Do More With Less: Composable Approaches to CDP + Customer Engagement

A woman is sitting at a computer, she's using composable approaches to CDP and customer engagement to do more with her data

Key Takeaways

  • Composable models leverage data cloud and data ingestion foundations that typically already exist to some degree within your organization, and are growing in maturity. Connect with your technology and data leadership to uncover more regarding current state and plans in these areas, and discuss the composable mode with them.
  • Data teams – if you are planning or already have a customer data cloud foundation in place, engage with your marketing team to explore ways they can leverage the investment that you are making. The value of your efforts will be more widely realized across the organization.
  • Marketing teams – engage with your data teams to explore these topics as chances are they already have a customer data cloud foundation in place that is being used by other parts of the business. Beware of 3rd party CDP platform investments, which are significant, as you may already have the “best of breed” foundations already in place within your data & analytics organization.
  • This new approach to B2C MarTech is vastly more agile, secure, and cost effective – resulting in transformational, highly positive impacts on customer engagement and experience. If you want to watch and hear a first hand account, check out this recent talk from the ​​Sr. Director, Global Engagement Marketing & Acquisition at OpenTable.
  • Reach out to Snowflake or to MessageGears or to me directly to talk through these topics and learn more. MessageGears has scores of customers that have adopted the approaches described in this article, and we can guide you on the right approach for your organization. You also can review a recent MessageGears webinar on this topic.

Themes for 2023 and Beyond

Doing more with less has always been popular with boardrooms and executive leadership, especially the CFO’s office. This year is shaping up to be one where this theme will especially ring true across entire organizations; a year likely to be marked by increased financial and budgetary scrutiny, and significant pressure for measurable return on investment.

At the same time, consumer expectations with respect to relevant, timely, data-driven customer experiences have never been higher. Consumer facing brands in particular, across all industry verticals, are competing more than ever for relevance, loyalty, and share of wallet. Brands that get customer engagement right will take market share, and emerge stronger.

Given the above twin priorities of cost management and delivering exceptional experiences, we find that marketing, marketing technology (MarTech), marketing operations, and data & analytics teams are facing a bit of a dilemma. These teams are asking themselves, “how do we work more closely together to deliver exceptional customer experiences while also managing down hard and soft costs, reducing risks, and avoiding or reducing technology sprawl and redundancy?” These conversations inevitably lead to discussions around customer data platform capability models and customer engagement platform capability models.

New Composable, Modular Models for CDP and Customer Engagement Capabilities

To tackle these twin challenges, there is a new, emerging customer data and customer engagement model that leading, data-driven brands have been adopting, and that MessageGears, along with our customers, has been at the forefront of since our inception. This model is often referred to using terms such as “warehouse-native,” “connected application,” “modular,” and “composable;” with “composable” emerging now as the most commonly used.

The composable model was born from the massive evolution that has taken place over the last five plus years in the data warehouse, also now known as the data cloud, space (I will use the term data cloud in the remainder of this article). Snowflake has been the leader in this evolution.

In essence, data clouds like Snowflake are vastly more performant and elastic than legacy alternatives – meaning they can handle massive query volumes and scale up or down without skipping a beat. Back in the days of Teradata, Netezza and the like, this simply was not the case without being cost prohibitive.

In addition, Snowflake leads the market in data sharing and data clean room capabilities, opening up secure, advanced, near-real-time use cases for marketers in the areas of identity resolution, data enrichment, and secure data collaboration with partners.

As a result, a whole new approach to SaaS MarTech has emerged — where a new breed of “warehouse-native” applications either connect to, or feed a centralized data cloud that owned and managed by the brand (or in some cases a trusted advisor, like a Data Consultancy). This simply means that these “connected applications” themselves do not store any of the brand’s data. They either feed the brand’s data cloud with data, or they connect directly to the brand’s data cloud as the primary data source to power their application’s specific use cases. The brand’s data cloud itself is the “hub” for everything, and the brand’s data cloud connects to and powers everything around it.

A true warehouse-native connected application (like MessageGears) is one that does NOT require a brand to map and copy data from the brand’s data cloud environment to a 3rd party SaaS solution provider’s data cloud. All of this data friction is removed. Any SaaS deployment model that requires a brand to map and copy data – even if via setting up data shares – is known as a traditional SaaS managed or packaged application model, which adds hard and soft costs.  

Many practitioners and thought leaders have written or spoken on these new models, terms, and topics, including Omer Singer, Luke Ambrosetti, and Jacques Corby-Tuech, to name a few. I would encourage searching for and reviewing their materials, and following them. You may also contact me directly, and I can point you to other material.

This composable, modular, connected app model, well implemented with cross-functional collaboration, delivers disruptive, revolutionary value that more than delivers on the twin priorities described above. This new approach to B2C MarTech is vastly more agile, secure, and cost effective – which results in transformational, highly positive impacts on customer engagement and experience.

Breaking Down Components of the Composable CDP and Customer Engagement Capability Model

So what does this composable, modular model even look like in practice? Below is a common reference architecture that we leverage at MessageGears when engaging with brands and with partners. This reference architecture is highly informed by actual customer data and customer engagement models that are in place across the brands and partners that we work with.

A common reference architecture that we leverage at MessageGears when engaging with brands and with partners.

While I am not going to go through every pillar of this capability model in great detail, I will comment briefly on specific areas within the model. Note also that the solution provider logos listed in the diagram illustrate “warehouse-first” industry-leading examples. There are often many great options as well as more custom approaches in each area.

Foundational Horizontal Layer – The Brand’s Data Cloud

This is the core of the composable model. All customer, transactional, and behavioral data flows into and lives here, with transformations, data model standardizations and universal identifiers. Chances are your brand’s data team already has (or plans to have) some level of a centralized customer and transactional data cloud in place, as many other parts of the organization demand and leverage this kind of data for planning and analytics purposes, in addition to the marketing team. As a marketing team, it is often simply a matter of asking your CTO or data & analytics leadership for what already exists, what is in the works, and how can the marketing team leverage it. For more on this layer, I would suggest reading Modern Marketing Starts with a Customer 360 on Snowflake written by Lourenco Mello and Luke Ambrosetti at Snowflake.

Pillar #1: Data Collection

Most brands with a million or more customers already have data teams that have put into place the processes for ingesting transactional and behavioral data from point-of-sale, website, and mobile applications using ETL and behavioral data ingestion and transformation tools. Behavioral data may be a gap in some organizations, and there are great best of breed solutions to specifically solve for this from folks like Snowplow and RudderStack, among others.

Pillar #2: Identity, Transformation &Enrichment (includes Snowflake Marketplace)

This area has the potential to be highly transformative relative to traditional ways of solving for identity resolution and data enrichment, and is perhaps the least well known and understood at this point. Best practice approaches that are emerging for identity resolution are for brands to do some of the deterministic identity resolution themselves, natively in their data cloud, using tools like DBT, Coalesce, or Truelty; and then to tap into 3rd party identity graph providers via the Snowflake Marketplace, leveraging data sharing and data clean room capabilities, for the more advanced probabilistic identity resolution use cases. 

As for data enrichment/data append use cases, this can be done in near-real-time, again via the Snowflake Marketplace leveraging data sharing and data clean room capabilities. Brands have many data providers to choose from, and can mix and match depending on the need and cost structures. With these approaches, as with the entire composable model, the brand’s data cloud remains the system of record for all first party customer data.

Pillar #3: Machine Learning, AI & Decisioning

Many brands with a large customer base have data science teams that are already using tooling to perform some level of propensity, predictive, next best offer or next best action models that run against their first party data. The output of these models live in the data cloud. There are also solution providers that accelerate the development of some of these models.

Pillars #4 + #6: Segmentation, Orchestration, Campaign Management + Channel Activation

There are stand-alone solutions, like MessageGears Segment, that deploy as modular, connected apps that run on top of the brand’s data cloud.  These solutions should include robust visual audience builders, segmentation data append capabilities, and the ability to activate audiences to 3rd party channels like Facebook, Google, direct mail, as well as traditional SaaS marketing clouds and customer engagement platforms, etc. Often the 3rd party channel activation component itself is referred to as reverse ETL these days.

Pillar #5: Native Channel Execution (email / push /  in-app / SMS)

Enterprise-class alternatives to traditional SaaS marketing clouds, ESPs, and customer engagement platforms exist today that deploy as warehouse-native connected applications, and are built to render and deliver highly personalized and timely messaging at massive scale. The MessageGears Message product is a leading example that has been in-market since the days of Hadoop, and is leveraged by some of the largest brands in the world.

Crawl / Walk / Run Approaches to Composable Models

You don’t need to have all of the components of this model in place, or have high levels of maturity in each area, in order to generate meaningful business impact and operational cost reductions. Common foundational approaches we see at MessageGears are for brand marketing teams to validate with the technology team that a customer data cloud foundation exists or is in process, and to establish an internal working partnership.

From there, brands can then start down the composable path by bringing in a visual audience segmentation and an audience activation tool, like the MessageGears Segment product, that runs “on top” of the brand’s customer data cloud as a connected app, often replacing Unica or similar 3rd party traditional SaaS audience building or segmentation type solutions. 

This approach enables marketers to create advanced visual audience blueprints and nodes, provide for audience meta-data enrichment, and to then be able to activate audiences to any third-party endpoint, including traditional SaaS customer engagement platforms, also often referred to as marketing clouds. Here is a functional summary of the MessageGears Segment product as a point of reference.

A functional summary of the MessageGears Segment product as a point of reference.

The other components of the composable model as described in this article can be matured and layered in, and other traditional SaaS tooling can be swapped out over time with connected applications to drive improved customer engagement and operational cost reduction.

Finally, I’ll leave you with a quote from the CEO of Snowflake that summarizes the changes in the data cloud space that allow for the new warehouse-native composable approaches to B2C SaaS MarTech discussed in this article.

A quote from the CEO of Snowflake that summarizes the changes in the data cloud space that have enabled the new warehouse-native composable approaches to B2C SaaS MarTech discussed in this article.

About the Author

Walter Rowland

Walter Rowland is the SVP Growth and Partnerships at MessageGears. Walter has worked in Sales, Marketing and Partner leadership roles at high-growth technology firms his entire career. Walter holds an MBA in finance from Columbia Business School, and an undergraduate degree from Harvard.