Vista is embracing data mesh to work with and manage data (see this earlier blog post). A critical part of a successful mesh is interoperability, which ensures that data living in different domains and teams can interact. Lego sets are a good analogy – individual pieces might differ greatly, but there is a universal way they fit together, enabling a nearly infinite number of outputs. In data this “connectivity” of products and datasets can take many forms, from naming conventions to APIs, and implementation is ultimately a balancing act between standardization and flexibility.
Interoperability value proposition at Vista
Finding that balance is important since interoperable development can create efficiencies and drive more relevant and connected customer experiences. Imagine if orders, site sessions, and content all tagged products differently, there would need to be incremental work to stitch each method together for customer journey insight. Vista has a standard and interoperable product hierarchy that is beneficial for building cross-domain products, common understanding, and decision-maker confidence. However, this hierarchy product still needs the flexibility to account for Manufacturing having a legitimately different way of grouping products than what is needed for customer experience analysis and modeling.
Aside from product data, the dimension of channel pops up in site data, financial data, attribution methods, marketing performance, and enterprise reporting. Developing a universal hierarchy benefits each team owning these areas individually, as they aren’t burdened with managing something ad hoc themselves. Beyond that, it makes the lives of stakeholders using products from different domains and teams much easier, as channel concepts stay standard. Ultimately and critically, the alignment of this terminology across domains speeds up algorithmic personalization efforts, which must leverage a variety of data sources containing channel information, to assess customer channel preferences.
Building the Channel Mapping product
To create standard channel naming and categories, the Channel Platform Enablement (CPE) team – containing analysts and data engineers – transformed data across multiple platforms and services. Maintaining and optimizing the “Channel Mapping” product also requires cross-functional collaboration with software engineers, other data product teams, and channel specialists. Upstream, CPE coordinates priorities with Technology on increased automation in channel and campaign set up which in turn stabilizes metadata capture about marketing placements, including channel names, targeted funnel stage (capture/convert demand), format (30sec TV spot, print marketing brochure, welcome email, etc.), country, and language.
CPE also manages this channel information historically, aiming to deliver on the highest priority sources to increase the potential of the product and serve more use cases. By mapping all these channel data to a standard taxonomy, more comparisons are readily available (ex. year over year) and downstream products can use a variety of keys to join onto the product to pull current naming. The data product is not perfect, but the aim is continuous improvement, which reflects how taxonomies can also shift and change with business goals.
Interoperability should be intentional and top of mind in development. Putting in this effort upfront makes it easier for different products to effectively and scalably communicate in the same language without incurring avoidable misalignments, rework, or proliferation of nonstandard stitching data layers. Misalignments are a distraction at best and misleading at worst and navigating them is not where analysts or users should be spending their time. Instead, focusing on common and accessible language and investing in interoperable ways of working can increase confidence in decisions, reduce the time and effort needed to come to shared meaning, and expand ideas for customer benefit from single to multi-domain thinking.