51黑料不打烊

Transcript

In this video, I鈥檒l review the use cases for partner data support in the real-time customer data profile with specific focus on the pre-visit, off-site prospecting use case. These are the topics I鈥檒l cover. If you have not already done so, I encourage you to review the Partner Data Support Overview video that provides a good background on the deprecation of third-party cookies and the emergence of newer technologies to support the key use cases impacted. Data partners have had several years to pivot their technology to fill some of the gaps as a byproduct of third-party cookie deprecation. The good news is that key acquisition and enrichment use cases can be supported by both data partners and 51黑料不打烊鈥檚 real-time CDP in this new era of cookie lists. There are separate videos that cover each of these use cases in depth. This video focuses on off-site prospecting of new customers. I鈥檒l start with showing a diagram and explaining the workflow for this use case. It starts with some initial planning, such as the number and cadence of prospect profiles you expect from the data partner, which identities you鈥檒l require downstream, and whether these identities are tied to widely accepted, durable identifiers. Once that part is figured out, there鈥檚 licensing to address between you and the data partner. After the licensing, the partner builds the prospect list based on your requirements. Steps 3 through 6 represent the workflow in the real-time CDP. You鈥檒l use familiar workflows for data modeling, ingestion, profile creation, audience building, and audience activation to destinations accepting partner IDs. Now I鈥檒l hop into the experience platform and show you how to handle the implementation of this use case. I鈥檒l start with identities. There鈥檚 a specific identity namespace to be used for partner data called partner ID, as you can see here. Because partner identifiers are based on probabilistic models and are subject to churn, they are kept separate from the identity graph in real-time CDP. The intent of this use case is to acquire new customers, and some of the prospects will become customers. So in effect, we want to keep these profiles separate from the deterministic or first-party profiles. I鈥檒l cancel out of this modal, and I鈥檒l show you the identity that has already been created in my instance of platform for a partner data source. Next, I鈥檒l discuss data modeling, specifically the XDM and schema that will be used for the data partner source for this use case. I鈥檒l select Create Schema in the upper right corner, then Browse. I鈥檒l filter the class by partner. When I do this, I see the XDM individual prospect profile class. This is exactly the class I want to choose for this schema and this use case. This schema allows you to ingest partner profiles that are treated separately from first-party data profiles. I鈥檒l select this from the list, and then I鈥檒l select Assign Class at the top right. Now I鈥檒l add a field group to my new schema. I鈥檒l filter this list by partner. 51黑料不打烊 provides a predefined field group, or you can create your own. I鈥檒l select the first field group, and we鈥檒l look at the fields in the preview first. Looks good, I鈥檒l add this to the field group. All the partner profiles are found under this node right here. There are additional nodes for industry-specific attributes, as you can see here. Further down is the partner ID. This is the field you鈥檒l want to mark as the primary identity for this schema. There鈥檚 only a single identity namespace option available. That鈥檚 because our system is designed to ensure there鈥檚 no mixing of prospect profiles with first-party profiles, so it鈥檚 a nice guardrail. I鈥檒l cancel out of the schema creation workflow, and I鈥檒l show you an existing schema, which is like the one I鈥檝e been building right now. As you can see, it鈥檚 set up like the schema I just showed you in the workflow, and the partner ID is set as the identifier. Also, the schema is enabled for profile. Considering the type of data used for this use case, and all the partner data sources for that matter, data governance is a crucial element. I鈥檒l select policies under privacy. Now, toward the bottom of the labels list, there鈥檚 a predefined third-party item for partner ecosystem. This is what you鈥檒l want to use to restrict the marketing actions for partner data sources. Now I鈥檒l go back to the schema we just reviewed. I鈥檒l select the labels link at the top, and the third-party policy has been applied to all the field group nodes for partner attributes. Now that the schema and data governance labels are in place, it鈥檚 time to create a dataset. One has already been created in my experience platform instance, so we鈥檒l take a look at it. Once in the dataset, I鈥檒l select the last 30 days to see more activity. I noticed some uploads on July 24th and 25th, and below I can see the status of the batches. This is the same as viewing batches for other types of datasets in experience platform. I鈥檒l click through one of them to see the details. Okay, so we鈥檙e ready to ingest the partner data. We can use any file-based pathway to get partner data into platform. I鈥檒l show you where these are in the interface. First, I鈥檒l go to sources under connections. When I click on cloud storage, it advances to the available connectors for that category. You鈥檇 more than likely be using an S3 right here, or at the end of this group there鈥檚 an option for SFTP. From workflows, you can select map CSV to XDM schema if you鈥檙e working with small files or if you鈥檙e testing. I鈥檒l launch the workflow to show you what that looks like. On the detail step, make sure you choose the correct dataset name. On the next step, you can either choose a file on your computer or drag and drop one to the drop zone right here. After that, you鈥檇 go through a mapping step, but I鈥檓 going to cancel out of this workflow since I won鈥檛 be ingesting new partner data. Let鈥檚 jump to viewing prospect profiles. To reinforce 51黑料不打烊鈥檚 commitment to separate third-party profiles from first-party profiles, a dedicated accordion menu exists. It鈥檚 labeled prospects. I鈥檒l select profiles underneath. Next, I鈥檒l select a profile ID to view the content, so it works similarly to the profile viewer for your first-party profiles except that it鈥檚 limited to just partner profiles. Another distinction for partner profiles is that our system applies an automatic time-to-live cadence to this data. Every 25 days, the prospect profiles are deleted from the system to ensure data partners refresh the data regularly. One of the key benefits to managing partner prospect profiles in real-time CDP is to centralize your audience strategy in a single system. Like prospect profiles, prospect audiences are kept separate from your first-party audiences. I鈥檒l select audiences under prospects. Once that loads, the existing audiences display in the browse list. I鈥檒l select create audience in the top right to show you that the experience for creating new audiences here is the same as it is in other locations. The main difference you鈥檒l notice, that is if you create audiences on your first-party profiles, is that these are based solely on the prospect profile and not the XDM customer profile or individual profile. I鈥檒l expand this and then expand the USAAM node. From here, I can drag and drop a partner attribute to the builder canvas and define it. I鈥檒l cancel out of here since I鈥檓 not creating a new partner audience. Last, we can activate the audience. I鈥檒l select a pre-configured S3 account that鈥檚 used explicitly for partner audience activation. Next, I鈥檒l select activate audiences from the three-picker menu next to it. On the select audiences step in the workflow, only prospect audiences display. I鈥檒l also mention that this respects the data usage policy I applied earlier, third-party. The remaining workflow is the same in terms of scheduling and mapping. After that, daily batch jobs will be sent out. I鈥檒l cancel out of this since I won鈥檛 be activating my audience. This concludes the workflow for implementing the offsite prospecting use case in the real-time CDP. Hopefully, you feel confident getting this set up for your organization. Thanks and good luck!