Register for "Diverse, Integrated, and Real Time: Delivering the Right Data for AI/ML Success" - Tuesday, June 18,1:00 pm EST

Evaluating Cloud Connectors for the Customer 360 Data Program

ABSTRACT: This blog, the second in a series, recommends criteria for enterprises to evaluate cloud connectors and ensure the benefits outweigh the costs.

It might be time to reconsider how to architect the customer 360 data program.

Rather than relying exclusively on a monolithic customer data platform, some enterprises now make the cloud data warehouse their primary source of customer truth. Cloud data warehouses such as Snowflake and Google BigQuery offer a flexible platform to integrate fast-rising volumes of multi-structured customer data right alongside data for related business functions. The customer data platform still provides logic and machine learning to profile customers, resolve identities, and predict their next move.

This shift requires a new architectural element: cloud connectors such as Lytics Cloud Connect, which delivers customer data straight from the data warehouse to touchpoints such as digital ad platforms. This blog, the second in a series, recommends criteria for enterprises to evaluate cloud connectors and ensure the benefits outweigh the costs. The first blog in the series explored digital advertising as a case study for cloud connectors. The third and final blog will recommend guiding principles for success with this new architectural approach.

First, a quick recap. A customer 360 data program curates rich profiles of customers—i.e., who they are, and what, why, and how they buy. These profiles enable enterprises to make sense of target buyers. They win their business and loyalty based on intimate knowledge of requirements, influences, and behavior. Here are the five stages of the C360 program, and the ways that they rely on various architectural elements.

  • Integrate. The data engineer and marketing operations manager use the data warehouse to integrate multi-structured customer data from various sources into merged files with common formats.
  • Profile. The marketing operations manager uses the customer data platform to resolve customer identities and profile them based on their digital activities. They store the profiles in the data warehouse.
  • Activate. The marketing operations manager and/or DevOps manager uses the cloud connector to activate profiles by feeding them to targets such as digital advertising platforms.
  • Synchronize. Data engineers, marketing operations managers, and DevOps managers continuously synchronize updates from targets back into their profiles in the data warehouse.
  • Analyze. Marketing operations managers and data analysts use business intelligence (BI) tools to study customers and track campaign performance. They might use ML models to assist their analysis.

Stages of the Customer 360 Data Program

These steps vary in some cases. For example, Lytics Cloud Connect can serve as an activation layer on top of the data warehouse, without the need for a customer data platform to create profiles. In this scenario it queries sales records within the data warehouse, for example to find and upsell customers that purchased a given product recently.

Weighing your options

Costs. This architectural approach has costs. Data engineers need to manage the pipelines that move data between the customer data platform and data warehouse. MarketingOps managers need to implement cloud connectors, then work with data engineers to synchronize updates across sources, targets, cloud connectors, and the data warehouse. Tasks like these require time and training.

Benefits. Enterprises make the effort because the benefits can outweigh the costs. By integrating higher volumes and varieties of data, they can build more accurate customer profiles. By activating customer profiles on more targets—in less time—they can engage more prospects for a given amount of effort. These benefits help increase lead generation, lead quality, and revenue. In addition, by curating customer profiles alongside other functional data in the data warehouse, marketing and analytics teams can uncover new opportunities to make the business more customer-driven.

Evaluation criteria for cloud connectors

To ensure the benefits outweigh the costs, enterprises need cloud connectors that are both efficient and effective. Marketing and data leaders should evaluate cloud connectors according to four criteria: ease of use, ecosystem support, compliance capabilities, and performance and scalability.

Ease of use

  • Cloud connectors should not require significant time or effort to administer. Ask the following questions to evaluate their ease of use.
  • What skills and training do MarketingOps managers and DevOps engineers need to become productive with these connectors? What level of SQL scripting knowledge do they require?
  • What level of automation do the connectors provide to guide implementation, configuration, task execution, and monitoring?
  • Do they provide operational notifications—for example, regarding system errors, utilization thresholds, or task completion status?
  • Do they synchronize target updates—for example, click-throughs and conversions for digital ads—back to profiles contained within the data warehouse?
  • How much time and effort is required to implement changes—for example, to add/remove an audience or target, or adjust the frequency of updates?

Ecosystem support

  • Cloud connectors should integrate with an ecosytem of popular targets, data warehouses, customer data platforms, and file formats. Ask the following questions to evaluate their ecosystem support.
  • Do the connectors integrate with common targets? This includes marketing automation platforms such as Hubspot and Adobe Marketo Engage, digital ad platforms such as Google Ads and Facebook, and customer-relationship management platforms such as Salesforce.
  • Do they integrate with targets such as Snowflake Data Cloud, Google BigQuery, Azure SQL database, and Amazon Redshift?
  • Do they integrate with customer data platforms such as Lytics Decision Engine and Segment?
  • What level of customization is required to integrate with homegrown elements and tools?
  • Do they support customer profiles in file formats such as Excel, CSV, Apache Avro, JSON, and Apache Parquet?

Compliance capabilities

Cloud connectors should help your team comply with legislation such as the General Data Protection Regulation (GDPR) in Europe and the California Consumer Privacy Act (CCPA) in the United States by controlling the usage of personally identifiable information (PII). Ask the following questions to evaluate how cloud connectors affect your compliance with regulations like these.

  • Do the cloud connectors offer role-based access controls to ensure only authenticated users perform only authorized tasks?
  • Can users “mask” or obfuscate columns containing PII such as names or social security numbers before they activate profiles and send them to a target?
  • Do the cloud connectors log events and user actions for auditing purposes?

Performance and scalability

Cloud connectors should deliver customer profiles to targets and synchronize updates at a speed and scale that meet business requirements. Ask the following questions to evaluate their performance and scalability.

  • Do the connectors meet performance requirements such as latency and throughput of data delivery? Do they support a sufficient number of concurrent targets and users?
  • Do they automatically scale to support workload spikes? For example, they might add compute resources or storage capacity to support rising numbers of profiles or ad responses.
  • What infrastructure resources—storage, compute, network connections, etc.—do the connectors require to operate on premises or in the cloud?
  • Can the connectors operate within configurable ranges—for example, with resource throttling and threshold-based alerting—to avoid cost over-runs for compute consumption?

So, has the time come for your enterprise to re-architect its customer 360 data program? You can answer this question by evaluating cloud connectors according to their ease of use, ecosystem support, compliance capabilities, and performance and scalability. If the answer is promising, look to establish a beachhead of success and expand from there.

Kevin Petrie

Kevin is the VP of Research at BARC US, where he writes and speaks about the intersection of AI, analytics, and data management. For nearly three decades Kevin has deciphered...

More About Kevin Petrie