One of the main reasons that companies adopt Snowplow is to build a single customer view. For many of our users, Snowplow lets them for the first time join behavioral data gathered from their website and mobile apps with other customer data sets (e.g. CRM). This simple step drives an enormous amount of value.
However, this is just the beginning. Most companies engage with users on a very large number of channels - not just via mobile apps and the web. To give just a few examples, companies will often engage with their users via:
- Outbound messaging, e.g. email or push notifications
- Display advertising
- Telephone calls (e.g. as part of a customer acquisition, service or support)
Incorporating data from these channels into your single customer view is essential if you want to built a complete picture of your users’ complete journeys.
In this post, we’re going to look at how you can use Snowplow to track user behaviour from some of these other channels, with a view to building a more complete customer view - looking particularly at email, push messaging, telephone and display advertising.
How does Snowplow enable companies to build single customer views?
Before we dive into some of the different data sources that we need to build our single customer view, it is worth recapping on how companies use Snowplow to build single customer views. The process is as follows:
- Track and collect event-level data from different places where you engage with your users, e.g. your website and mobile app, via our event trackers. With each event, record as many user identifiers as are available (e.g. 1st party cookies, 3rd party cookies, IDFVs, IDFAs, machine fingerprints, login IDs, Twitter handles, Facebook IDs, email addresses). At collection time we do not need to decide who is performing a particular action, but we need to collect as much evidence as possible so that we can decide that later on in the pipeline
- As part of a separate data modeling process downstream, build a map or graph of the different user identifiers that belong to the same individual user
- Use that map to assign each event to a single user, and for each user stitch together the different events that occurred across different channels, to build a consolidated view of each user. This is the single customer view
- Optionally, join that table with other customer data sets (e.g. CRM) that you have available in your data warehouse
Note that the above approach is very diffferent from that taken by packaged analytics vendors (Google Analytics, Adobe Analytics, Mixpanel, Kissmetrics, Keen.io et al) because packaged vendors insist that you decide at data capture time who a user is, whereas Snowplow enables you to defer that decision, and continue to evolve the logic for making that decision as
- you learn more about your users, and
- you start stitching data together from more and more channels.
The first key element, then, to building a complete single customer view is to strive to track events across all of the different touchpoints and channels where you interact with your users. The second key element is to ensure that events from a single user on one platform can be mapped together with events from a single user on another platform, so that your view of each customer is assembled from the totality of their interactions across all platforms.
Tracking event level data from locations outside the web and mobile
Broadly speaking, we define two categories of software system which can power your user engagement channels:
- Systems built by you
- Systems provided by third-parties (for example SaaS vendors)
In the case of (1), you track event-level data using Snowplow Trackers. In the case of (2), you use webhook integrations to enable you to track event-level data from the third party systems.
For many engagement channels, you can opt to engage users using either your internally-built systems or third-party systems, and this will determinate how you use Snowplow to track those events. Let us dive into some example channels to see this in action.
Email is a very widely used channel. Whilst some of our users use third party systems to manage their email, including SendGrid, Mandrill or Mailchimp, others choose to build their own systems, often employing tools like AWS Simple Email Service.
If you’ve built your own system using e.g. Java or Python, you can track email related events in Snowplow as follows:
- Track email send events directly from your server-side email sending application. This means recording an event when an email is sent to an individual
- Embed a Snowplow pixel in the email to track if that email is opened
- Decorate any links in the email to your website or app with querystring parameters so that you can spot users clicking through emails into your website or app. If you can decorate the links with a hashed version of the email address, you can then map email addresses to cookie IDs, enabling you to join email event data with your web event data
For users who want to use one of the great SaaS email marketing services, many of them now support webhooks. This means that as events such as email sends occur, the email provider will send granular event-level data describing those events to an HTTP endpoint of your choosing. By selecting your Snowplow collector as that HTTP endpoint, you can grab those events as part of your Snowplow data pipeline - either to load into your data warehouse, or even to act on in real-time via Amazon Kinesis.
This means you see not just what your users have done in your website and mobile app, but which emails they have opened, which they have ignored, and which they have clicked on.
Push messaging is an incredibly powerful channel that puts your message very prominently in users most important device: their cell phone.
If you have written your own push notificaiton system, you would integrate our trackers to record each push notification that is sent to each of your users.
We are delighted that as of release 75, you can now ingest all your event-level Urban Airship data straight into your Snowplow alongside all your mobile and web behavioral data. That means Urban Airship users can automatically track the following events from Urban Airship, alongside all other event data:
- Application opens and first opens
- Application closes
- Message sends
- In-app messages displays
- In-app messages expires
- In-app messages resolutions
- App uninstalls
That is incredibly valuable data to store alongside the user’s web, mobile and email engagement data.
For a large number of companies, the traditional telephone is a vital customer channel, both for customer acquisition and customer support.
There are companies running Snowplow who track calls using our trackers embedded in their own telephony systems, and others who have made use of our webhook integration with CallRail. You would typically record that a call had taken place, who initiated the call (the user or the company), the purpose the call, the length of the call and any outcomes, for example if a transaction was completed, or a support ticket raised.
A large number of Snowplow users track ad impression and ad click events through Snowplow, enabling them to identify which ads a user has seen and clicked on prior to arriving at a website or app and making a transaction; this kind of tracking enables Snowplow users to build sophisticated attribution models.
Since release 72, it is also possible to use Snowplow to do click tracking. Our Clojure Collector supports redirecting users through a link - so you can track clicks-outs from ads by redirecting them through the collector which records the click before sending the user on to the ultimate click destination.
Building a complete picture of all the different events on all the different touchpoints you engage your users is a big job, one which should include interactions in the physical as well as the digital world. We are working hard to extend Snowplow to make it easier to do just that. We would love to hear from you if there are additional integrations you’d like to see us roll out.
Interested in collecting and acting on all your event data, not just data from apps and the web?
Then get in touch with the Snowplow team!