It seems that users will be very tempted to write additional "session" information, for example some kind of visitor id to the logs. They would need that to, for example, correlate which people visit the homepage with which referrer then buy a product. That's a core use of most analytics products.
Considering the raw logs are timestamped and public, that then sounds like a rather large privacy hole for the users.
I would imagine some kind of pre-aggregation happening, and then making available aggregated data instead of raw logs.
I'm not sure I'd go with the idea (yet). The power of Google Analytics is in the "Analytics" pieces. This seems to be only the data gathering piece. Granted that is also important. But this would need a ETL & BI system or something like that in-front of it to be useful.
It is a great start.
I also really enjoy looking through the source code for projects like this and seeing how other coders do things.
I don't think there is an answer. Fake clicks can only be determined by a 3rd party -- or else sites could host the ads and report whatever numbers they want.
The only way forward that I can see is anonymous aggregation by a 3rd party that is not the advertiser or the website, and which has incentives to protect the users.
Imagine a system with dozens of 3rd party collectors that anonymize browsing data. The data is submitted by users who have a plugin for their collector. The collectors then sell the anonymous data to the advertisers. Advertisers can then compare data from different collectors to determine fake clicks, and what the level of payment should be.
There are two big problems with a scheme like this. First, anonymous data is much less valuable to advertisers. I think this can only be solved by technical means; making anonymous data the only information available. Second, why should users have a plugin that tracks them? Access to content seems like the obvious answer; users with a plugin (being tracked by a collector) could be allowed to read the NYT and the NYT website could be paid by the collector.
Realistically, I don't see either of the problems being resolved.
At least for me, the real value that GA provides is being up all the time and giving me peace of mind that hits won't get dropped. Running my own analytics infrastructure is possible, but would get in the way of what my site was actually trying to do.
What the open source analytics world needs is better infra, not better collection.
Important to distinguish between a blogger's utility from getting session information and the huge privacy risk of sending everyone's reading to a central provider (GA).
I use piwik and I have it set to respect do-not-track; I assume that means if someone visits from a private tab in FF, I won't even see a hit (piwik experts -- correct if wrong).
But I also benefit greatly from the analytics from non-private browsers -- seeing that users are, for example, clicking on the footnotes incentivizes me to deliver thorough arguments in the future.
If this gains widespread use then the transparency points will make it far worse for privacy than using Google.
Why? Google have access control over the data or view it in anonymised form when crunching. But this data will open to anyone and would allow people to correlate visitor sessions not just on a site but across other sites that use this, to the degree that you could start deanonymising the internet activity of people.
It's a strange world in which we can say, "At least today only Google has the data and they've made it hard to get to".
Some of us just don't want to live under Google's umbrella.
Personally, I have no interest in sacrificing my privacy for "free" services. I'm not important, nothing I do is controversial or of any significance, but having my activities logged and monitored by a giant all-seeing eye is still uncomfortable. It's disturbing, to me, how accepting of these things people have become.
20 comments
[ 3.0 ms ] story [ 49.7 ms ] threadConsidering the raw logs are timestamped and public, that then sounds like a rather large privacy hole for the users.
I would imagine some kind of pre-aggregation happening, and then making available aggregated data instead of raw logs.
This is one of the main reasons for fair analytics to exist.
It is a great start.
I also really enjoy looking through the source code for projects like this and seeing how other coders do things.
It's possible to build any type of chart/fancy dashboards on top of it
I'd love to have suggestions on how improve this particular aspect :)
The only way forward that I can see is anonymous aggregation by a 3rd party that is not the advertiser or the website, and which has incentives to protect the users.
Imagine a system with dozens of 3rd party collectors that anonymize browsing data. The data is submitted by users who have a plugin for their collector. The collectors then sell the anonymous data to the advertisers. Advertisers can then compare data from different collectors to determine fake clicks, and what the level of payment should be.
There are two big problems with a scheme like this. First, anonymous data is much less valuable to advertisers. I think this can only be solved by technical means; making anonymous data the only information available. Second, why should users have a plugin that tracks them? Access to content seems like the obvious answer; users with a plugin (being tracked by a collector) could be allowed to read the NYT and the NYT website could be paid by the collector.
Realistically, I don't see either of the problems being resolved.
What the open source analytics world needs is better infra, not better collection.
Is there an OS alternative to Google Analytics that's comparable in featureset?
https://piwik.org
Important to distinguish between a blogger's utility from getting session information and the huge privacy risk of sending everyone's reading to a central provider (GA).
I use piwik and I have it set to respect do-not-track; I assume that means if someone visits from a private tab in FF, I won't even see a hit (piwik experts -- correct if wrong).
But I also benefit greatly from the analytics from non-private browsers -- seeing that users are, for example, clicking on the footnotes incentivizes me to deliver thorough arguments in the future.
Why? Google have access control over the data or view it in anonymised form when crunching. But this data will open to anyone and would allow people to correlate visitor sessions not just on a site but across other sites that use this, to the degree that you could start deanonymising the internet activity of people.
It's a strange world in which we can say, "At least today only Google has the data and they've made it hard to get to".
Personally, I have no interest in sacrificing my privacy for "free" services. I'm not important, nothing I do is controversial or of any significance, but having my activities logged and monitored by a giant all-seeing eye is still uncomfortable. It's disturbing, to me, how accepting of these things people have become.