to Netlify Analytics](https://www.netlify.com/products/analytics/ https://www.netlify.com/products/analytics/). They don't provide a lot of demographic information (even browser used, afaik) but provide sufficient information for my purpose. edit: I should mention that this a paid feature ($9 a month). edit2: I do notice a higher number of traffic when compared to some free analytics which work on the client side, and respect Do Not Track.
- oarsinsync 7y ago
  > Nobody looks at logs anymore? Apparently not, as one other comment on this site lists one of the apparent benefits of using cloudflare as that they provide "the exact numbers on how many people requested or visited my site" https://news.ycombinator.com/item?id=21243135 https://news.ycombinator.com/item?id=21243135
  - Spivak 7y ago
    (Not speaking to CF specifically.) I think we need to nail down what we mean by unique visitors -- we're talking about an estimation which pools from many different sources to calculate as best we can the number of different real human visitors. There is good data in the server logs but if that's your only source of data then your estimation is going to be coarse. This can be good enough for certain applications, and is plenty for sites with simple access models but you do lose information compared to what you get from client-side tracking.
- diafygi 7y ago
  Does anyone have a list of self-hosted server-side analytics tools with conversion funnel tracking? The only self-hosted one I can find that has a funnel feature is Matomo.
- LinuxBender 7y ago
  > Nobody looks at logs anymore? Nope. It blew my mind as well. I used to work for a company that spent about 70 million per quarter on advertising and they most certainly wanted every .. last .. access log .. line. They would scrutinize the way I did log rotation to ensure nothing was lost. At my current gig, the web marketing team didn't even know what access logs were. I don't think they left out a single tracker on the site.
zimbatm 7y ago
Or just actively punishing Firefox because blocking tracking is also blocking ads, which is their revenue model. Why spend money optimizing for a browser that doesn't generate revenue? If Firefox becomes the bastion of privacy sensitive people it will become more and more like Tor users, all tainted with the same labels. I mean it's already the case that recaptcha will more likely trigger on Firefox than Chrome, asking for multiple rounds of checks. Like visitors existing Tor exit nodes, in a bit less worse.
- dmortin 7y ago
  > Why spend money optimizing for a browser that doesn't generate revenue? Maybe sites will move to a model that you'll be offered to pay for their services if you block trackers. So you'll have a choice of paying with your data or paying with your money. Either way someone has to foot the bill at the end. I guess in this case most users will let trackers work instead of paying.
  - re4ctor 7y ago
    That then gets into the ethics of should privacy only be for the wealthy
    - lotsofpulp 7y ago
      And that would get into the ethics of not allowing non wealthy people to pay with their data.
    - mrmonkeyman 7y ago
      Paying 2 dollars is not for the "wealthy". Nice try.
    - oldmanhorton 7y ago
      That's fair, but before computers, not much was available for free. Getting a newspaper would generally cost money, as would using the payphone to communicate, as would so many other things that are now freely available. So, while its definitely a problem, as long as it's cheap enough for "most" people to afford it, then I think we aren't really regressing? Perhaps if we consider information on the internet to be vital to daily life, a program like food stamps could be implemented? It sounds rather over engineered but it's the only idea I've ever heard for working around the "privacy is only for the wealthy" concern.
akerro 7y ago
I used to have piwik instead GA 2-5 years ago, according to piwik my website had 23% more visitors than on GA. https://matomo.org/ https://matomo.org/
- toastal 7y ago
  10 years later and still no PostgreSQL support.
  - dewey 7y ago
    https://github.com/matomo-org/matomo/pulls https://github.com/matomo-org/matomo/pulls
  - akerro 7y ago
    Also no support for ELK and plain-text-folder-managed data structure, so what? do one thing and do it well.
- bobviolier 7y ago
  I think one of the things that GA is doing is actively removing bots visits because they skew country/browser/etc. data. Maybe Matomo is not doing that (so well)? Another thing is that people blocking GA trackers will also not show up in GA, but will (probably) show up Matomo.
- nolok 7y ago
  > according to piwik my website had 23% more visitors than on GA A fair share of that is probably because GA is much better at identifying the same user across multiple browsers / devices. If they identify two browsers as the same person on site A, they can know they're the same person on site B, even if site B on its own doesn't have enough to figure it out (so piwik can't know it).
  - m463 7y ago
    or because GA is a well-known static target for blocking.
thrwn_frthr_awy 7y ago
This sounds like a risk for using Google Analytics to me. If your analytics suite isn’t representative of your site’s traffic and users, maybe you should look for a more ethical platform/solution?
- jakemal 7y ago
  Are there more ethical platforms that don't get blocked by default for Firefox users?
  - chmod775 7y ago
    Self-hosted analytics like Matomo for instance. https://github.com/matomo-org/matomo https://github.com/matomo-org/matomo
    - Spivak 7y ago
      Which do get blocked since they're still trackers. The protection isn't just against 3rd party trackers. And if you obfuscate your Matomo client script to get around the user's preference for trackers you're not being very ethical.
      - floatboth 7y ago
        Seems like Firefox's built-in protection is mostly about 3rd-party trackers? Even the description for "Tracking content" (the not-obviously-3rd-party category) talks about "external ads, videos, and other content with tracking code from other companies. For example, a website may embed a video from a video platform". And this category is not blocked by default, only in strict mode.
      - thrwn_frthr_awy 7y ago
        > And if you obfuscate your Matomo client script to get around the user's preference for trackers you're not being very ethical. I'm not sure I agree, but you made interesting point and I'm glad you brought it up! Personally, I don't think I would classify that as unethical. Ethics need to be looked at with intent in mind, and if you are comfortable being honest about what you are doing, I _think_ I would be okay with this. Not sure, though–I do have doubt.
giancarlostoro 7y ago
That's only effective against client-side analytics. There's also the user-agent from the client sent to the server. This is sent in most (all?) requests.
- Someone1234 7y ago
  Client side analytics is the most prevalent though because it provides far more useful, actionable, information. Everywhere I've worked we have used it in one form or another. It will take some time for people to adapt to server side analytics since a lot of places won't have that solution in place. Or more specifically a solution in place that overlaps with what client side analytics is meant to be providing already (they may have other server side analytics, like performance and hits).
  - giancarlostoro 7y ago
    I get that, but in regards to getting raw browser marketshare, client-side code isn't the only way.
dblohm7 7y ago
That’s already been happening for quite some time now.

bn7t 7y ago

Website owners can just use the user agent in weblogs to determine this.

dajonker 7y ago
Yes they can. But in my experience, at big organizations, the people using Google Analytics are not developers and will have no idea what weblogs are or how they could use them to determine the real browser statistics.
- NotASithLord 7y ago
  I'm a dev and I have no idea what weblogs have to do with this discussion. Quick search didn't do anything to clue me in either.
  - 0xffff2 7y ago
    Which part is confusing? The browser sends lots of data to the server with each request. The server usually logs at least some of that data. Among other things, that data can be used to generate a subset of the data that is typically gathered by GA and other client-side analytics tools.

zaarn 7y ago

Nobody looks at logs anymore? Strange, those are very useful. Google Analytics isn't the only analytics tool and as blockers have become more prevalent relying on GA as your sole source of truth is becoming increasingly desperate to worship Google. Ideally you run a simple analytics tool alongside that is lightweight and doesn't get blocked (ie, privacy friendly) as well as running a log analyzer on your apache/webserver logs.

geraltofrivia 7y ago
I host my personal site (https://priyansh.page https://priyansh.page) on Netlify and they give me analytics computed from server side logs. [Link to Netlify Analytics](https://www.netlify.com/products/analytics/ https://www.netlify.com/products/analytics/). They don't provide a lot of demographic information (even browser used, afaik) but provide sufficient information for my purpose. edit: I should mention that this a paid feature ($9 a month). edit2: I do notice a higher number of traffic when compared to some free analytics which work on the client side, and respect Do Not Track.
oarsinsync 7y ago
> Nobody looks at logs anymore? Apparently not, as one other comment on this site lists one of the apparent benefits of using cloudflare as that they provide "the exact numbers on how many people requested or visited my site" https://news.ycombinator.com/item?id=21243135 https://news.ycombinator.com/item?id=21243135
- Spivak 7y ago
  (Not speaking to CF specifically.) I think we need to nail down what we mean by unique visitors -- we're talking about an estimation which pools from many different sources to calculate as best we can the number of different real human visitors. There is good data in the server logs but if that's your only source of data then your estimation is going to be coarse. This can be good enough for certain applications, and is plenty for sites with simple access models but you do lose information compared to what you get from client-side tracking.
diafygi 7y ago
Does anyone have a list of self-hosted server-side analytics tools with conversion funnel tracking? The only self-hosted one I can find that has a funnel feature is Matomo.
LinuxBender 7y ago
> Nobody looks at logs anymore? Nope. It blew my mind as well. I used to work for a company that spent about 70 million per quarter on advertising and they most certainly wanted every .. last .. access log .. line. They would scrutinize the way I did log rotation to ensure nothing was lost. At my current gig, the web marketing team didn't even know what access logs were. I don't think they left out a single tracker on the site.

zimbatm 7y ago

Or just actively punishing Firefox because blocking tracking is also blocking ads, which is their revenue model. Why spend money optimizing for a browser that doesn't generate revenue? If Firefox becomes the bastion of privacy sensitive people it will become more and more like Tor users, all tainted with the same labels. I mean it's already the case that recaptcha will more likely trigger on Firefox than Chrome, asking for multiple rounds of checks. Like visitors existing Tor exit nodes, in a bit less worse.

dmortin 7y ago
> Why spend money optimizing for a browser that doesn't generate revenue? Maybe sites will move to a model that you'll be offered to pay for their services if you block trackers. So you'll have a choice of paying with your data or paying with your money. Either way someone has to foot the bill at the end. I guess in this case most users will let trackers work instead of paying.
- re4ctor 7y ago
  That then gets into the ethics of should privacy only be for the wealthy
  - lotsofpulp 7y ago
    And that would get into the ethics of not allowing non wealthy people to pay with their data.
  - mrmonkeyman 7y ago
    Paying 2 dollars is not for the "wealthy". Nice try.
  - oldmanhorton 7y ago
    That's fair, but before computers, not much was available for free. Getting a newspaper would generally cost money, as would using the payphone to communicate, as would so many other things that are now freely available. So, while its definitely a problem, as long as it's cheap enough for "most" people to afford it, then I think we aren't really regressing? Perhaps if we consider information on the internet to be vital to daily life, a program like food stamps could be implemented? It sounds rather over engineered but it's the only idea I've ever heard for working around the "privacy is only for the wealthy" concern.

akerro 7y ago

I used to have piwik instead GA 2-5 years ago, according to piwik my website had 23% more visitors than on GA. https://matomo.org/ https://matomo.org/

toastal 7y ago
10 years later and still no PostgreSQL support.
- dewey 7y ago
  https://github.com/matomo-org/matomo/pulls https://github.com/matomo-org/matomo/pulls
- akerro 7y ago
  Also no support for ELK and plain-text-folder-managed data structure, so what? do one thing and do it well.
bobviolier 7y ago
I think one of the things that GA is doing is actively removing bots visits because they skew country/browser/etc. data. Maybe Matomo is not doing that (so well)? Another thing is that people blocking GA trackers will also not show up in GA, but will (probably) show up Matomo.
nolok 7y ago
> according to piwik my website had 23% more visitors than on GA A fair share of that is probably because GA is much better at identifying the same user across multiple browsers / devices. If they identify two browsers as the same person on site A, they can know they're the same person on site B, even if site B on its own doesn't have enough to figure it out (so piwik can't know it).
- m463 7y ago
  or because GA is a well-known static target for blocking.

thrwn_frthr_awy 7y ago

This sounds like a risk for using Google Analytics to me. If your analytics suite isn’t representative of your site’s traffic and users, maybe you should look for a more ethical platform/solution?

jakemal 7y ago
Are there more ethical platforms that don't get blocked by default for Firefox users?
- chmod775 7y ago
  Self-hosted analytics like Matomo for instance. https://github.com/matomo-org/matomo https://github.com/matomo-org/matomo
  - Spivak 7y ago
    Which do get blocked since they're still trackers. The protection isn't just against 3rd party trackers. And if you obfuscate your Matomo client script to get around the user's preference for trackers you're not being very ethical.
    - floatboth 7y ago
      Seems like Firefox's built-in protection is mostly about 3rd-party trackers? Even the description for "Tracking content" (the not-obviously-3rd-party category) talks about "external ads, videos, and other content with tracking code from other companies. For example, a website may embed a video from a video platform". And this category is not blocked by default, only in strict mode.
    - thrwn_frthr_awy 7y ago
      > And if you obfuscate your Matomo client script to get around the user's preference for trackers you're not being very ethical. I'm not sure I agree, but you made interesting point and I'm glad you brought it up! Personally, I don't think I would classify that as unethical. Ethics need to be looked at with intent in mind, and if you are comfortable being honest about what you are doing, I _think_ I would be okay with this. Not sure, though–I do have doubt.

giancarlostoro 7y ago

That's only effective against client-side analytics. There's also the user-agent from the client sent to the server. This is sent in most (all?) requests.

Someone1234 7y ago
Client side analytics is the most prevalent though because it provides far more useful, actionable, information. Everywhere I've worked we have used it in one form or another. It will take some time for people to adapt to server side analytics since a lot of places won't have that solution in place. Or more specifically a solution in place that overlaps with what client side analytics is meant to be providing already (they may have other server side analytics, like performance and hits).
- giancarlostoro 7y ago
  I get that, but in regards to getting raw browser marketshare, client-side code isn't the only way.

dblohm7 7y ago

That’s already been happening for quite some time now.