Overview
Once Helix has identified all the records from various data sets related to the same individual, it combines the best data from those sources into a single record. But determining which data is "best" can depend on several factors, and it won't always be the same answer every time. For example, the best source for one person’s email address might be Archtics, but for another person it could be Mailchimp. In the Assemble Golden Record tab, you can define the selection rules (tiebreakers) Helix uses to make these decisions.
Tiebreaker rules
The left column contains a list of all the defined Golden Record fields. Click a field's name to view or change the tiebreakers. Each tiebreaker examines one attribute across multiple data sets.
Choosing the "Data Set" attribute allows you to manually order the data sets from most to least reliable. For example, you could choose to take a user's cell phone number from Archtics if it's filled out (not null) there, then fall back to Marketo as the next-best option, and so on. Only if the value in the prioritized data set is null will Helix move only to the next data set. Therefore if there is poor data entry or a blank value in the prioritized record, this value will persist to the Golden Record. If this is not your intent, please see the best practices for creating Helix data sets and mapping them to the Golden Record.
If two or more data sets have the same value for an attribute, then Helix uses the next tiebreaker rule (if any) to decide which of those sources to use. If there is still a tie after the final tiebreaker, Helix will arbitrarily choose one of the remaining data sets. Therefore, if you plan on using the data set prioritization for your tie breaker rules, then KORE recommends adding a second rule for when there are more than one records from the same source within the golden record group. A great option is to use the ‘Created On (Source)’ and ‘Update On (Source)’ timestamp (date time). However, it is important that you have mapped these fields in your helix data set definition.
"Created On (Source)" is a timestamp (date time) that represents when a record was added to the source platform. Select this attribute in descending order to take information from whichever data set was most recently created. For example, a customer fills out a webform from Formstack and submits a new email address for their preferred contact method. This Formstack record will have a more recent created on timestamp than their Archtics record. The Golden Record would use the new email address from Formstack because that record would have a more recent timestamp for "Created On (Source)".
"Update On (Source)" is a timestamp (date time) that represents when a record was added to or last updated within it's source platform. Select this attribute in descending order to take information from whichever data set was most recently updated. For example, a customer who receives newsletters might submit a new email address to Marketo, but Archtics would still have the old one if they haven't changed it with Ticketmaster yet. The Golden Record would use the new email address from Marketo because that record would have a more recent timestamp for "Update On (Source)".
Caution: Most platforms will overwrite the updated date time whenever any value on that record changes. Therefore, that date and time value may not indicate a field’s value that changed is the one you are populating on the Golden Record. Right now Helix does not support individual update date time values for each golden record field so even if the source system does support this level of detail, Helix is not able to process more than one update date time field for each record.
Caution: An attribute might seem numeric without actually having a numeric data type. For example, a phone number attribute might be stored using the string data type to accommodate symbols like parentheses. It would therefore not be available to choose as a tiebreaker.
Except for "Data Set", each attribute can be evaluated in either ascending (“asc”) or descending (“desc”) order.
- Ascending: Oldest date / lowest number first
- Descending: Newest date / highest number first
Add or edit tiebreakers
First, select a Golden Record field in the left column. Next, use the + (plus) button to add another tiebreaker, then select an attribute and order from the drop-down lists. You can also click and drag the move icon to reorder the tiebreakers. Use the trash icon to remove an attribute from the list of tiebreakers.
Next, open the Order drop-down list and select ascending or descending. Or, if you are adding a Data Set tiebreaker, open the Order drop-down list to view the available data sets and drag them into the desired sequence. Hover the mouse cursor over the Order field to view the list.
Click the Save button when finished.