Austin Hay on the Ultimate Guide to MarTech

Austin Hay on the Ultimate Guide to MarTech

transcriptlennys-podcastmartechgrowthmarketing-technologyattributiondata-infrastructuretools

Austin Hay on the Ultimate Guide to MarTech

Source: Lenny’s Podcast Speaker: Austin Hay Date: ~2022–23 Link: https://www.lennysnewsletter.com/p/the-ultimate-guide-to-martech-austin

Key ideas

  • MarTech as a discipline. Marketing technology lives at the crossroads of product, growth, engineering, and marketing. The role manages first-party (homegrown) and third-party tools, platform architecture, data schema, and contracts/compliance. At 30–40 people it’s a shared responsibility; around 100–150 employees is the critical mass where a dedicated owner is needed. It fits differently in B2C (under growth/CMO) vs B2B (under revenue operations), and centralised vs decentralised models each have trade-offs.
  • Build AND buy — not build vs buy. The modern MarTech mantra: buy a third-party tool for 90% of the use case, then build a custom layer on top for the remaining 10% that creates differentiated velocity. This is how companies get the best of both — rapid commodity capability from vendors plus proprietary advantage from internal engineering investment.
  • B2C stack evolution: CDP → warehouse-native. From 2016–2020, the standard B2C stack was a Customer Data Platform (CDP) at the centre — one SDK, data fanned out to analytics, ad networks, and email tools. Post-2020, cheap warehousing (Snowflake) plus reverse ETLs (Census, Hightouch, Rudder Stack) created a warehouse-native alternative: ingest into a data warehouse, model in dbt, activate via reverse ETL. Engineering-heavy teams should move this way; resource-constrained teams should stick with CDP.
  • Attribution under pressure. 2010–2020 was the “golden age of deterministic matching” — IDFA and pixel-based tracking made it easy to tie ad spend to installs. IDFA deprecation (iOS 14+) and cookie loss have made deterministic attribution impossible for significant audience segments. Teams now model probabilistically: build attribution models for 30% of the audience and extrapolate to 100%. The infrastructure response is to capture all UTMs and first/last-touch parameters from day one — retrofitting later is painful and data is irretrievable.
  • MarTech as quarterback function. The best MarTech leaders are engineers-turned-operators who drive change through cross-functional persuasion — upward (VP of product for platform changes), laterally (rev ops for CRM changes), and downward (data teams for schema). The unsexy but high-leverage work is contract design: knowing growth trajectories and designing cost structures so that tooling cost per user decreases as the company scales, rather than compounding linearly.

Overview

Austin Hay — Head of MarTech at Ramp, former VP of Growth at mParticle, early employee at Branch, Reforge instructor — provides a comprehensive practitioner’s guide to marketing technology as a discipline and function. Topics: what MarTech is and how it differs from growth, when to hire and where to place the role (B2B vs B2C, centralised vs decentralised), B2C and B2B stack architectures and their evolution, CDP vs reverse ETL decision, attribution models (first touch, last touch, MTA, MMM), the probabilistic data era post-IDFA, team structure, day-to-day responsibilities, and favourite tools (Segment, Braze, Snowflake, Hightouch/Census, Amplitude, Salesforce, HubSpot, Marketo).