Counting instead of tracking
fcused.at
fcused.at
That said, counting and not tracking as a concept is awesome, and more thought can go in that direction and I expect it will be fruitful.
Tracking isn’t inherently bad. It’s the history of malicious and overzealous ad networks that make it so. Also that analytics trackers don’t have a “we won’t sell your data even if we are bought out” philosophy. I wish we could accept that on a site by site basis understanding behavior is very valuable, and can greatly improve the experience of visitors, but unfortunately this gets coupled with ads, marketing and other internet four letter words that make it hard for folks to see it in a positive light.
I work for an analytics company. I totally respect folks who blindly (or thoughtfully) hate them and hate JavaScript - it’s clear nobody has built a system that is consciously safe and secure from the bottom up in a way that a website owner could have total confidence in and that’s a bummer, but I wish the HN community would make an effort to have a discussion about this rather than knee-jerk rage any time someone talks about analytics or tracking.
I bet people would do that if you frame it right. E.g. "help us improve our website! You will get to see our new website first and we will see from you interaction how we can improve. It would be a great service to us."
Marketings job is to sell this - they sell all kind of things why not sell tracking
I’m not sure tracking can be solved well; any information about it is deliberately vague, because it forms your api contract, and you want to swap out the rules from under the user’s feet, because that data is just so damn useful that it can really be sold off anywhere. I find it difficult to believe marketing can get trust out of someone who doesn’t: at best, it can continue as it currently does, keeping people who don’t think about it from thinking/caring/distrust about it. The whole thing is so opaque, unregulated and nonverifiable that its not even clear that you won’t just flat out lie about how tracking data is being used / sold to. Or even what exactly is being tracked.
I’m pretty sure if you don’t trust tracking, you don’t trust any of its various forms. It might be too much effort to avoid, or you don’t care enough, or don’t really expect anything to go wrong for you from it (the average person leads an average life; there really is nothing to hide, at least from most current web-tracking) but I think the trust gets lost in total.
Also, yes, you could argue that analytics for e.g. A/B tests are actually "improving the website/service". Yet I'd still prefer to opt out of them, because those A/B tests usually encourage making the software more addictive (erm, "engaging") and less useful ("only single-digit-% of users use this feature, so let's drop it"). I don't trust the motivations of those companies, so I don't believe they'll use A/B tests for my benefit as a user/customer.
What's the difference between being bought out and selling the data? Why would a company like that be bought out if not for their data?
Who gets to use the data.
For example, a retail place that tracks your data and uses it to improve its selection and/or offer helpful things to you doesn't necessarily sell to the bidder that wants to use it to manipulate you into buying financial products. Plus 37 other companies that want the data.
If someone buys the retail company, however, they get the data. They may or may not have been bought for the data. I should mention that the data is a package deal. You aren't getting the fruity pie filling without taking the crust and packaging as well.
I care about hits per page (ideally by humans, unique or not), the time it takes to generate the page server side, and how long it takes until the DOM is finished loading. I throw that into buckets, i.e I have hourly counts for the last few weeks, daily counts for the last few months, and so on; the only stuff that will be kept "forever" are yearly counts. I still rarely look at it, but at least now it takes up little space.
As a consequence of this you can end up having SEO people destroying a site that is designed for humans to read as they still have quaint notions about keyword stuffing. Because these people have no real knowledge of code or the products the website hopes to sell they have a less than useful effect. They will track that one conversion and give themselves brownie points for their SEO and disregard the thousands of others who didn't get past the wall of SEO keywords to get as far as a converting page.
What surprises me is how poor the server side analytics tools are. Server logs should be great and what SEO people look at side by side with their frontend analytics. They should be doing this for basic sanity reasons, to see if the 'hits' add up or whether some slow loading pages burdened with scripts that never fire is mangling the statistics. Or even to check if some developer has 'gone rogue' and stopped the scripts being loaded on pages that matter to them.
Yet despite the obvious utility of server side tracking - not least the no cookie message - nobody is doing a good job of providing decent software for it that would pass muster with SEO types.
I do have one question: how would something like this work when your webpages are being cached or when you're behind a CDN? As far as I can see none of this can work unless all requests hit your web server.
Does anyone know of any solutions that work similarly, but would allow for CDN's and caching? I'm aware that these will probably need JS, but I'm fine with that.