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

Aug 13, 2023 Episode Page ↗
Overview

Austin Hay, Head of Marketing Technology at Ramp, delves into MarTech's role in growth, distinguishing it from growth PMs and advising on when to hire a MarTech specialist. He shares insights on team structures, essential tools, and strategies for future-proofing attribution in a complex data landscape.

At a Glance
67 Insights
1h 24m Duration
15 Topics
6 Concepts

Deep Dive Analysis

Defining Marketing Technology (MarTech) and its Role

Distinguishing MarTech from General Growth Roles

Signs Your Company Needs a MarTech Person

Organizational Placement of MarTech Roles in B2B vs. B2C

A Day in the Life of a MarTech Professional

Marketing Technology vs. Marketing Operations

Evolution of B2C Marketing Technology Stacks

Understanding Reverse ETLs and their Role

Evolution of B2B Marketing Technology Stacks

Challenges and Best Practices for Attribution

Emerging Tools and Probabilistic Data in Marketing

Hiring a MarTech Professional: What to Look For

Useful Frameworks for Work and Life

Lightning Round: Books, TV, Products, and Life Motto

Austin's Recommended Golden MarTech Stack

Marketing Technology (MarTech)

MarTech is an amorphous, cross-functional discipline at the intersection of product, growth, engineering, and marketing. It involves managing first-party homegrown solutions and third-party tools, focusing on people, process, systems, and platforms, similar to a product manager for internal systems.

Marketing Operations (MarOps)

MarOps typically refers to a function that is not always technical, often involving a systems or business analyst. This role focuses on setting up campaigns, sending email blasts, debugging, and performing analytics work like SQL queries, which are semi-technical but not engineering-based.

Reverse ETL

Reverse ETL is a capability that allows data to be moved from a data warehouse (like Snowflake) to various marketing and advertising tools. It enables businesses to activate their modeled data stored in the warehouse for different purposes, essentially building a custom CDP functionality.

Probabilistic Matching/Attribution

This is a method of making marketing decisions with incomplete or aggregate data, especially in the context of increased privacy restrictions (like iOS 14). Instead of deterministic, one-to-one identification, it involves creating models for a percentage of the population and extrapolating those findings to the whole.

Build and Buy Framework

This framework suggests that when considering technology solutions, companies should not limit themselves to an 'either/or' build versus buy decision. Instead, they should explore how to combine buying a third-party tool (to get 90% of the way there) with building custom solutions on top (for the remaining 10%) to achieve optimal outcomes.

Thinking Gray

A decision-making philosophy from Stephen B. Sample's 'The Contrarian's Guide to Leadership,' which advises against making quick, black-or-white decisions. Instead, it encourages delaying a decision for as long as possible until it absolutely must be made, allowing for more information and potentially a better outcome.

?
What exactly is Marketing Technology (MarTech)?

MarTech is a cross-functional discipline that manages first-party and third-party platforms and systems, focusing on people and processes. A MarTech professional acts like a product manager for these internal systems, ensuring they drive growth and efficiency.

?
What is the difference between a growth role and a MarTech role?

At small startups, growth professionals often handle MarTech tasks. As a company scales (around 100-150 people), MarTech becomes a specialized, centrally owned function responsible for data flow, system schema, procurement, and legal aspects of tools, whereas growth focuses on user acquisition and optimization.

?
What are the signs a company needs to hire a MarTech person?

Signs include existing systems becoming too complex or overwhelming for one person, an inability to move forward with business plans due to tool changes, or a lack of understanding of how data flows through tools, leading to inefficiency and potential liability.

?
Where should a MarTech person report in an organization?

For B2C, MarTech often reports into a CMO or Head of Growth, serving the growth team. For B2B, it's messier, potentially living in RevOps, IT, product ops, or even engineering, depending on the business model (pure B2B vs. B2B2C) and the technical leadership of the function.

?
What is the difference between Marketing Technology and Marketing Operations?

Marketing Technology is typically an engineering-based role, often performed by someone with an engineering background who designs and builds systems. Marketing Operations is usually less technical, focusing on campaign setup, email blasts, debugging, and analytics, often performed by systems or business analysts.

?
How has the B2C MarTech stack evolved?

Initially, from 2016-2017, the stack centered around CDPs (like Segment or mParticle) to collect and distribute user data. Post-2020, with cheaper warehousing, the trend shifted to storing and modeling data in a warehouse (e.g., Snowflake) and using Reverse ETLs (e.g., Hightouch, Census) to activate that data in various tools, allowing for more custom CDP functionality.

?
What is the best way to approach attribution in the current privacy landscape?

The best practice is to design your infrastructure from day one to support multi-touch attribution (MTA), even if you start with first or last touch. This involves collecting all relevant URL parameters (referrer, UTMs, ad network IDs) and storing them as both user attributes and events, including first and last campaign information, to enable future modeling.

?
What are the key challenges for attribution in mobile apps and websites today?

For mobile apps, iOS 14 (ATT) has made deterministic one-to-one matching difficult, pushing towards aggregate and probabilistic data. For websites, browsers stripping URL parameters and cookie blockers mean a higher percentage of paid traffic is incorrectly attributed to organic, as initial user journey data is lost.

?
What qualities should one look for when hiring a MarTech professional?

Key qualities include intellectual curiosity, scrappy engineering skills (JavaScript, Python, API requests), and strong cross-functional team-playing abilities. They should be able to manage upward, laterally, and downward, acting as a 'quarterback' to persuade different departments.

1. Adopt “Build and Buy” Strategy

Instead of a rigid “build versus buy” approach, embrace a “build and buy” strategy: purchase third-party tools for core functionality, then build custom solutions on top to achieve higher velocity and unique capabilities.

2. Adopt “Tools Solve Problems” Mindset

Consistently remind yourself and your team that tools exist solely to solve problems, fostering a problem-first approach rather than a tool-first mentality in MarTech strategy.

3. Use PPS Framework

When facing a challenge, apply the Problem, People, System (PPS) framework: first identify the ‘Problem,’ then the ‘People’ involved, and finally, what ‘System’ it impacts, to avoid prematurely jumping to tool-based solutions.

4. Practice “Thinking Gray”

Adopt the “thinking gray” philosophy by intentionally delaying decisions for as long as possible until absolutely necessary, fostering patience and allowing for better, more informed choices in both work and life.

5. Develop Probabilistic Attribution Skills

Familiarize yourself with probabilistic matching and attribution as a crucial skill set, as marketers increasingly rely on models to extrapolate insights from partial data (e.g., 30% of a population to 100%).

6. Future-Proof Tool Setup Minimally

When setting up or picking tools, always consider their long-term implications (1 year out) and take minimally invasive actions to mitigate future catastrophic risks, balancing future-proofing with immediate product market fit needs.

7. Prioritize Cross-Functional Management

When hiring MarTech professionals, prioritize candidates with strong upward, lateral, and downward management skills, as they will need to persuade and coordinate across multiple departments (RevOps, Product, Data) to achieve goals.

8. Set Up MTA Infrastructure Early

Prioritize setting up the correct data infrastructure for multi-touch attribution (MTA) from the beginning, as it’s not overly complicated and prevents future data gaps when you need MTA results.

9. Scrutinize SaaS Contracts for Scale

Pay close attention to SaaS tool contracts, terms, liability exposure, and future scaling costs, as managing these effectively can prevent significant financial losses and provide high leverage.

10. Centralize MarTech Ownership

As your company scales past 100-150 people, transition from a “village approach” where everyone manages tools to a centrally owned MarTech function to ensure data flow, schema, and legal compliance are properly managed.

11. Design Future-Proof MarTech Systems

Actively design and advocate for a 1-2 year vision for your MarTech system, considering financials, contracts, and operational efficiency to ensure long-term scalability and cost control.

12. Automate MarTech Administrative Tasks

Automate administrative tasks such as PII requests, tool admittance, and permissions to free up MarTech time and prevent issues like accidental mass emails from interns.

13. Prioritize MarTech Cost Efficiency

Set internal goals to continuously reduce the total cost of tooling per user or per seat as the business grows, striving for lean and efficient operations.

14. Implement First & Last Touch UTM

Store UTM parameters and ad network IDs (e.g., GCLID) locally on the device as “UTM first campaign” and “UTM last campaign,” updating the “last” value with each new visit to enable multi-touch attribution.

15. Design Data for Multi-Touch Attribution

Collect user attributes for both first and last attribution, and fire page view events with first and last UTMs, enabling your data warehouse team to coalesce all touchpoints for comprehensive MTA analysis.

16. Align Taxonomy with Tool Capabilities

When designing your data taxonomy, consider the limitations and capabilities of your third-party tools, as many only support user and event objects, influencing how you can structure and store data.

17. Control Schema with Data Warehouse

Build your own data warehouse for unlimited schema control (product, user, event), as most third-party B2C tools (except Snowplow) limit you to predefined user and event objects.

18. Consider Tool Object Orientation

Design your website and app tracking with the object orientation (user, event) of your third-party tools in mind, as this impacts how data can be collected and utilized.

19. Design Data Movement Process

Establish a clear methodology or process for how and when data should be moved between different tools and your data warehouse to avoid haphazard data management and system issues.

20. Optimize Ads with Full-Funnel Data

Enhance ad optimization by tying top-of-funnel data to bottom-of-funnel opportunity data (e.g., ideal value of an opportunity) and sending these “synthetic events” back to ad networks, rather than just optimizing for clicks.

21. Invest in Early Security Measures

Make small, early investments in security measures like SSO (e.g., $2,000 and two days of setup) to prevent larger security problems and future IT overhead.

22. Implement Safeguards Against Duplication

Establish safeguards to control SaaS tool sprawl, ensuring teams have necessary tools without introducing duplicative technology, which helps manage costs and simplifies the tech stack.

23. Hire for Intellectual Curiosity

Prioritize intellectual curiosity when hiring MarTech professionals, as the field constantly evolves, requiring a strong willingness and ability to learn new tools and technologies quickly.

24. Seek Scrappy Engineering Skills

Look for candidates with scrappy engineering skills (e.g., JavaScript, Python, API literacy) who can understand and solve technical problems, even if they aren’t full-stack engineers.

25. Hire Tool-Agnostic Problem Solvers

Prioritize hiring MarTech professionals who are tool-agnostic and view tools as solutions to problems, rather than defaulting to previously used tools.

26. Ask “MarTech System Audit”

Ask candidates to outline their approach to auditing a new MarTech system and proposing changes within a week to identify tool-agnostic problem-solvers who think backward from the problem.

27. Ask “How Did You Prepare?”

Use the interview question “How did you prepare?” to assess a candidate’s thinking process, planning ability, seriousness, and intellectual curiosity.

28. Self-Teach Core MarTech Skills

Acquire essential MarTech engineering skills (web/backend programming) through self-teaching or online academies, as a 6-month investment can make you “pretty dangerous” in the field without needing a formal degree.

29. MarTech Leaders Must Be ICs

Recognize that all marketing technologists, including leaders, must maintain individual contributor (IC) skills and be the most senior technical experts on all first- and third-party systems.

30. Place B2C MarTech within Growth

For B2C companies, position the MarTech function alongside or reporting into the Head of Growth or CMO, as its primary customer is the marketing/growth team.

31. Hire Technically Strong MarTech

Ensure your MarTech lead is a strong technical architect or operator to effectively represent the function to product and engineering organizations.

32. Avoid Decentralized MarTech

For most startups, avoid a decentralized MarTech model (e.g., a systems person in every org) as it can create more systems and spread resources too thin, unless you are a massive company.

33. Align MarTech Placement by Leader

Determine MarTech’s organizational placement (e.g., ops, product, engineering) based on the leader’s primary skills; an ops-focused leader might fit in a decentralized ops role, while a technical architect/PM leader might lead a standalone team.

34. MarTech Supports UA Campaigns

MarTech’s role is to support user acquisition teams by providing the technological infrastructure and expertise to optimize campaigns, improve CAC, and achieve other growth metrics.

35. Empower Growth Engineers for UA

Recognize that growth engineers with strong technical skills can be “superhuman” in user acquisition, setting up tools and running campaigns end-to-end without needing a dedicated MarTech person.

36. Distinguish MarTech from MarOps

Understand that MarTech typically involves an engineering background, focusing on technical implementation, while MarOps often involves systems/business analysis without necessarily requiring an engineering background.

37. Align MarTech Goals with Growth

Tie MarTech goals directly to the growth and user acquisition teams’ objectives (e.g., CAC goals) to ensure alignment and support their success.

38. Develop Persuasion Skills

Cultivate strong persuasion and salesmanship skills to effectively communicate the importance of MarTech problems and secure necessary resources and buy-in from other teams.

39. Plan Email Marketing Migration

Anticipate and plan for major email marketing tool migrations (e.g., from a small tool to Braze or Marketo) as the company scales, aiming for a safe transition without losing revenue over a 6-month initiative.

40. Utilize CDPs for B2C Data

For B2C businesses, leverage Customer Data Platforms (CDPs) like Segment or MParticle to collect user data, tie it to actions, and distribute it to ad networks, email tools, and product analytics.

41. Centralize Data in Warehouse

Adopt a data architecture where all data is stored and modeled centrally in a data warehouse (e.g., Snowflake), then use reverse ETL tools to activate and move that data to various platforms.

42. Leverage Reverse ETL for Stacks

For companies with an advanced engineering culture, existing data warehouses (like Snowflake), and modeling capabilities (like DBT), prioritize using reverse ETLs to activate data from the warehouse.

43. Center B2B MarTech on Salesforce

For B2B businesses, design your MarTech stack with Salesforce at its core, as it typically serves as the central hub for lead capture, CRM, and other related tools.

44. Strategize B2B2C User-Company Mapping

For B2B2C models, carefully plan how and when to tie individual user data to company or entity objects, considering tool placement and data flow to manage top-of-funnel complexity with CRM needs.

45. Carefully Map HubSpot to Salesforce

If using both HubSpot and Salesforce, meticulously plan and implement how data is mapped between them to avoid significant data management challenges, as there’s no single “good solution” for this integration.

46. Choose B2B2C Data Flow Strategy

Select a clear data flow strategy for B2B2C, such as collecting user/event data in Amplitude/Segment and merging it directly into Salesforce, or routing it through HubSpot first.

47. Monitor Threads for Ad API

Keep a close eye on Threads’ development, particularly how quickly they establish an advertising API and whether it integrates with Meta’s existing architecture, to understand new ad opportunities.

48. Explore Reddit for Ad Conversions

Investigate Reddit as an emerging advertising channel due to its opening conversions API and embedded ad formats that blend seamlessly with user posts.

49. Adapt to Aggregate App Attribution

Acknowledge that app ad attribution is now largely aggregate and deterministic matching is no longer feasible, requiring new approaches to measurement.

50. Account for Browser URL Truncation

Be aware that browsers are increasingly stripping out URL parameters, leading to more paid traffic being misattributed as organic, and adjust your measurement strategies accordingly.

Understand that third-party cookie blocking means users remain anonymous until login, leading to a lack of pre-conversion attribution data and potentially misclassifying paid traffic as organic.

52. Prioritize MTA & Probabilistic Modeling

For most businesses, focus on improving multi-touch attribution (MTA) and probabilistic modeling before investing in Media Mix Modeling (MMM), as they often lack the data or readiness for MMM.

53. Conduct Geo-Based Testing

For offline advertising (e.g., billboards), conduct geo-based testing to isolate campaign impact by controlling for confounding factors in specific geographic areas.

54. Recognize MarTech Overwhelm Signals

Look for signs like an overwhelmed growth engineer, lack of time, or increasing system complexity as indicators that it’s time to consider hiring a dedicated MarTech person.

Ensure someone is responsible for understanding data flow, schema, procurement, and legal aspects of MarTech tools to avoid infinite liability and manage contracts effectively as the company scales.

56. Identify MarTech Need by Sprawl

Consider hiring a MarTech person when you accumulate many tools and feel a need for greater efficiency in connecting data and backend infrastructure for growth measurement and execution.

57. Address MarTech Pain Points

Recognize “pain points” like inability to execute due to tool knowledge gaps, setup issues, or fear of impacting existing systems as key drivers for needing dedicated MarTech expertise.

58. Build Around Purchased Tools

Create financial models to demonstrate that buying a third-party tool at the lowest cost and then building custom solutions around it can be more cost-effective and faster than building the entire system from scratch.

59. Leverage Vendor Commitment

When building custom solutions on top of a purchased third-party tool, leverage the vendor’s commitment to your success to potentially achieve accelerated outcomes and support.

60. Apply “Thinking Gray” to People

Extend the “thinking gray” concept to evaluating people; avoid making quick judgments and allow yourself grace to not decide about individuals until a decision is truly required, leading to more nuanced understanding.

61. Cultivate Appreciation & Empathy

Cultivate appreciation by understanding the challenges people face in their daily lives, which fosters empathy, makes you more grateful for good moments, and enhances business relationships.

62. Understand People Beyond Work

Strive to understand colleagues and individuals beyond their work persona, delving into their personal drivers and challenges, which deepens appreciation and improves interactions.

63. Golden B2C MarTech Stack

For B2C, use Amplitude (CDP/product analytics), Customer.io (email, upgrade to Braze), Snowflake (data warehouse), Hightouch (reverse ETL), and AppsFlyer (mobile attribution) for an optimized MarTech stack.

64. Golden B2B MarTech Stack

For B2B, use Amplitude (product analytics), Branch (web attribution), Salesforce (CRM), Snowflake (data warehouse), Hightouch (reverse ETL), and Customer.io (email, upgrade to Braze) for an optimized MarTech stack.

65. MarTech Drives Growth

Understand that the core role of a MarTech professional is to leverage technology and tools specifically to drive growth for the business.

66. Create “Life Frameworks” One-Pager

Compile a one-page document of useful life frameworks (just the words) to serve as a quick reference and aid in remembering valuable mental models.

67. Small Teams: Growth is MarTech

If you’re a small startup (30 people or less), recognize that your growth acquisition person likely is your MarTech person, handling tool setup and data management for ad networks.

Tools are just meant to solve problems.

Austin Hay

You buy the tool to get 90% of the way there, and then you build the cool thing on top with the other 10%.

Austin Hay

The ideal world is that you actually are growing as a business, making more money, hiring more people, acquiring more users, and your total cost of tooling per person goes down.

Austin Hay

From 2010 to 2020, we had the golden years of deterministic matching where, you know, it was very easy to run an ad and understand with precision who installed the app.

Austin Hay

I think Twitter like ruins people's careers. I've already seen multiple careers ruined by Twitter. Some people just don't know how to shut their mouth.

Austin Hay

Designing Attribution Infrastructure for Multi-Touch Attribution (MTA)

Austin Hay
  1. When a user comes to your website, collect the full URL and the referring URL.
  2. Collect all additional marketing parameters (e.g., GCLID, TikTok ID, Microsoft ID) from the URL.
  3. Collect all UTMs (UTM campaign, UTM medium) from the URL.
  4. Store these parameters locally on the device (e.g., in the browser) as 'UTM first campaign' and 'UTM last campaign'.
  5. Every time a person revisits, replace the 'last campaign' or 'last value' with the new one.
  6. Collect this user information as both a user attribute and an event, ensuring the user profile has both first and last attribution data.
  7. Fire off a page view event with first and last attribution data, allowing for deviation if multiple steps occurred in the middle.
  8. Optionally, set these parameters in first-party cookies and third-party cookies for tooling vendors.

Problem, People, System (PPS) Framework

Austin Hay
  1. Identify and clearly define the core 'Problem' that needs to be solved.
  2. Determine 'Who' are the people involved, affected, or needed for the solution, considering their permissions, training, and potential confounding factors.
  3. Only after understanding the problem and people, design or select the 'System' (tool or technology) to solve the problem.
2010 to 2020
Golden years of deterministic matching Period when it was easy to precisely understand ad installs via IDFA and PII.
100 to 150 people
Company size where MarTech becomes centrally owned Critical mass where a 'village approach' to systems and tools is no longer scalable.
30%
Percentage of population for probabilistic modeling The amount of population data needed to create a model that can be extrapolated to 100% when deterministic data is unavailable.
6 months
Investment for basic software skills Estimated time for someone to self-teach web or backend programming skills needed for MarTech.