When your only tool is a hammer, everything looks like a nail. I’ve been illustrating the point recently by asking whether every system I see is really a Customer Data Platform (CDP). The question comes up because nearly every customer management system builds its own customer database, which is one core function of a CDP. What distinguishes CDPs is that they make their database accessible to other execution systems and add some type of customer management intelligence. This intelligence ranges from behavior flags, segment codes, or predictive model scores to treatment recommendations to full-blown campaign management. Sometimes the enriched data is all that’s exposed to the execution systems, although usually the underlying customer profiles are available as well. Often the CDPs support just one stage of the customer life cycle, such as acquisition or retention: this in itself doesn’t disqualify a system, since I expect that they’ll expand in the future. The other key feature is that CDPs are designed to be run by marketers, not IT staff, even though IT will usually play a role in connecting to company-managed data sources.
I bring all this up partly to clarify that I'm actually being more selective than you might think in deciding what to call a CDP and partly because I’m writing today about Aginity, which refers to itself as a “customer insight appliance” but I think can rightly be classified as a CDP. This in turn matters because CDPs solve a critical problem – marketers’ need for better customer databases – so identifying the widest possible range of CDP vendors increases the chances of each marketer finding a solution that fits her requirements.
On to Aginity itself. Functionally, the system is organized into layers for data loading, database management and analytics, and data consumption, which is exactly the model you’d expect from a CDP. Where it differs from most CDPs is the underlying technology. Aginity runs on a Netezza or similar "massively parallel processing" (MPP) data appliance that would typically run on-premise at the client, rather than being accessed remotely in a “Software as a Service” (SaaS) model.
Of course, most marketers couldn’t care less about this difference. They might care more if Aginity was a tool for IT departments, but in fact marketers can control most Aginity functions beyond the initial connections with source systems, and those connections require IT help even for SaaS systems.
Digging a bit deeper into the technical details (and feel free to skip the rest of this paragraph; it will not be on the final exam), Aginity uses a combination of relational and Hadoop data stores, which lets it add new data sources without formal data modeling. It uses a simple wizard that lets non-technical users add new data elements and expose them on a metadata layer. The system automatically generates scripts to load new data and distribute it appropriately on the data appliance. The system doesn't do the type of "fuzzy matching" needed to associate customer identities across different platforms when no direct link is available; it relies on the client or external partners to make those connections.
Once loaded, the data can be queried directly via SQL, typically using Aginity’s free Query Workbench, which is widely used for MPP databases throughout the industry. Or the data can be published using other Aginity tools that create data marts for external analysis and execution systems. The publishing tools can be run by a marketing analyst, although Aginity says most clients let IT staff use them so IT can enforce quality standards, governance rules, backup management, and similar best practices.
The net result is that Aginity can have a new customer database available to marketing in 90 days or less (often much less), compared with the six to twelve months this typically requires. It’s this speed and flexibility that make me consider Aginity a tool for marketers – and thus a CDP – rather than a tool for IT departments.
Aginity also provides some analytical and customer management features of its own. These include ability to add derived attributes such as lifetime value calculations and segment codes to customer records. These attributes can call on any data gathered by the system, a critical advantage of a CDP. Customer lists can be fed to external systems for direct execution, such as sending an email, or can be loaded into data marts that external systems access with their own segmentation and campaign management tools. Aginity currently provides a range of analysis features including dashboards, profile reports, and segment migration over time. It relies on external systems for advanced analytics such as predictive modeling and plans tighter integration with such systems to allow more precise control over customer treatments.
Aginity was founded in 2006 as a service firm to assemble data for analytics and marketing execution. Its current product, first released in January 2012, is based on tools it developed as a service agency. The company’s clients are concentrated among large retailers but include some ecommerce, manufacturing, and other industries that handle large amounts of customer data.
I bring all this up partly to clarify that I'm actually being more selective than you might think in deciding what to call a CDP and partly because I’m writing today about Aginity, which refers to itself as a “customer insight appliance” but I think can rightly be classified as a CDP. This in turn matters because CDPs solve a critical problem – marketers’ need for better customer databases – so identifying the widest possible range of CDP vendors increases the chances of each marketer finding a solution that fits her requirements.
On to Aginity itself. Functionally, the system is organized into layers for data loading, database management and analytics, and data consumption, which is exactly the model you’d expect from a CDP. Where it differs from most CDPs is the underlying technology. Aginity runs on a Netezza or similar "massively parallel processing" (MPP) data appliance that would typically run on-premise at the client, rather than being accessed remotely in a “Software as a Service” (SaaS) model.
Of course, most marketers couldn’t care less about this difference. They might care more if Aginity was a tool for IT departments, but in fact marketers can control most Aginity functions beyond the initial connections with source systems, and those connections require IT help even for SaaS systems.
Digging a bit deeper into the technical details (and feel free to skip the rest of this paragraph; it will not be on the final exam), Aginity uses a combination of relational and Hadoop data stores, which lets it add new data sources without formal data modeling. It uses a simple wizard that lets non-technical users add new data elements and expose them on a metadata layer. The system automatically generates scripts to load new data and distribute it appropriately on the data appliance. The system doesn't do the type of "fuzzy matching" needed to associate customer identities across different platforms when no direct link is available; it relies on the client or external partners to make those connections.
Once loaded, the data can be queried directly via SQL, typically using Aginity’s free Query Workbench, which is widely used for MPP databases throughout the industry. Or the data can be published using other Aginity tools that create data marts for external analysis and execution systems. The publishing tools can be run by a marketing analyst, although Aginity says most clients let IT staff use them so IT can enforce quality standards, governance rules, backup management, and similar best practices.
The net result is that Aginity can have a new customer database available to marketing in 90 days or less (often much less), compared with the six to twelve months this typically requires. It’s this speed and flexibility that make me consider Aginity a tool for marketers – and thus a CDP – rather than a tool for IT departments.
Aginity also provides some analytical and customer management features of its own. These include ability to add derived attributes such as lifetime value calculations and segment codes to customer records. These attributes can call on any data gathered by the system, a critical advantage of a CDP. Customer lists can be fed to external systems for direct execution, such as sending an email, or can be loaded into data marts that external systems access with their own segmentation and campaign management tools. Aginity currently provides a range of analysis features including dashboards, profile reports, and segment migration over time. It relies on external systems for advanced analytics such as predictive modeling and plans tighter integration with such systems to allow more precise control over customer treatments.
Aginity was founded in 2006 as a service firm to assemble data for analytics and marketing execution. Its current product, first released in January 2012, is based on tools it developed as a service agency. The company’s clients are concentrated among large retailers but include some ecommerce, manufacturing, and other industries that handle large amounts of customer data.
