One of the corollaries to this is that if you are basing your decisions on analytics, you are probably under-representing people who block them, who often tend to be a non-representative subset of your users. I have seen multiple projects do things like stop supporting features that their analytics “showed nobody was using” but failed to realize that a significant portion of their (often technical and very vocal) userbase did in fact use the feature.
Oh, and by the way, responding to these kinds of issues with “if you wanted your voice to be heard, you should have turned on analytics” is inexcusable.
War story: we turned off the “print” button on MSN because it was unused. Users in Japan complained - they were printing articles to read on the train. Turned it back on in that geo
When I turned off my Windows 10 Telemetry, I expressed a clear choice that I'd rather not give my data, and it's fine for my voice to not be heard. I just really don't care about Windows 10 that much. On the other hand, I turned on all the telemetry for OctoPi, Steam, and a bunch of other softwares I care about. I want my voice to be heard for those cases.
Serial numbers filed off: an author announced on his newsletter that he was going to turn on the "feature" of his mail service that would automatically unsubscribe anyone who didn't read the newsletter.
It tracked this, of course, by assuming that the mail would be opened in a web browser that would make requests for images.
He cancelled that decision:
"Logged opens for each newsletter are between 53% and 60% -- but an experiment a while back revealed that hundreds of you aren't logged, for various reasons around email security and that one guy who reads everything on a 1980's greenscreen monitor. Clicking of links depends massively on what's in the newsletter, plus the note in the previous sentence, but has gone as high as 41% of readers on one week this year."
"Oh, and by the way, responding to these kinds of issues with “if you wanted your voice to be heard, you should have turned on analytics” is inexcusable."
Authentic question: how would a business know this otherwise in an actionable and effective way?
> Oh, and by the way, responding to these kinds of issues with “if you wanted your voice to be heard, you should have turned on analytics” is inexcusable.
Some folks seem to think analytics is the only way to get feedback from users.
Just ask me. Survey, email me directly, anything but siphoning all my data off to Google in the process.
I don't understand why first-party analytics and JavaScript aren't more popular, but I suppose it's a matter of cost in the serverless and freemium world. But for VC stuff... There are plenty of pre-built analytics kits out there that can run from your site and you would have none of these problems.
Depends on the service, but if they are ad supported, people who block analytics probably also block ads, so they don't really have any incentive to cater to them.
Analytics should be one of your sources. Server logs is another. I know of a company that has a whatsapp group for people who've contacted them via support, and one more for "fans". Each of these bring novel perspectives, and its up to the org to extract signal out of them.
I run a non-tech eCommerce site in North America that does $6-8m a year and I tested two 90 day periods year over year and found 12.4% of transactions didn’t show up in Google Analytics. I haven’t segmented them by geo yet but about 75% of my sales are in North America. There was only a 0.1% difference year over year.
1. Number of visitors as recorded in Google Analytics
2. Number of loads of a 1x1 pixel served on a different domain
They see higher numbers for (2) than (1), and attribute the difference to users blocking Google Analytics.
I don't see them describing how they excluded bot traffic, however, and for my sites the majority of hits I get are from bots. Only some bots run JS, so I suspect their numbers for blocking users are thoroughly diluted by these bots.
(Disclosure: I work for Google, speaking only for myself)
It's not just two numbers of total hits the article is comparing.
The author extracted the browser information from the server logs (presumably from the User-Agent header i guess?). If they were able to do this, i'd assume they also filtered out bots from the tally :)
Yeah, the numbers in the article are so far off my intuition that I'm happy to latch onto any explanation for why they're weird. Being unable to effectively discount bots seems likely.
I run a geek-tech blog in Hebrew. I used analytics for a while in the past, but the numbers were completely off. There were posts with almost more comments than visitors based on the analytics.
Took me a while to realize that most of my readers block analytics since they're super privacy-saavy. I shut the tool down, it's useless for some crowds.
Nowadays I also doubt if it's ethical for any crowd.
What's interesting to me about the number of people who block Google analytics, is the number of people working with those analytics in product management, marketing or SEO, that are apparently unaware of anyone being able to block Google Analytics.
This is why I have analytics on the JavaScript side and server side. I can calculate those who block my analytics and at the same time capture a lot of the relevant information
We do ~1.1m sessions per month according to GA, comparing to our Clofudflare visitor data real number is ~35% higher. I think CF is going to be as accurate as possible and both say the filter bots out.
I recently got tired of Google Analytics (for a number of reasons) and switched to Fathom Analytics.
Now, Ghostery says I have "zero trackers" on my personal site.
Fathom doesn't collect personal info about my visitors. They just show me my aggregate metrics (popular pages, top referrers) which is all I need anyway.
Unfortunately it's just a matter of time. I have no doubt they will get added to uBlock pretty soon. Most people don't make a difference between tracking anonymously for data aggregation and tracking individual users. Many are paranoid and think that any type of tracking is bad.
Try Firefox mobile with ublock origin. I wish more Android users knew about this.
Sidenote: ads on sites degrade the experience so badly, whenever I browse without adblock I am honestly shocked at how bad it is, and I am amazed others don't seem to care.
Root users like myself use tools like AdAway to managing hosts blocking. This will help cover a lot of ads in apps as well. You also have alternative WebView implementations like Bromite (although I wish it were easier to install and worked on Android 10) that can give you ad blocking in apps using web views too.
This is also missing the part where you can run Android without the Play Store using microG and F-Droid + Aurora. It's not a huge number, but there are still a lot of people blocking ads in one form or another--even if it's as simple as using Firefox or Brave for basic browsing.
On the Samsung Galaxy you just install Samsung Internet and from within it you can add AdBlock Plus (and most of the other adblockers). They are delivered through Samsung's apk store.
It saved the mobile surfing experience for me (was using chrome on the mobile before)..
DNS66 from F-Droid works well. It routes all web traffic through a tunnel applying hosts file(s) to block domains associated with ads and other dodgy cruft.
Turned it off via PiHole, so my ChromeOS laptops and Androids are covered at home, still trying to get Wireguard set up on the PiHole to allow remote VPN, while suffering double-NAT/CGNat through my AT&T hotpsot as home connection. Turns out it's not a slam-dunk!
FWIW the numbers for my blog are broadly similar, with a larger sample taken when I had a blog post on HN, with about 2x higher unique user numbers from Cloudflare analytics (server side) than Google Analytics.
Is there a free server-side analytics service that’s not a pain to set up?
I’m using Netlify, but enabling analytics is 9$/site/month, so I tell myself that GA is just fine.
Not only advanced users. At least as of 2019 we for example have pihole installed LAN wide at home and Lockdown on all portable devices. Also recommend or installed on devices of anyone I know, technical or not and of course my number one recommendation to anyone is to install Firefox and avoid chrome.
This. Does it take into account scraping bots? I remember looking at my server logs many years ago, finding a surprising amount of Chinese and Russian visits for my personal website. Eventually I realized it was Baidu Yandex type stuff
I don't think there is an effective way to distinguish bots from real users when a UA is faking what it is. IP filtering or other measures can help, but you ultimately rely on the UA strong from the UA.
I refuse to implement client-side analytics; for one thing, it is a waste of bandwidth, and for the other thing, it is not how software should be designed, and for the other other thing, it is unethical.
I do log some stuff sent to the server, as well as some stuff about the response (such as the HTTP status code, data size, timing, etc), although I do not sell this data to anyone else, and this logged data is reduced further if the client sends a "DNT:1" header.
But decisions about how to make something would normally be based on actual comments by the users, rather than analytics, I should think.
How many users who block GA are also going to not be honest about the OS or browser they claim to be? I suspect that number is going to be a much higher percentage than of those who don't block GA...
I'd imagine most users who block GA/ads on desktop would also want to on mobile, but can't just because it's so difficult to set up an adblocker on mobile.
A HOSTS file or blocking DNS server will easily do that, my whole network in fact has GA and a bunch of other crap blocked this way. On the other hand, setting up a MITM proxy/VPN is much much harder on mobile. However I am surprised at the 0% for Android and 17% for iOS blocking GA --- I was expecting it to be the opposite, with the former being historically much less of a walled-garden than the latter.
On the other hand, perhaps everyone who blocks GA and uses Android is in the aforementioned situation of not saying that they're using Android --- they may be reporting a Linux or some other user-agent.
I find it quite surprising how many companies don't invest enough in having a certain maturity in terms of web analytics, despite they heavily invested in CRM, machine learning and other fields.
Ad-blockers are still not considered a fact in data gathering nor in QA or, in the best case, just a minor singularity.
Despite different sources agree on a +30% of usage, at least in Europe, they're still ignored.
Still many business decisions are based on solutions like GA.
If decisions were to be taken on any other source, let's say polls, marketing studies, etc., after being warned of a 30% of uncertainty, people would be very reluctant to choose A over B. This is not the case for web analytics, it seems.
Now technology has evolved and any medium size company should harvest their own data (yes, stop giving away some of your most valuable data to third parties, please!).
would it be naive to say that sites can get around the blockers by proxying their analytics through their own domain? Isn't the lazy way of sending everything directly to ga from the client the cause here -- and any thoughtful site owner should be able to circumvent this if they truly care to?
I have started some weeks ago this project: https://github.com/iris-analytics
It's a small JS that gathers data and sends it to a Go backend to then be stored in ClickHouse.
Although there's lots to do, we use it in production successfully.
Remember ClickHouse was born precisely for web analytics where a single instance can handle hundreds of millions of inserts per day with no effort.
I did this because stats say adblocker penetration in Europe is beyond 30% and this would give us real time insights with no sampling and ad-hoc queries.
If you want to help me out, you are very welcome!!!
[+] [-] saagarjha|6 years ago|reply
Oh, and by the way, responding to these kinds of issues with “if you wanted your voice to be heard, you should have turned on analytics” is inexcusable.
[+] [-] RyJones|6 years ago|reply
[+] [-] thanhhaimai|6 years ago|reply
Just curious, what's so inexcusable about that?
[+] [-] dsr_|6 years ago|reply
It tracked this, of course, by assuming that the mail would be opened in a web browser that would make requests for images.
He cancelled that decision:
"Logged opens for each newsletter are between 53% and 60% -- but an experiment a while back revealed that hundreds of you aren't logged, for various reasons around email security and that one guy who reads everything on a 1980's greenscreen monitor. Clicking of links depends massively on what's in the newsletter, plus the note in the previous sentence, but has gone as high as 41% of readers on one week this year."
[+] [-] johnnyg|6 years ago|reply
Authentic question: how would a business know this otherwise in an actionable and effective way?
[+] [-] nkrisc|6 years ago|reply
Some folks seem to think analytics is the only way to get feedback from users.
Just ask me. Survey, email me directly, anything but siphoning all my data off to Google in the process.
[+] [-] cl0rkster|6 years ago|reply
[+] [-] Swizec|6 years ago|reply
You can't just follow analytics. You have to understand your users.
[+] [-] cortesoft|6 years ago|reply
[+] [-] SquareWheel|6 years ago|reply
It's true though. If you want to be recognized, you can't be incognito. That's like refusing to vote and then complaining about politics.
[+] [-] kludgekraft|6 years ago|reply
[+] [-] kevin_thibedeau|6 years ago|reply
[+] [-] furtive808|6 years ago|reply
[+] [-] jefftk|6 years ago|reply
1. Number of visitors as recorded in Google Analytics
2. Number of loads of a 1x1 pixel served on a different domain
They see higher numbers for (2) than (1), and attribute the difference to users blocking Google Analytics.
I don't see them describing how they excluded bot traffic, however, and for my sites the majority of hits I get are from bots. Only some bots run JS, so I suspect their numbers for blocking users are thoroughly diluted by these bots.
(Disclosure: I work for Google, speaking only for myself)
[+] [-] ianhawes|6 years ago|reply
[+] [-] epidemian|6 years ago|reply
The author extracted the browser information from the server logs (presumably from the User-Agent header i guess?). If they were able to do this, i'd assume they also filtered out bots from the tally :)
[+] [-] ben0x539|6 years ago|reply
[+] [-] jjohansson|6 years ago|reply
[+] [-] anjakefala|6 years ago|reply
[+] [-] neiman|6 years ago|reply
Took me a while to realize that most of my readers block analytics since they're super privacy-saavy. I shut the tool down, it's useless for some crowds.
Nowadays I also doubt if it's ethical for any crowd.
[+] [-] technion|6 years ago|reply
[+] [-] lettergram|6 years ago|reply
[+] [-] saagarjha|6 years ago|reply
[+] [-] TomGullen|6 years ago|reply
[+] [-] mijustin|6 years ago|reply
Now, Ghostery says I have "zero trackers" on my personal site.
Fathom doesn't collect personal info about my visitors. They just show me my aggregate metrics (popular pages, top referrers) which is all I need anyway.
[+] [-] going_to_800|6 years ago|reply
[+] [-] mariomeissner|6 years ago|reply
[+] [-] gradschool|6 years ago|reply
Don't forget about all those smartasses spoofing their user agent just to make your life even harder.
https://addons.mozilla.org/en-US/firefox/addon/random_user_a...
[+] [-] lousken|6 years ago|reply
[+] [-] twhb|6 years ago|reply
[+] [-] zizee|6 years ago|reply
Sidenote: ads on sites degrade the experience so badly, whenever I browse without adblock I am honestly shocked at how bad it is, and I am amazed others don't seem to care.
[+] [-] ben0x539|6 years ago|reply
[+] [-] toastal|6 years ago|reply
This is also missing the part where you can run Android without the Play Store using microG and F-Droid + Aurora. It's not a huge number, but there are still a lot of people blocking ads in one form or another--even if it's as simple as using Firefox or Brave for basic browsing.
[+] [-] l33tman|6 years ago|reply
It saved the mobile surfing experience for me (was using chrome on the mobile before)..
[+] [-] system2|6 years ago|reply
https://addons.mozilla.org/en-US/android/addon/adblock-plus/
[+] [-] xorzarle|6 years ago|reply
[+] [-] wideasleep1|6 years ago|reply
[+] [-] trishume|6 years ago|reply
[+] [-] ronyfadel|6 years ago|reply
[+] [-] onyva|6 years ago|reply
[+] [-] Narkov|6 years ago|reply
[+] [-] cerved|6 years ago|reply
[+] [-] Ayesh|6 years ago|reply
[+] [-] zzo38computer|6 years ago|reply
I do log some stuff sent to the server, as well as some stuff about the response (such as the HTTP status code, data size, timing, etc), although I do not sell this data to anyone else, and this logged data is reduced further if the client sends a "DNT:1" header.
But decisions about how to make something would normally be based on actual comments by the users, rather than analytics, I should think.
[+] [-] userbinator|6 years ago|reply
I'd imagine most users who block GA/ads on desktop would also want to on mobile, but can't just because it's so difficult to set up an adblocker on mobile.
A HOSTS file or blocking DNS server will easily do that, my whole network in fact has GA and a bunch of other crap blocked this way. On the other hand, setting up a MITM proxy/VPN is much much harder on mobile. However I am surprised at the 0% for Android and 17% for iOS blocking GA --- I was expecting it to be the opposite, with the former being historically much less of a walled-garden than the latter.
On the other hand, perhaps everyone who blocks GA and uses Android is in the aforementioned situation of not saying that they're using Android --- they may be reporting a Linux or some other user-agent.
[+] [-] pachico|6 years ago|reply
[+] [-] darepublic|6 years ago|reply
[+] [-] IggleSniggle|6 years ago|reply
[+] [-] pachico|6 years ago|reply
If you want to help me out, you are very welcome!!!
[+] [-] einpoklum|6 years ago|reply