Aug 13, 2023 · 1h 24m · lennys-podcast
The ultimate guide to Martech | Austin Hay (Reforge, Ramp, Runway)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this in-depth masterclass, MarTech veteran Austin Hay and host Lenny Rachitsky break down modern marketing technology, exploring data stack architectures, attribution modeling in a privacy-first world, and organizational strategies for scaling growth infrastructure.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Lenny holds 14.8% of the talking time here. How this is scored →
Lenny as informed peer 2.6 Guest teaching 6.6 Guest disagreement 0.6 Lenny pushing back 0.3
speaking balance: gold is Lenny, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 56:05 Austin pushes back against hasty MMM adoption
Austin directly challenges the recent industry hype around MMM, explaining that most businesses lack the data foundation and actually need basic MTA and probabilistic models.
Hardest push from Lenny ▶ 1:09:54 Lenny defends corporate synergy jargonWhen Austin expresses disdain for the word 'synergy', Lenny humorously pushes back to defend the term as useful and clear.
Biggest teaching moment ▶ 45:25 Austin's step-by-step masterclass on tracking attribution parametersAustin walks through a granular, highly technical blueprint for capturing first-touch, last-touch, and middle-touch UTM parameters to future-proof attribution pipelines.
Lenny holds their own ▶ 48:59 Lenny synthesizes third-party data warehouse constraintsLenny distills Austin's complex explanation into the fundamental core principle of data warehouse taxonomy and custom event schemas versus rigid third-party schemas.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Lenny as informed peer | Guest teaching | Guest disagreement | Lenny pushing back | Why |
|---|---|---|---|---|---|---|
| Defining MarTech and the Scope of the Role | 1 | 6 | 0 | 0 | Lenny opens by openly admitting he does not fully understand MarTech. Austin gives a comprehensive definition framing it as a hybrid between product management and engineering systems. | |
| Growth vs. MarTech and the 100-Person Scaling Inflection Point | 3 | 7 | 1 | 1 | Lenny asks how MarTech differs from general growth and attempts to synthesize Austin's answer. Austin schools him in detail on the organizational inflection point around 100-150 people where community ownership breaks down. | |
| Organizational Alignment: B2B vs. B2C and Reporting Structures | 2 | 7 | 1 | 0 | Lenny inquires about org chart positioning. Austin walks through a detailed matrix contrasting B2B vs. B2C and centralized vs. decentralized reporting structures. | |
| The Dual Role: High-Leverage Administration and Systems Architecture | 3 | 6 | 0 | 0 | Lenny asks whether a MarTech hire operates as an IC or manager. Austin explains the dual necessity of deep technical IC competence and cross-functional leadership, as well as contract management. | |
| MarTech vs. Marketing Operations and Supporting User Acquisition | 3 | 6 | 1 | 0 | Lenny asks how paid user acquisition interacts with MarTech. Austin clarifies the difference between MarOps (operational, business analyst) and MarTech (engineering/technical), citing Ramp's custom conversion pipelines. | |
| Goal Setting, Unit Economics, and Major Platform Migrations | 2 | 6 | 0 | 0 | Lenny asks how MarTech professionals are evaluated on goals. Austin explains unit economics, tool cost efficiency models, and large platform migrations. | |
| Evolution of the Data Stack: CDPs, Warehouses, and Reverse ETL | 4 | 8 | 1 | 1 | Lenny asks for starting tool recommendations and prompts Austin for reverse ETL examples. Austin delivers an exhaustive masterclass on CDPs, data warehouses, dbt, and reverse ETL architectures. | |
| Sponsor Segment: Brave Search API | 4 | 9 | 1 | 1 | Following the sponsor read, Lenny asks about attribution. Austin dives deep into multi-touch attribution (MTA) vs. mixed media modeling (MMM), giving an exact technical recipe for capturing UTMs and ad click IDs. | |
| Privacy Shifts, Probabilistic Attribution, and New Ad Platforms | 3 | 8 | 2 | 0 | Lenny brings up post-ATT attribution challenges. Austin educates Lenny on probabilistic attribution modeling, browser tracking parameter stripping, and why most businesses are not actually ready for MMM. | |
| Hiring and Interviewing High-Impact MarTech Talent | 2 | 6 | 0 | 0 | Lenny asks about hiring profiles and red flags. Austin discusses assessing intellectual curiosity, testing for tool-agnostic problem solving, and looking past false flags like resume gaps. | |
| Core Operating Frameworks: PPS, Build-and-Buy, and Thinking Gray | 3 | 7 | 1 | 1 | Lenny enthusiastically initiates a discussion of mental frameworks. Austin introduces PPS, Build-and-Buy, and Steven Sample's Thinking Gray principle, while Lenny playfully defends the word synergy. | |
| Lightning Round: Recommended Books, Shows, and Life Mottos | 2 | 4 | 0 | 0 | In the lightning round, Austin provides personal recommendations, favorite tools like Cal.com, and an emotional reflection on hardship and appreciation. | |
| Austin Hay's Recommended 'Golden Stack' for Modern Growth | 2 | 6 | 0 | 0 | Austin lays out his prescriptive Golden Stack tool combinations for B2C and B2B, followed by an anecdote about attempting to fly drones in Washington DC. |
Statements from this episode (25)
Hay: MarTech operates as product management for internal and external platforms
“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, because, you know, marketing technology can mean a collection of third party tools, whi…”
Hay: Startups need dedicated systems owners at 100 to 150 employees
“Usually around, I would call it, like, a hundred to a 150 people is the critical mass where you can't just have a village approach to systems and tools.”
Ramp splits revenue operations into business and marketing technology
“At ramp, we're big enough and we're a B to B company, but we have a B to C top of funnel where we try to acquire users and get them to fill out our application to get a credit card. We have like a distinct revenue operations team that's broken into business te…”
Hay: High-velocity MarTech means buying 90% of tools and building 10%
“That's how you actually have a lot of velocity, is thinking about not just build versus buy, it's build and buy now. So you buy the tool to get 90% of the way there, and then you build the cool thing on top with the other 10%.”
Hay: Most startups should avoid decentralized operations models
“As you add more operational people, it just creates more systems, and so, like, unless you're a massive company where you need that type of scale, I think, like, most startups, like, should avoid that decentralized model.”
Hay: Top MarTech leaders require recent hands-on operating experience
“You have to be an IC in that you are the most senior technical expert on all first party and third party systems. So you have to know really well how third party tools work, and you don't know that without doing the work yourself. So I do find that like some o…”
Hay: Cross-functional influence matters more than team size when hiring MarTech
“So what I would say is like more important to look for than how many people has this person managed is how well can they manage upward, laterally, and downward?”
Austin Hay: MarTech is engineering-based while MarOps is semi-technical operations
“Marketing technology has tech in it, so it's usually an engineer or somebody with an engineering background doing that function. Marketing operations is usually not always technical. It's maybe a systems analyst or a business analyst, you know, could be somebo…”
Ramp optimizes ad campaigns on bottom-funnel opportunity values via synthetic events
“And actually one of the coolest and most fun projects that we worked on early when I joined ramp is we were optimizing. We're trying to get top of funnel data all the way down to the bottom of the funnel and tie it with opportunity data. So we could send that …”
Hay: Warehouse ownership costs dropped around 2020, reducing data team size needs
“The thing about it, though, that I think really changed around 2020 is that the cost of ownership of warehousing became much cheaper. And so, twenty-twenty-one, you start getting the place where, like, yeah, it actually makes a lot of sense and is really easy …”
Hay: Advanced engineering teams using Snowflake and dbt should adopt reverse ETL
“If you are an advanced engineering Culture, and you are cutting edge, and you're going to do a bunch of modeling in dbt, and you already have Snowflake, you should move towards a model of using a reverse ETL.”
Hay: Companies should design data tracking systems for multi-touch attribution from day one
“And so in my mind, the best practice that everybody should stand up from day one is to basically design the system for MTA, and then use whatever makes sense as you grow.”
Hay: Snowplow is the only CDP allowing custom schema control
“Most third party B to C tools don't allow you to control the schema. There's only one CDP I know that does that. That's Snowplow. The rest are, there's a user object and an event object. So you can either stuff data as a user property onto the user object, or …”
Austin Hay: Deterministic mobile ad tracking via IDFAs is effectively dead
“From 20 10 to 20 20, 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. Maybe you don't know their name, but you actually would know their IDFA and you could tie…”
Hay: Browser URL parameter stripping inflates organic web traffic counts
“Browsers now are stripping out those URLs we talked about. So you're just seeing a bigger and bigger percentage of people being counted as organic that actually came from a paid advertisement because when they got redirected to your website, they just, the bro…”
Hay: Most businesses are not ready for Marketing Mix Modeling
“Most of the time businesses are not ready for MMM. They actually just need MTA, and they need better probabilistic modeling.”
Hay: Breaking into MarTech requires six months of coding, not a CS degree
“You don't have to go get a software engineering degree. You can teach yourself. I am self-taught. You can take a coding academy online. I think you get enough knowledge through being able to do web programming or some type of back-end programming So I would sa…”
Hay: Never join companies hiding financials from director-level hires
“One of the red flags whenever I'm approaching a company to work for them is I'll ask for their company financials and like a company that's not willing to divulge their financials to like a director level or above person. I don't want to do business with, beca…”
Hay: Systems thinkers should delay decisions as long as possible
“And one of his tactics is this concept of thinking gray, which is actually to not decide for as long as you possibly can before you have to decide. And it's really challenging because it involves this little thing called patience, which I do not have a lot of …”
Hay: Asking candidates about recent hardship encourages vulnerability and calms them
“Tell me about the most difficult or challenging thing you've overcome in the last year in your life. It doesn't have to be work related. It can be personal. And I think it's like a great way to just like reset the atmosphere Make people dig a little bit deeper…”
Hay: Cal.com is superior to Calendly for multi-calendar sync and modern UX
“I've been a big Calendly user for a long time, but Calendly is pretty expensive. If anyone from Calendly is listening, you want to give me a promo, cool, but it's very expensive, and then I also just found that, like, it is not always graceful at syncing multi…”
Hay reveals the ideal golden tech stack for B2C marketing
“If I was a B to C business, I'd buy Amplitude for my CDP, I'd buy customer IO, and maybe I'd upgrade to Braze in the future. I'd put everything in Snowflake. I'd buy high touch to reverse ETL, that data out to my ad networks. For attribution, probably AppSlyer…”
Austin Hay: Branch is better for web-only attribution
“If it's a web only business, probably use branch because branch is better for web.”