Announcing Streamlined RampID Files (5/31/19)
Starting June 12, 2019, we’re moving toward an even more people-based identity graph by removing redundant RampIDs when we deliver RampIDs for PII-based data files to our customers and destination partners. Removing these redundant RampIDs will greatly streamline these files.
Note
Updated: Starting September 18, 2019, we'll also remove redundant RampIDs when we deliver RampIDs for CID-based data files for LiveRamp Activation or multiple-RampID Measurement Enablement customers.
Overview of the Change
If you’re a LiveRamp Activation or multiple-RampID Measurement Enablement customer who receives RampIDs, you may have noticed that you receive many more RampIDs in the files we send you than the number of records you sent to us. For example, you might have sent us 200,000 records from your CRM, but we sent you back 800,000+ RampIDs.
Since a number of these RampIDs are redundant and do not provide additional value in terms of reach or match rate, starting June 12, we will no longer be including these RampIDs in the files you receive from us.
The benefits include:
Less overlapping redundant information: Streamlined match of RampID to customers’ CRM systems
Smaller RampID files: Faster processing times, faster transmission times, and lower storage costs
Details of the Change
Derived RampIDs are LiveRamp identifiers that are generated for any PII touchpoint, such as name and postal, email, or phone, that LiveRamp has yet to match to a complete set of PII. Derived RampIDs currently make up the majority of RampIDs we deliver to you. These RampIDs start with “Xi”.
After we confidently resolve that PII touchpoint to an individual, we convert the derived RampID to a maintained RampID (our identifier that represents an individual that we fully recognize). Maintained RampIDs start with “XY”.
Currently, our systems always return a derived RampID even if the PII touchpoint now resolves to a maintained RampID. Starting June 10, these redundant derived RampIDs will no longer be included in files we return to customers or delivered to our destination partners that receive RampIDs.
This will have minimal impact on reach and match rates, and minimal impact to your ongoing workflows.
FAQs
Which regions are affected by this change?
This change impacts U.S. data only. Customers in Canada and Australia will also be impacted as their files are built using the same technology and the same ingestion system.
Does this change affect input files that don’t contain PII?
This change only impacts PII input files (files including name and postal, email, and phone), and does not affect input files with device identifiers (such as cookies or mobile device IDs).
Why did we make this change?
Customers have asked why they receive so many RampIDs per record and if there’s a way to determine the value of each RampID. Since at least 50% of the RampIDs don’t carry any value (i.e., derived RampIDs that map to a maintained RampID), we decided to remove them.
Some customers also have expressed that they have problems dealing with large RampID files in their system and that it takes a long time for LiveRamp to transfer the files. This will speed up processing.
LiveRamp seeks to continually improve our identity graph. The ideal state is when a record with several touchpoints resolves to only one maintained RampID. Removing redundant RampIDs is the first step in this direction. More improvements will follow.
What are the advantages of this change?
Fewer RampIDs for customers that ingest RampIDs
Closer to our goal of having a single maintained RampID per record
Anticipated faster processing times in LiveRamp system
Faster delivery of RampID files (less time to transfer the file over the Internet)
Lower storage cost for customers who store RampIDs in a data lake or measurement system
How does this change impact measurement workflows?
Measurement customers who send PII-based files and get back multiple-RampID files will receive 50-70% fewer RampIDs. The redundant derived RampIDs didn’t carry any additional identity information, so this will make it easier and faster for customers to analyze the data.
Customers that have long measurement cycles might experience a minor loss in overlap (~1% per month).
Why might a file still contain more than one RampID per record?
Multiple RampID might still be returned when LiveRamp does not have enough evidence that several touchpoints of a record belong to the same person. LiveRamp will always be conservative in consolidating touchpoints into one record.
Some situations that result in multiple RampIDs are listed below:
Some users have so-called “garbage emails” that they only use for certain websites. It is unlikely that we can add those emails to a maintained person in our identity graph since users don’t provide more PII touchpoints to that email.
In up to 5% of the cases, we see PII from outside of the U.S. or typos/errors in emails or phones which result in a different (derived) RampID.
A shared touchpoint (e.g., a landline phone number shared between multiple household members) that does not get resolved together with a name can result in a different RampID that usually does not align with a RampID from a NAP (name and postal) or an email that belongs to a single record.
A touchpoint that LiveRamp sees for the first time (someone moved, changed names, got a new phone, create a new email) will always resolve to a different RampID at first. Over time LiveRamp will merge that touchpoint to an existing, maintained RampID.
For questions about this change, contact your LiveRamp rep for more information.