Dual Zone at it’s core is about optimizing workload management and the roles and responsibilities of teams. It also assumes you are able to reframe CDP as a capability versus a technology, and then agree to split them up into two Zones:
Zone 1 is typically a Modern Data Stack architecture built on a Cloud Technology and focuses on Identity Resolution and Data Management.
Zone 2 is typically a packaged CDP software solution that is a licensed SaaS solution that focuses on Journey Orchestration and Campaign Activations.
Often when I have an initial conversation on Dual Zone, the first reaction I get is “Oh so it’s basically a hybrid approach”. I try to politely nod but in my head I’m screaming “BUT IT’S SO MUCH MORE THAN HYBRID!”
Why Dual Zone is so much more than Hybrid is that when consultants propose a “hybrid approach” that is the epitome of classic consultant-speak. It says something and nothing all in the same breadth. If you have a spectrum of options to choose from, “hybrid” can mean an infinite number of things that sit along the line in-between those two fixed end points. That’s not what Dual Zone is at all.
Dual Zone is a specific approach and a philosophy on how to optimize the systems and the teams involved in standing up a Customer Data Platform (CDP) capability. And to be even more clear, it’s not simply optimizing CDP technology but how your organization uses data and insights to deliver orchestrated experiences across channels and customer touchpoints.
With that, here are some guiding pillars of Dual Zone:
KEY VALUE DRIVERS OF DUAL ZONE
Optimize Cost — how do you optimize for total cost of ownership
Increase Efficiency — how do you execute with greater speed and fewer resources
Improve Scalability — how do you grow the capability across organization from a process, people, and tech perspective
Achieve Sustainability — how do you find balance and repeatability as you scale
KEY OPERATING PRINCIPLES OF DUAL ZONE
Think about an Experience Op Model, NOT CDP Op Model— the emergence of Data and AI has forced IT and Business to work together to achieve success, propelling a converge of teams. Borrowing Data Mesh / Data Fabric concepts, teams who had multiple degrees of separation now need to come together and be in lockstep so that there are quality data products, quality data science models, and quality customer experiences being delivered. That will 100% require change in how we organize. The good news is this is a not a completely new concept! Front-end Designers and Back-end Developers went through this same process about a decade ago when Creative teams would design front-end experiences extremely difficult to implement. It then led to the idea of multi-disciplinary teams (SCRUMs) who can work together in rapid fashion to communicate quickly to solve problems. Now is the time to introduce Data teams, Data Science teams, and InfraOps teams into the mix. The path is already set, don’t re-invent the wheel!
Avoid Data Duplication — Do not simply copy data from Zone 1 into Zone 2. That defeats the purpose of a Dual Zone approach. Raw data should primarily sit in Zone 1 and Zone 2 should only be loading data required to enable use cases which means the data loaded should be pre-processed in Zone 1 and aggregated for consumption by Zone 2. Do that and you will have fewer headaches about overspending on storage and synching issues.
Assess the True Cost of Ownership— High compute activities should happen in Zone 1. Though creating calculated attributes and insights as well as AI/Model can be done in Zone 2 (depending on the CDP software) — it shouldn’t. Taking a FinOps mindset to analyze TCO you will find much more value offloading as much of those activities as possible to Zone 1 which typically also has very transparent pricing.
Design Identity Resolution in Two Stages — Customer identity should be established in two stages. Zone 1 should establish the customer core identity using identifiers that are true proxies for identity (e.g. Name, Address, etc) and Zone 2 should build off that identity. Zone 2 should establish the customer enhanced identity that also serves as the “Actionable Customer Profile” using identifiers that can potentially over/under collapse profiles (e.g. cookies, device IDs) and therefore it must be used cautiously and evaluated on company by company basis based on their risk threshold of being wrong.
Segment is a Team Sport — Segmentation design and development requires coordination between both Zones. Zone 1 should establish the the supporting profile attributes necessary for business to create segments. And Zone 2 should be where governance is established on a common segmentation strategy on what segments get created and who can activate against them against the organization. On the flip side, Exploratory Data Analysis and Segmentation Analysis can be performed in Zone 2 but you will have a richer data set to analyze from in Zone 1 which is where it that analysis should be performed.
Let the Data Fly Wheel be a Flywheel — Dual Zone may seem like a very linear, left to right concept and might suggest everything has to flow through Zone 1 before it flows to Zone 2. That is 100% not the case. Both Zone 1 and Zone 2 can ingest data. Zone 2 should prioritize ingesting real-time, evenstream data that is required for real-time personalization to avoid the latency penalty that is required to go through Zone 1. Regardless of whether a data source flows into Zone 1 or Zone 2, it’s a closed loop system which will feed each other. So at the end of the day the decision of where a data source should land first should totally be dictated by the use case.
Build Reports Where The Data Lives — Zone 2 has their own native reporting capabilities and should use that for journey analytics and whatever OOTB reporting it provides. However, avoid extending beyond that because heavy BI reporting should happen in Zone 1 since it provides much more flexibility and access to granular data sets (see principle #1) which you can then roll up as needed.
Cleanrooms Can Live In Multiple Places (for now) — Enterprise cleanrooms should logically sit in Zone 1. Advertising cleanrooms today can potentially sit in Zone 1 as well (depending on the cloud you are on) but can also be treated as another form of adtech that Zone 2 integrates with. In the future, as DMP finally goes away (though it feels like a Jason movie at this point where it won’t quite completely die yet) my belief is that CDP vendors in Zone 2 will establish cleanroom capabilities and merge that “new adtech” directly into the CDP stack just because it makes too much sense to do it.
In full transparency, these value drivers and principles in my Dual Zone Manifesto might change over time as things evolve in the marketplace but I will update them as needed.
Comments