Powerful AI and data solutions, such as Data Cloud, Marketing Intelligence, and Agentforce, bring unprecedented versatility to marketing. However, introducing these advanced technologies comes with complexity and risk. Implementing a data platform or agentic AI requires careful planning to avoid compromising existing data, workflows, and integrations. Decisions made early on can significantly impact platform usage in the long term. As a result, marketing is increasingly becoming something that needs to be architected.

There is a difference between Well-Architected and, well… architected. Salesforce’s Well-Architected framework remains the ecosystem gold standard for building adaptable, scalable, and usable solutions. Well-Architected is not unique to Salesforce, as it is based on the AWS framework of the same name. Other companies, such as Google and Microsoft, also have their versions. Due to Salesforce’s decision to discontinue its Well-Architected Team, concerns have arisen over the future of Salesforce architecture. 

Well-Architected still has its place in the Salesforce ecosystem, perhaps now more than ever. I think it is invaluable for one domain that has not been consistently architected in the past: marketing. This article explores how Well-Architected can help you optimize your marketing technology stack.

The Fundamentals of Salesforce Well-Architected

What makes Salesforce Well-Architected so powerful is its practical yet comprehensive approach. With constant hurry and exceedingly high demands, Salesforce Architects need a tool that can be applied easily while ensuring that all key areas are covered.

The framework is designed to help you build robust, scalable, and maintainable solutions on the Salesforce Customer 360 Platform. It is based on three core principles: Trusted, Easy, and Adaptable.

Trusted

Marketing cannot work without trust. Solutions must ensure that customer data is handled with care, complying with privacy regulations and maintaining data integrity. This involves implementing security measures, such as encryption and access controls, to protect sensitive information. Marketers can conduct audits to ensure that data is trustworthy and compliant.

Easy

There is beauty in simplicity, and complexity should be avoided whenever possible. Marketers can gain a lot from automation, but only if it is easy to use and maintain. A shared taxonomy, naming convention, and folder structure help streamline the marketer’s workflow. 

Ease of use applies equally to customer-facing assets such as web forms, email templates, and webinars. Plus, these should be designed with accessibility and UX best practices in mind.

Adaptable

The marketing landscape constantly evolves, with new channels, technologies, and consumer behaviors emerging regularly. A Well-Architected marketing solution must be flexible enough to adapt to these changes. 

This involves building modular systems that can be easily updated or extended. Marketers should avoid interdependencies between platforms to make it easier to switch tools. Additionally, planning for contingencies is good risk management; build fall-back solutions for outages.

Implementing Well-Architected Marketing

Marketing hasn’t traditionally been a core domain for IT architects. The rise of marketing operations has contributed to the technologization of marketing, leading to the widespread adoption of data and analytics platforms across the marketing domain. 

As marketing platforms inevitably become more connected internally (to CRM systems and data lakes) and externally (to ad tech and web analytics platforms), architecting them becomes challenging. Instead of inventing a new framework specifically for marketing, let’s look at how Well-Architected can tackle these issues.

Identify Patterns and Anti-Patterns

One key aspect of Well-Architected is the management of patterns and anti-patterns. Patterns are the intentional processes and practices that lead your architecture to its desired state. Meanwhile, anti-patterns, such as ad hoc manual tasks or missing security procedures, are things to avoid. 

Working with patterns and anti-patterns allows marketers to flesh out objectives and areas for improvement. When correctly listed, categorized, and prioritized, they facilitate stakeholder collaboration.

In the marketing domain, customers dictate some patterns and anti-patterns. For example, a pattern might be personalized email campaigns based on customer behavior data. The goal of this would be to boost customer engagement and conversion rates. 

An anti-pattern could be the reliance on generic, one-size-fits-all marketing messages that fail to resonate with most customers. Documenting these helps develop your marketing and flags issues in the underlying processes and capabilities.

Streamline Marketer and Customer Workflows

One of the primary purposes of architecture is to make people’s lives easier. Well-Architected marketing is characterized by the minimum number of clicks and steps required to complete a task. Say a marketer wants to create an email campaign, add members, and activate them as a segment. 

Is the process Well-Architected if they need to do everything manually, shuffling between record pages and apps? Could parts of the process be automated using a Flow, Apex trigger, Prompt Template, or agent?

Consider a similar situation from the customer’s perspective. Marketers are eager to gain customer insights through web forms because they enable more precise segmentation and personalization. The problem is that no one enjoys manual data input, least of all paying customers. 

Marketers can alleviate this pain by inferring information from customer behavior or by connecting to new data sources to gain the same information. When a form is needed, marketers should make the forms interactive and progressive to minimize customer effort at each touchpoint.

Catalog and Prioritize Data Sources

When it comes to data, marketers should follow the KonMari method: if something doesn’t “spark joy,” get rid of it. Of course, it isn’t as simple in real life, where enterprises have varying uses for data, all of which are not visible. 

What an architect can do, however, is to catalog different data sources to form a holistic understanding of the organization’s data. This allows the architect to find redundancies, overlaps, and white spaces. 

Moreover, a data catalog makes it easier to prioritize different data sources. This is especially important when using Salesforce Data Cloud, as the platform allows users to prioritize contact point data for activations. As always, it is best to avoid building a data catalog from scratch and opt for a prebuilt solution instead.

Optimize Data Ingestion and Activation

There is much talk of real-time in marketing. However, it’s important to distinguish between real-time data and real-time marketing. Real-time data means collecting, ingesting, unifying, and transforming data into an actionable form as quickly as possible when a data event has occurred.

Real-time marketing doesn’t necessarily mean marketing actions should be triggered immediately as data enters the system. Instead, it refers to the capability to activate marketing in any situation whenever appropriate, with the most up-to-date data. This may mean waiting five minutes, two hours, a week, or even six months, depending on the use case.

Since marketing is about activating instead of storing data, marketers should be frugal about data ingestion. Are there immediate use cases for a data source? If not, don’t ingest it. What are the latency demands for the data? Does it have to be available in milliseconds or minutes, or is it okay to pull the data in hourly batches? 

Using Data Cloud, you can save significant credits by ingesting data in batches instead of streaming. Similarly, segment runs have a cost associated with them, so it is best to avoid unnecessarily frequent refreshes. 

To summarize, “just in case” is not the approach you want to take with marketing data – opt for “just enough, just in time” instead.

Plan for Risks and Contingencies

Risks and contingencies are sometimes used interchangeably to mean the same thing. However, there is a subtle but important difference. 

Risk management is about preventing bad things from happening. Planning for contingencies is designing a plan B when a situation does occur. Both are essential for business continuity, and marketing is no exception. 

Marketers should seek to identify data, operations, and security risks to avoid being blind sided by them. For the business-critical risks, having a contingency plan in place is key. 

For instance, if marketing is responsible for user onboarding automations, and an API outage prevents data from entering Data Cloud, marketers should have a plan of action to either jump-start the onboarding process manually or trigger it retroactively for the affected accounts.

Final Thoughts

Hopefully, you will be inspired to examine your marketing technology from an architectural perspective. 

However, I want to issue a warning: it is almost as damaging to over-architect a solution as neglecting it altogether. Architecture takes time and effort, and organizations should balance what is considered best practice and what is practical. 

Marketers are more concerned with building campaigns and customer journeys than designing robust and scalable systems. Considering this, I advise using a vertical, rather than horizontal focus when architecting marketing solutions. This means planning for end-to-end use cases, rather than complete layers. 

For example, a marketing architect should enable a customer onboarding experience, rather than designing the complete data ingestion layer before moving on. Marketing architects who think in layers instead of use cases risk becoming like the Greek Sisyphus, doomed to roll a boulder up the hill but never quite reaching the top. 

Ultimately, it’s better to finish something imperfect rather than seek perfection.