Announcing Protocols Transformations – clean your data as it’s collected

It should come as no surprise that data quality issues are a growing concern among organizations today, both large and small. A recent Forrester report quantifies the impact. It found that, on average, 32% of marketing teams' time is spent managing data quality, while 26% of their marketing campaigns suffered due to the quality of that data. It’s no wonder executives are starting to take notice, with 82% of companies making data quality a top priority. 

So how do companies improve data quality? A well-defined tracking plan that maps business objectives to user actions is a great start. Segment’s Tracking Plan feature was launched as part of our Protocols product last year to proactively help customers improve data quality. Since then, it’s helped hundreds of companies do just that. 

However, it’s a simple fact of life that mistakes happen. 

Even the most exacting data collection processes are subject to human error and organizational complexity, and a single implementation error on a business-critical event, like Lead Captured, can cost your business hundreds of thousands of dollars.

Previously, the only means of fixing bad data was to ask engineering to fix the instrumentation code, a lengthy process that could take weeks if not months. 

For example, let’s say you’re an e-commerce retailer and have followed the Segment e-commerce Spec to track Order Completed events. In your latest app release, a new engineer changes the event name to orderCompleted, a naming convention they used at their previous company. 

This casing change seems innocuous, but it sends your organization into a tailspin as funnel reports, marketing campaigns, and attribution models all break. After a brief moment of panic, you create a JIRA ticket, beg engineering to push the fix, and wait patiently for your customers to update their app version. If you’re lucky, a month later, your data collection has returned to normal.

We’ve heard cases such as the above repeatedly over the past year, so today we’re excited to announce the launch of Transformations, now included in Protocols, to help customers fix bad data immediately.

Transform your data to move faster

With Transformations, you can change data as it flows through Segment to either correct bad data or customize the structure of an event. The next time you see an event come through with the wrong event or property name, instead of waiting for your next app release, you can immediately apply a fix. And if you find yourself needing a way to change an event to satisfy a destination requirement, transformations provide an easy way to map event and property names.

Transformations opens up a slew of interesting new use cases for our customers. Let’s dive in!

Conform to Segment specs with Transformations

Tracking plans can be a somewhat new concept for teams to become acquainted with, so we've developed sample tracking plans for a variety of industries and use cases.

• SaaS Tracking Spec →

• Mobile Tracking Spec →

• E-commerce Tracking Spec →

• Video Tracking Spec →

These specs also allow Segment to map key events to the unique requirements of each Segment destination. For example, if a video analytics destination expects to receive Video Content Started with a specific set of named properties, we can build out destinations to format the payloads accordingly based on the video spec. 

For many businesses, there are often situations where events don’t conform to a Segment Spec, which in turn causes issues with your destinations, Transformations makes it incredibly easy to update those events to conform with the spec. To reuse our earlier example, with Transformations you can change an event name from completed_order to Order Completed to conform to Segment’s e-commerce spec.

Example transformation for Order Completed event

Transform your data for destinations

Segment currently integrates with 300+ destinations – business tools or apps that you can connect via Segment with the flick of a switch. Each tool expects to receive data in slightly different formats, which our integrations handle gracefully. 

However, there are unique scenarios in which a destination requires a specific property key or event naming convention to function properly. When multiple destinations require different conventions, customers are forced to send the same property or event twice, with different destination-relevant names. 

With Protocols Transformations, it’s now easy to align your event collection with your business model, and use transformations to map event and property names to satisfy the unique requirements of destinations like Google Analytics, Sailthru, and Iterable. 

Scale data transformations across the company

If you’ve used data transformation solutions before, you may have noticed that transformations can quickly become unmanageable as individuals look to solve their myopic problem. When researching transformations, one customer told us that in their current system they have a transformation that changes an event like so:

1st Transformation:    order_updated >  orderUpdated

2nd Transformation:  orderUpdated   >  Order Updated

3rd Transformation:   Order Updated >  order_updated

Without strong controls and visibility, every individual will create transformations to suit their specific needs, even when those needs conflict.

In the example above, it’s reasonable to think that the Google Analytics owner might create a transformation to map all the relevant events to Order Completed at the Source. That change would immediately break the analyst’s queries and limit their ability to analyze granular user activities.

To limit these issues and make transformations scale, we built in smart controls and high visibility. Specifically, each transformation rule must have a distinct combination of source, type, destination, and event name. 

This ensures:

• A single event can only be transformed once, preventing situations like  order_updated > orderUpdated > Order Updated from happening

• Property level transformations are always grouped by the original event name which consolidates property-level transformations under a single event

• Transformations can be sorted and filtered quickly across any of the distinct dimensions

To improve visibility, we’ve added advanced filters and sorting functionality so you can see all transformations for a source, destination, event, etc. We also will soon include Destination Filters and Privacy Controls in this view to minimize the risk of conflicting operations.

Protocols Transformations list view

At Segment, we believe that all businesses should be able to collect data once and be able to send it everywhere it needs to go. Transformations is a further extension of this idea – simply collect data to Segment once, and we will help get it everywhere it needs to go, and in perfect shape for your business needs.

Ready to transform your data? Click here to request access to Protocols. Or if you’re a Protocols customer, visit your workspace to give Transformations a test drive.

You can also learn more about Transformations by checking out our docs.