The ultimate guide to Martech | Austin Hay (Reforge, Ramp, Runway)

Marketing technology, or Martech, has evolved from a simple support function into the central nervous system of modern growth organizations. From 2010 to 2020, marketers enjoyed the "golden years" of deterministic matching—where tying an ad view to an app install was precise and relatively simple. Today, privacy changes, cookie blockers, and fragmented user journeys have rendered that playbook obsolete.

To navigate this complexity, companies need more than just tools; they need a disciplined approach to infrastructure, data, and strategy. Austin Hay, Head of Marketing Technology at Ramp and a former leader at Runway, mParticle, and Branch, offers a masterclass in building scalable marketing systems. Whether you are an early-stage founder or leading growth at a unicorn, understanding the architecture behind your growth engine is no longer optional.

Key Takeaways

• Martech is a Product Role: A marketing technologist is essentially a product manager whose product is the internal system and platform, bridging the gap between engineering, marketing, and data.
• The "Golden Stack" Has Shifted: The modern stack leverages a data warehouse (like Snowflake) and Reverse ETL (like HighTouch) rather than relying solely on a monolithic Customer Data Platform (CDP).
• Build and Buy: Avoid the binary "build versus buy" trap. The most effective teams buy core tools to get 90% of the way there and build the final 10% to create a custom competitive advantage.
• Future-Proof Attribution: Start collecting first-touch and last-touch UTM parameters immediately and store them on user profiles. You cannot retrofit this data later when you are ready for Multi-Touch Attribution (MTA).
• Hire for Curiosity: The best Martech candidates are intellectually curious and "engineering scrappy"—capable of reading API documentation and writing basic scripts even if they aren't full-stack engineers.

Defining the Role: What is Martech?

Marketing technology is often misunderstood as simply "managing software contracts." However, in high-growth environments, it is a cross-functional discipline that sits at the intersection of product, growth, engineering, and marketing. It involves people, processes, systems, and platforms.

Austin Hay defines the role not as an IT function, but as a specialized form of product management.

"I think really the way to think about marketing technology is it's a product manager whose specific role and focus is the system or the third party or first party platform."

When to Hire a Martech Lead

In the early stages (under 30 employees), Martech is a "village activity." A growth manager or a founder hacks together tools to get the job done. This approach works until it doesn't. As a company scales past 100 to 150 employees, the "village approach" becomes a liability. Data flows become messy, contracts become inefficient, and technical debt accumulates.

At this inflection point, a dedicated Martech lead becomes essential to:

• Architect Systems: designing how data flows between the warehouse, CRM, and marketing tools.
• Manage Risk: ensuring PII compliance and managing data governance.
• Optimize Cost: preventing tool redundancy and negotiating scalable contracts.

Martech vs. Marketing Operations

It is vital to distinguish between Marketing Technology and Marketing Operations (MarOps). Marketing Technology is an engineering-adjacent role. These professionals understand APIs, data schemas, and backend architecture. Marketing Operations is typically an analyst or systems role focused on campaign execution, email blasts, and SQL queries. While they work closely together, the skill sets differ significantly.

The "Golden Stack": B2C and B2B Architectures

The tooling landscape has shifted dramatically since 2017. Previously, the standard architecture relied heavily on a CDP (like Segment) as the central brain. Today, with the cost of data warehousing dropping, the architecture has decentralized, placing the data warehouse at the center.

The Modern B2C Stack

For consumer businesses, the goal is to ingest high volumes of user events and activate them across ad networks and messaging platforms. Hay recommends the following "Golden Stack":

1. Data Warehouse: Snowflake. This acts as the single source of truth.
2. Ingestion/CDP: Amplitude. While known for analytics, Amplitude can serve as an ingestion point to collect user events and stream them into Snowflake.
3. Activation (Reverse ETL): HighTouch or Census. These tools pipe modeled data out of Snowflake and into your marketing tools, replacing the activation role traditional CDPs used to play.
4. Messaging: Customer.io for early stage, upgrading to Braze as complexity grows.
5. Attribution: AppsFlyer or Branch for mobile attribution.

The B2B Complexity

B2B stacks face a unique challenge: the "B2B2C" problem. You must track individual users (top of funnel) while mapping them to corporate entities (bottom of funnel). This requires reconciling two different object models.

• CRM: Salesforce remains the standard, despite its complexity.
• Marketing Automation: HubSpot is common, though reconciling HubSpot data with Salesforce data is a notorious pain point due to differing data structures.
• The Recommendation: Use a tool like Customer.io or Braze for B2B email marketing to maintain a high bar for user experience, rather than relying solely on clunky sales automation tools.

Solving the Attribution Puzzle

With the death of the cookie and the introduction of Apple’s App Tracking Transparency (ATT), deterministic attribution—knowing exactly which ad drove a specific user—is fading. Marketers must now rely on probabilistic modeling and triangulation.

Designing for Multi-Touch Attribution (MTA)

Many companies debate whether to use first-touch or last-touch attribution. Hay suggests this is the wrong debate. Instead, you must build an infrastructure that supports any model eventually. The most critical mistake companies make is failing to capture the data early on.

The Protocol for Future-Proofing:

• Capture Everything: When a user lands, grab the URL, referrer, UTMs, and ad network parameters (gclid, fbclid).
• Local Storage: Store "First Campaign" and "Last Campaign" attributes directly on the user’s profile in your database.
• Update Logic: If a user returns via a new channel, update the "Last Campaign" attribute while preserving the "First Campaign."
• Event Properties: Fire every pageview event with these parameters attached.

"Most people just don't do the work early on and then when they want to go back later and have MTA results, they don't have the data to do it."

By capturing this raw data from day one, your data science team can retroactively apply Multi-Touch Attribution or Media Mix Modeling (MMM) when the company is mature enough to need it.

Frameworks for Martech Success

Effective marketing technology is less about specific tools and more about decision-making frameworks. Hay utilizes several mental models to navigate the chaos of growing organizations.

1. People, Problem, System (PPS)

When a stakeholder asks for a new tool, the instinct is to jump straight to the "System." Resist this. Work backward:

• Problem: What is the discrete issue we are solving?
• People: Who is involved? Do they need training? Is this a process issue?
• System: Only after defining the problem and the people should you design the system.

2. Build and Buy

The "Build vs. Buy" dichotomy creates false conflict. The savvy approach is "Build and Buy." You buy a best-in-class third-party tool (like an A/B testing platform) to handle the commodity infrastructure. Then, you invest engineering resources to build custom layers on top of it. This allows you to move faster than building from scratch while retaining the flexibility to adapt the tool to your specific business logic.

3. "Think Gray"

In a fast-paced environment, there is immense pressure to make binary decisions quickly. Borrowing from Steven Sample’s The Contrarian’s Guide to Leadership, Hay advocates for "Thinking Gray"—delaying a decision until it is absolutely necessary.

In systems architecture, early decisions can lock you into rigid paths. By waiting until you have more information, you often avoid costly restructuring. It requires patience, but it prevents the "action bias" that leads to bloated tech stacks.

Hiring the Next Martech Leader

Finding talent in this niche is difficult because there is no "Martech degree." The best candidates are often self-taught.

What to Look For

• Intellectual Curiosity: The landscape changes monthly. Candidates must be obsessed with learning.
• Engineering Scrappiness: Look for someone who isn't a software engineer by trade but has taught themselves enough Python or Javascript to solve problems.
• Tool Agnosticism: Beware of candidates who only want to implement tools they have used before. A strong architect looks at the problem first, not their personal comfort zone.

The Interview Question

To assess strategic thinking, ask candidates: "You start tomorrow. By Friday, you need to write a report on what we should change. What is your process?"

A weak candidate will immediately suggest tools they like. A strong candidate will describe a discovery process: interviewing stakeholders, auditing the current data schema, and understanding the business goals before suggesting a single piece of software.

Conclusion

Martech is no longer just about plumbing; it is about strategy. As the "golden years" of easy data vanish, the companies that will win are those that treat their marketing stack as a product. By centralizing data in warehouses, adopting a "build and buy" mentality, and hiring for architectural thinking rather than just tool administration, organizations can build a growth engine that is resilient, scalable, and efficient.

Remember the golden rule: Tools are just meant to solve problems. If you focus on the problem first, the technology will follow.