Ask HN: What about site statistics when people are using adblocker? | Hacker News Reader

Ask HN: What about site statistics when people are using adblocker?

9 pointskevinsimper7 comments

A lot of people are using adblockers today and I understand why people do it. Nearly all third party trackers are or can easily be blocked by todays adblockere.

That limits the ways to track users and we should respect that, but I think the problem is only for thirdparty tracking, I don't think people have a problem with that the main website tracks usage.

That brings back memories like AWstats that analyses logs to figure out which are the most popular pages. http://www.awstats.org/

But is that the future? I think so, but I have not seen anybody else talk about it. It is not as sexy and it requires a lot more to install than simply copy-and-pasting a javascript snippet, but it can't be blocked!

What do you think? Would love to hear your opinion!

View on HN

I never actually did the math to look how much money i loose / how many percent actually use a adblock. But i've also seen people who use AdBlock but allow Analytics or have no AdBlock but something like Ghostery. So i am pretty sure the numbers are always wrong.

But next to Analytics i am a big fan of GoAccess (http://goaccess.io/) which also can do HTML exports if you prefer that over their very well done TUI.

The cool thing is there is basically no install because it just parses the access logs. It supports multiple formats and i even use it with a rails app (i probably simply changed the log format to something more standard tho). You just throw the binary on the server you just need it on and point to the log file.

Edit:// I just checked a old blog targeting people with security in mind. It suggests about 80% analytics blocking O_O

Also i we use Kibana (https://www.elastic.co/products/kibana) at work which coubled with Logstash and other software can be a very mightly log analyzer. But you have to do the configuration yourself.

Thank you for that sweet reply!

Yes you are right, there is so many combinations and as you found out a lot of people are using adblockers!

I did not know about goaccess, that looks really cool! I will check that out!

The problem with analyzing logs is that it is difficult to track unique visits purely from logs. Also differentiating bots that tries to look like real humans.

Kibana and logstash is awesome, but it is not the easiest to setup if you compare it to Google Analytics.

Yeah i figured at least 20% of my 80% are bots. If not more. And you are right, without Javascript its not really possible to have detailed informations like with Analytics. Following the user around should be possible with the usual access logs, but i have yet to find a software that supports this, and i can imagine its hard to create as there are so many different combinations of everything.

Its really interesting to think about. I can see no solution for this, just how it will get worse and worse over time. People are more and more scared about Trackers and block them sometimes in way to consequent ways, hence i even had users complaining about a broken design when they had some browser extension which removed every element containing social media names in the tags. Something like <section id="why-i-hate-facebook"> is appearantly not acceptable for some.

I only really see three solutions:

One is creating a service that uses a sdk that works into whatever webserver you are using and then slightly alters your html to input different detection methods. That would make it impossible to block or possible to detect a block. (index.html was loaded, but .js was not)

Allowing a tracker to be served from a subdomain. This way it is not thirdparty and should be okay.

Analytics over websockets.

---

But it will not be easy and will probably be a weapons race with adblockers.

The problem is really only with thirdparty trackers, that when have data across multiple sites can make a much deeper profile about you than a single site can make without you allowing it.

So if we can convince everybody about that, that would be great!

I was going to write about how one could fix this with CNAME'ing the tracker/ad services and beeing able to choose custom file names. AFAIK most blockers just block requests based on patterns and not actually look at the DOM (sure some do, but those are paranoid fuckers anyway).

You are right that the profile it creates can be scary, and honestly i block everything i can. I just happen to think about the business side as well.

Remember when we tracked with images? I dont even remember why, but in the early days all we did was a 1x1 image. Piwik even still supports tracking images, i dont know how good the data can get that way tho.

I have only one site where I have access to server logs, (I think). I am using Smugmug to host my photos. Their stats exclude my logged in sessions.

So I have about 1700 pageviews under SmugMug stats whereas only 63 pageviews on Google Analytics this month. GA shows only 24 sessions. Smugmug doesn't have sessions info. It is kind of hard to believe that so many people are using AdBlockers but I guess that is the future.

But since Smugmug doesn't provide sessions info, it is possible that there were a few people with AdBlocker are responsible for most of the views under Smugmug stats.

You can either create your own tracking solution, load one via 3rd party, or load a 3rd party via your own domain making it seem 1st party.

In general tho, most systems use a lot of JavaScript, and you'll need to figure out how to best capture the data and record it. Depending on the volume of traffic, you could easily create pretty sizable amounts of data.

I help do this kind of work, at work, and you need to realize you'll never get 1 to 1 data for all visitors, but you should try to get as large a sample as possible and use statistical reasoning when discussing issues.

Reply on news.ycombinator.com