I usually use two squares of toilet paper per wipe. Occasionally I’ll use three, sometimes accidentally and sometimes because I need three. Whenever I use three I wonder to myself how that will appear when the numbers are crunched. Then I remember there’s no analytics for my toilet paper.
> otherwise they don’t know how to improve the product.
Or, you know, they could just ask customers how the product is working for them. Automated data collection is a huge, huge problem and stampedes over civil liberties. We ought to control our computers, and not be controlled by them, which is why I vigorously oppose all non-free (proprietary, non libre) software.
I send error data from all of my web and mobile apps. We catch a LOT of bugs that we don't see in test due to browser differences or state differences. While this may be invasive, it's necessary to deliver a good customer experience.
I believe a gradient exists here. That's why we don't collect any usage information, but error reporting is something I've found life changing especially given how fragmented the experience is across different browsers and devices.
I see your point here, and then I count the number of times I have responded to survey requests or participated in nielsen ratings surveys, even when paid to do it.
I agree, most companies conduct feedback in a completely impersonal way. I tell my customers that my software doesn't contain any tracking or telemetry, and that human-to-human feedback (direct email) from users is the only way I can know of problems. I also make it very clear that I am open to all kinds of feedback. This is great because I get bug reports directly, can understand how people are using the program in their workflow, and can contextualize feature requests and build solutions that fit for everyone.
Users like this because they are actually being listened to by the software's author, there is no overhead to fixing problems, and they aren't being shoehorned into a database.
Of course direct feedback is good. But it's not scalable and you risk of running a very real case of bias towards whoever you're asking. Running telemetry on your software is not "evil" or "stampedes over civil liberties". You can most certainly run telemetry in a way that protects people's privacy and gives you insightful data on how your users operate and interact with your software. I know this, I've worked on an analytics service serving hundreds of millions of users. I'm building right now a feature-flagging/configuration/core metrics service that does exactly that: preserve user privacy and allow developers to learn about their users (not easy and there are some trade offs, like accuracy).
I can't count the number of times I have seen a new product or bit of software and though immediately "That sounds like a horrible idea, how useless" And then later eventually trying it and liking it.
Asking customers what they want can be useful sometimes but mostly people don't know what they want other than a slightly improved version of the thing they already like.
> Majority of consumer facing products collect usage analytics, otherwise they don’t know how to improve the product.
Producers managed to improve products for thousands of years without violating the privacy of their customer. Equating slightly inconvenient with impossible doesn't make for a great argument.
The improvements were stumbles in the dark by comparison — people didn’t even realise colour blindness existed until the industrial revolution.
That said, while I am fine with automated bug tracking for this sort of goal, I absolutely agree that all automated tracking needs to be opt-in rather than opt-out.
Yeah, while I'm not a fan of involuntary data collection it completely makes sense why firms do it. Unskewing data collected in voluntary and self reported scenarios is a treacherous and sometimes impossible undertaking.
Edit: Often times we're not talking about the difference between good data and really good data. We're talking about the difference between useful and useless data.
I don't agree it's a 0 or 1 decision, you can gather analytics in a privacy-preserving way, see e.g. plausible.io [1] or similar implementations that track users without collecting personal data.
I suppose your mattress phones home to tell its makers how often you sleep and with which frequency you do other vigorous physical activities on it. Only way to improve it.
I think that a mattress is a bit simpler than an application with hundreds of thousands of possible interactions. There are not so many different ways you can use a mattress.
There is a limit to what majority can decide for a person and it's called "rights of a person" designed exactly for that purpose.
So, for instance, majority in Nazi-Germany was ok with humiliating and eventually killing Jewish people and that fact doesn't make it justifed anyhow. Majority sometimes disrespect freedom of speech using bullying innocent comments advocating respect for a freedom which we observe recently even here in form of downvoting and it also have no justification what so ever.
>otherwise they don’t know how to improve the product.
they should ask and pay for participation if they wish to know, or they simply don't know, it's ok.
If I wasn't asked about it I consider it a spyware activity and we have no idea what they do with this information now or in the future.
