Are analytics good? | Hacker News Reader

Are analytics good?

todepond.com

87 pointssurprisetalk42 comments

There is a drive towards “no analytics”. For example, this wikiblog has no analytics.

view-source:https://www.todepond.com/wikiblogarden/social-media/analytic...

  <script data-goatcounter="https://todepond.goatcounter.com/count" async src="//gc.zgo.at/count.js"></script>

> GoatCounter is an open source web analytics platform

¯\_(ツ)_/¯

Hey, author here.

When I wrote that post, the site had no analytics. This article prompted someone to reach out and recommend goatcounter to me. It's lovely! And it avoids the bad points that I mention in the post. I highly recommend writing blog posts as a way of finding solutions!

By the way, you can see all the analytics at https://todepond.goatcounter.com

All best,

Lu x

For the simple reports it's generating I'm surprised this deployment of GoatCounter requires javascript in the client instead of just consuming the Location/ip/user-agent string from the server logs.

Feels gdpr violating… Not a functionality and tracking.

it's explicitly in compliance with GDPR, check it out: https://www.goatcounter.com/help/gdpr

> Other information such as the URL or Referer do not relate to an identified person.

This doesn't sound like it holds weight: Referer is a user-specified string, this could easily be attributed to a natural person by the use of additional information

The URL of this comment is [0]. I clicked the link above, if Referer is being tracked, it's almost guaranteed that you've collected my PII

[0] https://news.ycombinator.com/reply?id=38385378&goto=item%3Fi...

I'm not a lawyer but I'm fairly sure that writing "I didn't violate any law" doesn't necessarily mean that the statement is true.

A JS tag is easier to deploy than an agent on a server. I run GoatCounter on my Netlify static site (and I don't even have access to server logs).

Interesting blog post. I personally believe analytics are bad. Optimizing for analytics means:

(a) Taking the path of least resistance instead of trying to gauge engagement in other ways. Yes, you can still use other ways but most people don't.

(b) Optimizing for analytics means optimizing for an immedaite return rather than following your gut on what you like. The ultimate reason for analytics is so that you optimize for ad views for Google, rather than increasing quality. Google would never do anything unless it is in its own self-interest. They have zero social responsibility.

(c) I have disabled analytics on my blogs and now I find I write just what I want and what I find interesting and I don't bother with which posts get the most views.

What do you want? Views from interesting poeple or an interesting number of views?

> What do you want?

I want to know which pages and posts people find the most interesting and which remain popular over time. Just because I’m curious, that’s all.

It leads to interesting surprises, for example I wrote a pretty small post on my blog weighing the pros and cons of merging vs squashing first and for some reason it’s one of the top Google results and I get consistent views for something I felt was only minor.

My thesis is that at one point, people knew this about their creations through direct interaction. Analytics are a poor substitute, especially since they do not show the subtle aspects of engagement.

Our natural instinct for curiosity as you put it is precisely the tool used by big-tech to subvert our natural human instincts into becoming replaceable cogs of their technological kingdom.

Lovely site, btw :)

And congrats on your marriage!

[0] https://www.lloydatkinson.net

I feel the same way!

I pour out my heart for pieces like this[1] and all my traffic comes from stuff like this[2].

[1] https://taylor.town/pardon-2023

[2] https://taylor.town/chatgpt-custom

I've pretty much given up trying to please the various algorithms.

Lately, I write about things that make me feel the biggest feelings, plus things I think others would find helpful.

But making things is hard without instant feedback.

Thank you! Given I made it, it's hard for me to perceive how it really looks to someone else. Yours is good too!

Thank you for the congrats!

Indeed, I have no idea why some topics gain so much interest vs others.

> Optimizing for analytics means optimizing for an immedaite return rather than following your gut on what you like.

You don't "optimize for analytics", you optimize for an objective that you are measuring through analytics. If you are simply following your gut then you simply don't have a quantitative objective.

Note that Im not saying that not having a strategy is not a valid strategy per se, I'm just pointing out that analytics is a means to an end.

The point though is that the tail wags the dog: I don't believe for a second that most objectives aren't, in fact, being selected for being measurable via the available analytics (or even merely the appearance of being measurable via those metrics).

There are things that are worth knowing and there are things that are easy to measure and the overlap is disappointingly small.

TodePond makes superb videos on development, the creative process, and cellular automata.

[1] https://www.youtube.com/watch?v=Tpvcy1gt5yU&list=ULcxqQ59vzy...

My all time favorite youtube channel. I think this video is a good introduction to the vibe of the channel

https://www.youtube.com/watch?v=WMJ1H3Ai-qs

Hard to make a blanket statement about analytics being 'good' or 'bad'. They are there as one of many tools to provide a compass and let you know whether you're headed in the right direction.

I think there's a bit of nuance here between an artist's approach versus that of a product designer, for instance. Artists seek to express an idea, feeling, or concept. Analytics, which (in the case of Youtube) optimise for a certain outcome, can easily befuddle. Designers seek to solve a problem. Analytics can help shed light on whether the problem has been solved.

Hi I wrote this post! You're right - that's what I was getting at in the post :) I've seen lots of people say analytics are "bad" recently, and that's what inspired this post (as well as starting to write an analytics-free blog).

I think it's especially twisted when analytics companies (like google) try to manipulate me into changing from an "artist analytics user" into a "engagement engineer analytics user". That's bad.

Hi there, I wrote this article!

I'd recommend also the sequel to this article: "Good and bad analytics"

https://www.todepond.com/wikiblogarden/social-media/analytic...

Consider linking to your youtube channel. If not for self promotion, then at least as a courtesy to the reader.

The todepond youtube channel is a well-kept secret

Just learning about your channel now. You make some great videos! Thank you.

Analytics flattens user engagement to more narrowly defined single dimensional metrics. And, critically, it ignores the bidirectional nature of a creator's relationship with their audience.

It's like when politicians cite opinion polls to justify their platform but ignore the fact that they often play a huge roll in the formation of those opinions in the first place. IMO, having a platform means not only that you respond to and are shaped by your audience but also that you play an active role in cultivating it

> I can see what other videos you watch.

> I can see what you type into the search bar - even when it’s not for my video.

Wait, wait, wait, what!? How does the platform allow that?

Article writer here! Yeah it's really creepy. In the "Research" tab, youtube tells me what my viewers have been searching for - especially the searches that don't produce many results. Sometimes the search terms are very general (eg: javascript), and sometimes they're more specific (eg: how to [personally identifying task]).

It also shows me what my viewers favourite videos are. Again, the 'highest rated' ones are general. But as you look down the numbers, you can see more specific things.

Well. I was in the "analytics is mostly bad" camp anyways. But the things that you just told me are more of the "analytics are far worse than you think" kind. Like "how the hell aren't those people in jail right now"?

>Like "how the hell aren't those people in jail right now"?

Were you under the impression that things you typed in the search bar of YouTube (or Google or Facebook or...) were not tracked? What about this situation makes you believe someone should go to jail?

The reasonable assumption is that it’s tracked by YouTube and mostly relegated to server logs and automated analysis, not that the search queries will be displayed verbatim to the channel owners.

They don't filter out low-volume queries?? In AdWords search term report they would blank out any queries not searched by multiple different users which was enough to not leak anything personally identifiable..

Analytics should just be one point to weigh across others when making a decision. I don't fully think it's a binary decision to use it or not.

Consider analytics. Consider user feedback. Consider your gut. Consider stakeholder requests. Consider historicals. Consider your team. Consider your debt. etc etc

I like the concept of 'walkouts.' And I'd add they don't necessarily reflect badly on those being walked out on - sometimes they're just the result of a disconnect between what a person thought they were getting into and what it turned out they actually getting. As the author mentioned, walkouts occur more when there's no incremental cost in walking out, so I've seen them a lot at sessions for conferences that have a fixed registration fee - e.g., a presentation may be great but just over one's head. Imagine walkouts must really be prevalent in free MOOCs (how many college courses would you have dropped had there been no consequences?).

Hey author here. Yes I agree about the 'vagueness' of a walkout. For any given walkout, it could be for any reason. But when you put the numbers together, you can see clear patterns.

One person walking out means nothing. But ten walkouts at the same time tells you something. It feels like analytics. And those people wouldn't feel as free to do this if there wasn't an element of vagueness.

Analytics are just some metrics. Paying some attention to metrics is usually good.

But too much focus on metrics - vs. the actual long-term goals - is always bad.

After being an extremely data-driven dev/pm/founder I'm currently making a mostly offline app with close to no analytics (apple gives install numbers but that's about it). It's fun not to know how many people click on my button. It's like a whole world of concerns melt away and I'm forced to just have (good?) taste. No days spent constructing funnels and hunting drop-off.

Would not recommend for a business that's trying hard to optimize, but for a hobby project flying without instruments is pretty fun.

Analytics are good and useful in a broader context.

They often are a valuable signal which combined with a broader environment (news cycle, activity by marketing dept, socio economic factors, etc) can yield meaningful insights.

Nothing exists in a vacuum.

How do I know? Our customers use Wide Angle Analytics and are doing just that. Their feedback indicates that yes, analytics are good.

No, analytics are not good. They're bad, actually.

They don't solve the problems they claim to solve, and they create new problems.

>They're bad, actually.

"Measuring stuff is bad"?

More like "Privacy destruction and flagrant disrespect for users in the name of slurping up as much personally identifiable data as possible that you have no right to have or see is bad."

Analytics are good, just be aware of the trap of goodhart's law.

https://en.wikipedia.org/wiki/Goodhart%27s_law

Good for what?

Analytics are an instrument panel, it’s a set of gauges. How it looks depends on what you’re trying to accomplish and the conditions of the system.

Reply on news.ycombinator.com