Publicis Sapient

CUSTOMER DATA PLATFORMS

How do you manage anonymous profiles and distinguish records that cannot be associated with any identifiable individual?
Do you offer tagging services that track customer behaviors across digital properties (e.g., website, apps, etc.)?
Please provide a description of your APIs, event message/logs, and data stores which are currently accessed by external applications.
Please include an overview of any data models that you use today and how well these models are understood by business users.
Do you allow your business users to create derived or calculated attributes from your customer data?
Do you have the ability to extract customer information in bulk? If yes, is the data in a SQL or NoSQL database?

Please describe any current identity resolution and profile stitching.
Are you identifying customers across multiple devices? How? Are you using an identity vendor?
Can you retroactively merge historical data when an identifier is found after data has been loaded (e.g., a customer ID is associated with a previously anonymous cookie)?
Are there any limitations on your sources of identifiers or attributes that are being used for stitching?
Can you integrate with third-party solutions to help with identity resolution? If so, which ones? If not, which ones are priority to build connectors to?
Do you have a persistent customer ID or Golden Record? If you have multiple, what are their uses?

What out-of-the-box reports are available to your primary data consumers?
Describe your data warehouse and BI current state.
Do you have a sandbox environment with access to production raw and processed data for data analysis? Can you refresh this data without heavily involving IT?
Do you include tools/reports to automatically monitor system operations and performance? If so, please explain/describe these capabilities.
How do you determine statistical significance or test success? Are you using A/B testing systems?
Do you have an established process to productionize models and reports developed by your data scientists/analysts?
What are you using to track performance of marketing campaigns (e.g., personalization campaigns)? How is campaign attribution calculated?

Do you currently create custom audiences? If yes, please describe how this works, especially what aspects are business-user driven.
How is segment membership calculated for newly created segments? Is historical data included?
Do you allow a business user to obtain summary counts of audience segments? Can segments be compared against each other without having to heavily involve IT?
How frequently are your segments updated?
Can you pass audience segments to external platforms like digital media / ESPs / advertising platforms?
Does your current stack support predictive segmentation?
Does your current stack use a single segmentation system across campaigns (that is, the same structured segments across multiple channels e.g., web, email, mobile app, etc.)?

Please explain the scope of your team’s machine-learning capabilities.
What algorithms (generally) are your teams currently able to create?
Do you have an established process for taking the model to production? What is the process?
How frequently are machine-learning models updated? Can the algorithms respond to a customer’s live behavior?
What use cases is machine learning used for today? (e.g., recommendations for content, products, offers, banners, etc.)
Can you restrict or suppress the same recommendations from being served repeatedly to an active customer? E.g., a person who visits your site/uses your app three times in a week.

Describe how you currently personalize experiences across the following channels (if an additional system of activation is required, please note this):
- Mobile App
- Clienteling
- Mobile messaging/SMS
- Email
- Social
- Website
- Onsite Search
- Ad/Cookie Tracking Networks
- Call Center
- Chat
- Loyalty
- On location/store, etc.
Describe how you support multi-step campaigns, including: identifying the audience, leveraging business rules to build an audience, timing/triggering events, creating the experiences, pushing the data to other vendors to deploy, or deploying natively from you, etc.
Please list any control or oversight mechanisms you are using to suppress oversaturation (e.g., of a particular campaign) at the customer level.
Do you currently test campaigns before pushing them live?
In what ways can the customer profile be exposed to call center or physical business locations? Describe how the data can be exposed in web-based applications or other methods of exposure.

Please explain how you host your data platform (i.e., on premise, vendor hosted, cloud-based/SaaS). What is your technology stack?
How is deployment of your internal solutions handled (that is, centrally by IT or by business groups, or intermediaries such as a “Digital Center of Excellence” etc.)
How do business users access customer/marketing data today?
What is the current number of people at your company accessing customer/marketing data?
Are there permissioning levels that restrict certain business users from performing particular tasks or accessing certain data? If so, please describe the levels.
What skills are typically required by which types of data consumers (e.g., marketers, analysts, developers, executives, etc.)?
What is your customer database uptime, historically?
What is the typical time it takes for a data consumer to be granted a request for hashed or encrypted sensitive data for insights/analysis?

Please describe the support SLAs you have with your based Customer Data systems.
What training materials or user resources are available from your SaaS based Customer Data systems?
Describe the audit trail for user access.

Describe your approach to encryption for data in storage and in transit.
What best practices do you follow for security protocols around data storage? A list is fine.
What type of vulnerability assessments have been conducted on your customer data?
Do you support Single Sign-On (SSO) for your employees' tools?
What type of security is used for your employees' tools? Do you require any data security certifications? If so, which ones?
Which teams have access to PII?
Please explain internal processes for handling General Data Protection Regulation (GDPR) and California Consumer Privacy Act (CCPA) compliance and features intended to assist with Personally Identifiable Information (PII). Can your customer data systems support these processes?
Do you require your vendors to avoid co-mingling of customer data between clients? Are you participating in any data co-ops or 2nd Party data partnerships?

RAVINDRA NARLA
Data Engineering and Data Platforms
ravindra.narla@publicissapient.com
MAX KIRBY
Digital Identity and Data Privacy
max.kirby@publicissapient.com

publicis sapient

CDP Readiness Assessment | publicissapient.com