Introducing the Pixel tracker
- The purpose of the Pixel tracker)
- How it works
- Considerations when using the Pixel tracker with the Clojure collector in particular
- Next steps on the Snowplow tracker roadmap
Our aim with Snowplow has been to enables companies to track user events across all platforms and devices. That means enabling tracking offline events, as well as online events, and mobile events, as well as web events.
- Views of HTML emails
- Views of ecommerce products on 3rd party marketplaces
- Views of pages on 3rd party hosting sites e.g. Github
In these cases, you can use the Pixel tracker to track events directly into your Snowplow stack. Doing so enables you to analyse complete customer journeys: tying together data on the emails a user has opened with their subsequent web browsing behavior, for example.
page_urlthat the event occurs on or the
- Appends those data points as key value parameters on a query string
- Makes a GET request to your collector including the above querystring, so that the data relevant data is passed into Snowplow
page_url. Instead, you have to hardcode those values into the request string, and append those values onto an image request for the Snowplow tracking pixel.
- The number of unique visitors to a web page
- The number of events / page views (e.g. the number of times an email was opened)
Anatomy of a Pixel tracking tag
The standard Pixel tracking tag looks something like this:
There are several things to note about the tag:
1. It is a plain HTML image tag
<img src...> tag.
2. Only a handful of parameters is passed into the collector
The data points passed are:
3. The parameters are hard-coded
As a result, is it necessary to generate a unique tag for each individual web page / email newsletter that you want to track. To make it easier to generate the tag, we have created a wizard
The Pixel tracker works with both the Cloudfront collector and the cross-domain Clojure collector. However, there is an important difference between the way it works with each collector, that has implications for:
- What user data is captured
- Which services you should use the Pixel tracker with
When the Pixel tracker is used with the Cloudfront collector, the only data captured is:
- The name / value pairs stored on the query string i.e. the
- The data captured as standard by the Cloudfront collector i.e. the
useragentstring and the
timeof the event
This limits the scope of the analysis that can be performed with the data: if for example we’re using the Pixel tracker to track views of a README page on a Github repo, we can see how many times the page was viewed but not how many unique users viewed the page, because no
user_id has been set or stored.
In contrast, when using the Clojure collector with the Pixel tracker, a
user_id is set server-side, and saved to a cookie on the user browser. This provides better data for analytics: you can now analyse at the number of unique visitors to a web page.
However, you need to make sure that you are allowed to drop a cookie on a user, on a web page owned and managed by a partner or 3rd party service provider. It is your responsibility to ensure that you only drop cookies on web pages where the owners of the web page / service provider are happy for you to do so. There are many examples of providers who do not: for example eBay explicitly does not allow you to drop cookies on your listings pages. Snowplow takes no responsibility for your use of the Pixel tracker. It is your responsibility to ensure that you abide by the terms and conditions of any 3rd party services and hosting companies you employ this tracking technology on, and we urge extreme caution when deploying the Pixel tracker in conjunction wiht the Clojure collector on sites owned and operated by 3rd parties.
The Pixel tracker is only our second tracker: to fulfill our vision of supporting event-data collection across many more platforms, we need to launch a wide range of new trackers.
We are getting close to launching an Arduino tracker for Snowplow, which will enable data collection from physical events into Snowplow. As you might expect, mobile trackers (especially for Android and iOS) are high priorities oadmap, alongside other software trackers (e.g. Windows 8). It will take a lot of work (and trackers) to fulfill our vision of enabling data collection across any platform in Snowplow, but we are getting there steadily.