5 ms·

This is good to hear. I didn't think the percentages would be that high. "Data driven design" is kinda overrated anyway. It's better to actually listen to your

This is good to hear. I didn't think the percentages would be that high. "Data driven design" is kinda overrated anyway. It's better to actually listen to your users. Stats don't show what people think, and they're often tweaked to show what people want to see.

TeMPOraL 5y ago
Collecting telemetry is essentially doing an impromptu social science study, and it suffers from the same problems - collecting good data and making correct inferences is a difficult task, and it requires specialized skills. Social scientists are trained to do this, and yet they still fail more often than not[0]. A random business trying to run product development off automated metrics? They don't stand a chance. This is not to say the data is useless - it's just worth remembering that, unless you have strong statistical background, you're probably reading your data wrong, so your metrics should be treated as weak evidence at best. -- [0] - It's not snark, it's a corollary of the replication crisis. This is a hard job.
WJW 5y ago
"Listening to your users" is just a more manual form of data gathering though. Also, it's not like "hearing what you want to hear" has been invented by statisticians, that is super easy to do when talking to users as well. Some of the problems in talking to users: - people will often tell you what they think you want to hear rather than what they really think. (Very few users would ever tell an interviewing dev that they think the product is unsalvageable, the dev team wasted several months and that the user would never ever use it) - Users often don't know what they want because they don't understand the full range of possibilities. Sometimes they dream of tech that would violate several laws of physics, at other times they fail to realize that for-loops exist and just ask for a faster way to manually click a button a thousand times. - Fads affect users just as much as anyone else. If there has been a surge of news articles about how Google is doing fancy AI stuff, you can bet the bakery on the corner of the street will mention machine learning when asked what they want to improve their business.
- JohnFen 5y ago
  "Listening to your users" is a voluntary exchange with your users, not compelling data from them. The difference between the two is critical.
  - snowwrestler 5y ago
    It’s amazing how far the goalposts have moved on analytical data. I worked retail in The Time Before The Web and we collected and analyzed data from our customers. We had a door clicker to give us a daily record of traffic. We did detailed analytics on what items customers bought when, to inform inventory and staffing decisions. We did regular people-counting studies to evaluate our merchandising in different parts of the store. We did coupons in local papers for which we could directly track redemption. This was all in the service of making the store work better for our customers. Now, you might say that those are all anonymous aggregate piles of data. Well, so are most web analytics packages, including many client-side JavaScript.
    - JohnFen 5y ago
      No goalposts have moved. The difference between the two is that there's no question that counting heads walking through a door and watching inventory patterns are not personal to the customer. With telemetry, including web analytics, there is always a question because there is no visibility as to their actual practices. It all amounts to "trust us". And we've had more than enough experience with these things to know that such trust is frequently misplaced -- so none of it can be trusted. And GA especially cannot be trusted, and also happens to be pretty much ubiquitous. As a side note, it's become just as dangerous to be in a physical store as online now. People who prefer not to be spied on must pay in cash, be sure to put their phones into airplane mode before entering the store, and so forth. And with the increasing adoption of face recognition in stores, it becomes risky to even show your face. In other words, they have to be as "on guard" in the store as online.
hkai 5y ago
Why is it good to prevent analytics? How can it benefit the user? Why is it more important to hide the fact that you're interested in cars and Netflix, but not important to protect free speech, free enterprise and free association? I'm absolutely shocked by why someone would want to protect against a vendor selling me a new mobile phone when the real threat of social media is promoting left-wing extremism and censorship.
- scrollaway 5y ago
  I'll take your question at face value (even though you're very certainly a troll account given your post history). It's less about "hiding" or "preventing" anything, and more about putting the user in control of sharing those things. If I tell you I like figure skating, here on an online forum, I'm making a conscious decision to share that fact about my life. If a script completely silently figures out I like figure skating, because of websites I previously visited, that it was able to infer due to various data sharing setups (such as google analytics overprevalence etc), that is not a choice I made. It's a form of stalking. The problem is accentuated because it's not just one hobby, it's everything. Your hobbies, your location, gender, religion, income, political affiliation, etc. That it serves ads based on that is icky, but the problem is that it knows those things at all without me choosing to share that. And this is without even getting into "bad actor" territory. Privacy is control over what you share with others.
- perryizgr8 5y ago
  I think both are important. Personally I hate analytics because it makes my equipment work against me, or at least for someone else. Consider that there are websites and apps that are so overloaded with js and ads, that they will run sluggishly on anything but the latest phones and laptops. I would rather not allow that. I have no problem with static ads. I love watching creative ads on TV during a sports broadcast. Instagram ads are another example of advertising done well. Minimal distraction from the content, no slowdowns, easily dismissible. As for censorship, I agree that is a more important issue. But we can focus on more than one issue at a time. I will take a win on either front.
- approxim8ion 5y ago
ehsankia 5y ago
The percentage is misleading. it's not people intentionally blocking Google Analytics, it's anyone that runs any adblocker, which almost all block Google Analytics. I installed uBlock for my parents and they don't have the faintest clue what GA is. A significant portion of those 58% probably don't care either.

TeMPOraL 5y ago

Collecting telemetry is essentially doing an impromptu social science study, and it suffers from the same problems - collecting good data and making correct inferences is a difficult task, and it requires specialized skills. Social scientists are trained to do this, and yet they still fail more often than not[0]. A random business trying to run product development off automated metrics? They don't stand a chance. This is not to say the data is useless - it's just worth remembering that, unless you have strong statistical background, you're probably reading your data wrong, so your metrics should be treated as weak evidence at best. -- [0] - It's not snark, it's a corollary of the replication crisis. This is a hard job.

WJW 5y ago

"Listening to your users" is just a more manual form of data gathering though. Also, it's not like "hearing what you want to hear" has been invented by statisticians, that is super easy to do when talking to users as well. Some of the problems in talking to users: - people will often tell you what they think you want to hear rather than what they really think. (Very few users would ever tell an interviewing dev that they think the product is unsalvageable, the dev team wasted several months and that the user would never ever use it) - Users often don't know what they want because they don't understand the full range of possibilities. Sometimes they dream of tech that would violate several laws of physics, at other times they fail to realize that for-loops exist and just ask for a faster way to manually click a button a thousand times. - Fads affect users just as much as anyone else. If there has been a surge of news articles about how Google is doing fancy AI stuff, you can bet the bakery on the corner of the street will mention machine learning when asked what they want to improve their business.

JohnFen 5y ago
"Listening to your users" is a voluntary exchange with your users, not compelling data from them. The difference between the two is critical.
- snowwrestler 5y ago
  It’s amazing how far the goalposts have moved on analytical data. I worked retail in The Time Before The Web and we collected and analyzed data from our customers. We had a door clicker to give us a daily record of traffic. We did detailed analytics on what items customers bought when, to inform inventory and staffing decisions. We did regular people-counting studies to evaluate our merchandising in different parts of the store. We did coupons in local papers for which we could directly track redemption. This was all in the service of making the store work better for our customers. Now, you might say that those are all anonymous aggregate piles of data. Well, so are most web analytics packages, including many client-side JavaScript.
  - JohnFen 5y ago
    No goalposts have moved. The difference between the two is that there's no question that counting heads walking through a door and watching inventory patterns are not personal to the customer. With telemetry, including web analytics, there is always a question because there is no visibility as to their actual practices. It all amounts to "trust us". And we've had more than enough experience with these things to know that such trust is frequently misplaced -- so none of it can be trusted. And GA especially cannot be trusted, and also happens to be pretty much ubiquitous. As a side note, it's become just as dangerous to be in a physical store as online now. People who prefer not to be spied on must pay in cash, be sure to put their phones into airplane mode before entering the store, and so forth. And with the increasing adoption of face recognition in stores, it becomes risky to even show your face. In other words, they have to be as "on guard" in the store as online.

hkai 5y ago

Why is it good to prevent analytics? How can it benefit the user? Why is it more important to hide the fact that you're interested in cars and Netflix, but not important to protect free speech, free enterprise and free association? I'm absolutely shocked by why someone would want to protect against a vendor selling me a new mobile phone when the real threat of social media is promoting left-wing extremism and censorship.

scrollaway 5y ago
I'll take your question at face value (even though you're very certainly a troll account given your post history). It's less about "hiding" or "preventing" anything, and more about putting the user in control of sharing those things. If I tell you I like figure skating, here on an online forum, I'm making a conscious decision to share that fact about my life. If a script completely silently figures out I like figure skating, because of websites I previously visited, that it was able to infer due to various data sharing setups (such as google analytics overprevalence etc), that is not a choice I made. It's a form of stalking. The problem is accentuated because it's not just one hobby, it's everything. Your hobbies, your location, gender, religion, income, political affiliation, etc. That it serves ads based on that is icky, but the problem is that it knows those things at all without me choosing to share that. And this is without even getting into "bad actor" territory. Privacy is control over what you share with others.
perryizgr8 5y ago
I think both are important. Personally I hate analytics because it makes my equipment work against me, or at least for someone else. Consider that there are websites and apps that are so overloaded with js and ads, that they will run sluggishly on anything but the latest phones and laptops. I would rather not allow that. I have no problem with static ads. I love watching creative ads on TV during a sports broadcast. Instagram ads are another example of advertising done well. Minimal distraction from the content, no slowdowns, easily dismissible. As for censorship, I agree that is a more important issue. But we can focus on more than one issue at a time. I will take a win on either front.
approxim8ion 5y ago

ehsankia 5y ago

The percentage is misleading. it's not people intentionally blocking Google Analytics, it's anyone that runs any adblocker, which almost all block Google Analytics. I installed uBlock for my parents and they don't have the faintest clue what GA is. A significant portion of those 58% probably don't care either.