Julia Schottenstein
Product lead for dbt Cloud at dbt Labs, former VC at NEA, angel investor
Where to hear and read about Julia Schottenstein1 episode or test
Julia Schottenstein is product lead for dbt Cloud at dbt Labs, a former NEA venture investor who joined the company after losing the chance to invest, and she led its acquisition of Transform. The thing most worth learning from her is to build a simple, open product that can stand alone, because a viable independent path gives you the best options in every deal and pricing conversation.
we would be so lucky to have tech debt because that means people are using the product
M&A, competition, pricing, and investing | Julia Schottenstein (dbt Labs)
Julia Schottenstein: the story and the lessons
1Julia Schottenstein's path from NEA investor to dbt Cloud product lead
Schottenstein invested in technical startups at NEA before moving to product. At NEA she backed early-stage companies building for technical audiences: dev tools, infrastructure and data. She calls a move from VC into product unusual, since most people go the other way.
She sees investing and product as the same interests in different proportions: product goes much deeper on technology and gives markets less focus.
She first heard of dbt in 2019, when users called it an identity. Users described dbt as part of who they were, not just a tool. That strength of feeling gave her strong conviction about the company.
She lost the dbt investment, then asked to join the company. In 2020 founder Tristan Handy had a term sheet from another fund but also offered her a shot, and Sequoia won the deal. She asked to put in nearly 20% of her liquid net worth personally, and the board vetoed it.
She did not tell herself she had dodged a bullet. She saw dbt as a runaway train she wanted to board, and called Tristan a few months later to ask to join.
She carries three VC habits into product work. She invests time in a network of operators at companies slightly ahead of dbt Labs and asks how they handled open source, pricing and acquisitions. She is a T-shaped generalist who knows a little finance, business and product and goes deep where she specializes.
She also thinks in power laws and makes bets that bend the trajectory, though a PM can't have 50% of their work go to zero the way a fund can.
2How Schottenstein evaluates a company and spots product-market fit
She judges early companies on people, market, product and distribution. The questions are whether you trust the founder to lead, whether the market is growing with room for a new entrant, whether users show real spark, and whether there is a distribution edge. The same four tests apply whether she invests or joins.
With dbt, Tristan could paint a compelling future and also dig into daily analytics engineering work, and the chaos of cloud warehouses was the opening because dbt added order.
She thinks distribution is arguably more important than product. Getting to market is very hard, so she looks at a company's edge in ecosystem or distribution and at how it will sell.
No company scores 10 out of 10 on all four tests. If you join, she says to ask what you bring that covers the weakest dimension.
Distribution strengths differ between bottoms-up and enterprise companies. dbt started horizontal, and people could begin without ever talking to sales. For enterprise companies she checks whether the team can land a complex sale, has a background in the space and has a network.
Not every company should be product-led. Some sell top-down.
She reads product-market fit from users who can't stop talking about it. Users who bring the product up constantly and want to share it with teammates and people at other companies are the clearest sign. That love does much of the go-to-market work because users become evangelists.
3Why dbt won, and how Schottenstein ships
dbt came from founders feeling the pain as consultants. The founders ran Fishtown Analytics, a consulting firm, for almost two years and used dbt to serve clients. Whenever they hit a paper cut or a slow workflow, they built the fix into dbt.
The timing helped: cloud data warehouses were exploding in 2019, the year Snowflake went from a $4 billion to a $12 billion company.
Simplicity was dbt's strength, and openness made it spread. Anyone who knows SQL can do work once held by data engineers, with no sales conversation needed. Users shared it inside and across companies, and 20,000 companies now use dbt every week.
Revenue is reinvested in the community and product, so the flywheel spins faster, and the large user base draws partners to build for dbt.
She keeps the open standard free and charges for state and collaboration. The open source core is where you describe business logic. The proprietary cloud product covers stateful interactions and cross-team or structural collaboration.
Most lost dbt Cloud deals go to dbt open source, and the team is happy to lose to itself.
She ships good enough and treats tech debt as a sign of success. The first dbt Cloud scheduler was a naive for loop over a jobs table, with no distributed design of Go workers and RabbitMQ because there were no users yet. It has been rebuilt several times and now serves 8,000 companies and 10 million runs per month.
She repeats two sayings against perfectionism: worse is better, and tech debt is a champagne problem. The team was a little embarrassed by the first version, but it got the job done.
4Julia Schottenstein's M&A playbook, from the Transform deal
She says to plan for M&A while you still don't need it. A viable path to staying independent gives you a real alternative in every conversation: do nothing and stay the course. The best preparation is a strong offense in building the company.
Most companies have no path to staying independent forever, so she says they must plan for M&A.
She advises hurting two or three likely buyers while staying friendly. Find the area where you hold an advantage and make yourself impossible to ignore. Transform pressured dbt Labs with a strong semantic layer product while presenting itself as a partner, and dbt Labs later bought it.
"I would inflict pain on that potential buyer make it impossible for them to not notice you."
She bought Transform because it had what dbt Labs lacked. Transform built only the semantic layer, which lets companies define metrics once. dbt Labs had distribution and ecosystem but was behind on that product.
Transform's founders came from Airbnb, famous in data for its Minerva metrics layer, and Transform had already built much of the dbt integration, so the work after the deal was easier.
She says never shut the door on a possible buyer. Founders often refuse to talk to an incumbent or future buyer out of a too-competitive stance. M&A is about creating plan Bs, and you don't yet know if you can go the distance alone.
She builds buyer relationships long before a sale. Corporate development teams exist to meet every possible target, so take the meeting, say you aren't ready to sell, and ask for an intro to a product lead or GM who could sponsor a deal. A startup met Airbnb's head of product at a random South by Southwest party, and a year later it led to an acquisition.
Airbnb was never an obvious buyer, which is why she says to meet everyone.
When you must sell, she says to call a list of about a dozen. No more than a dozen companies will be a very good fit. Screen out those with disqualifying criteria, then call the rest.
With runway left, she narrows to two or three buyers and uses the inflict-pain approach instead.
Out of runway, she says to be plain and not clever. A short note works: the product was interesting, the plan didn't pan out, we want to keep the team together, we are running a process. Buyers mostly acquire the team, so have a data room and your people ready.
Everyone already reads "evaluating strategic alternatives" as a sale.
With runway, she never mentions M&A and talks partnerships. M&A is a dirty word if you plan to stay independent. Frame conversations as collaboration and knowledge sharing, and use "strategic partnership" as the code word every buyer understands.
She tells founders to lean on investors and not fear disappointing them. Investors know about 50% of early-stage portfolios return nothing, and would rather see you land at a great company that helps them later. Ask your VC for intros to possible buyers.
She adds that sometimes you should quit, though she hopes founders keep finding a way forward.
She expects the M&A market to revive soon. The market peaked in November 2021, and founders now accept those valuations won't return. Companies running out of cash or options will bring better assets to market.
She can't name a date but thinks "we're pretty close".
5How Schottenstein thinks about pricing and competition
She starts willingness-to-pay talks before the product is built. You don't choose whether to have a pricing conversation, only when. Having it late means sales is pitching to buyers who aren't excited enough to pay.
She credits Madhavan Ramanujam for the idea, and says zero interest rates let investors fund stars and usage so many companies never planned revenue. dbt Labs doesn't claim to be perfect at pricing.
dbt charges a small fraction of the value it creates, by design. Customers value dbt at 20 to 35% of their cloud data warehouse spend, and the price is a very small fraction of that.
One core value says the company cares more about value creation than value capture.
She ran dbt Labs' first pricing change while the stakes were low. It was the first pricing change in the company's history and taught the team its customers' price elasticity. The goal was to make pricing catch up with how people valued the tool.
Spreadsheets modeled the impact, but only customer conversations showed appetite. Conversion and churn were tracked closely and the team was largely happy with the result.
She asks relative-value and price-band questions, since people won't name a price. Customers compared dbt's worth to their cloud warehouse, then said which price is a no-brainer, which is fair or comfortable, and which is too expensive. Product and product marketing ran dozens of these conversations.
She set three pillars for competing: hold vision, grow the pie, lean into strengths. Most competitor criticism is noise, so keep your eyes ahead. Grow the market with partners and leave space for the ecosystem, holding ground only on the transformation standard and the semantic standard.
Transform's founder Nick Handel led the exercise of writing these down. The market grows as companies move from reporting and BI to ML, which trains on clean transformed data.
6How Schottenstein leads her team and hires
She had engineers act out an algorithm with rope so the team owned it. At an offsite, each engineer was a graph node and the rope was the edges, and they walked the new algorithm step by step. The change moved dbt DAG runs from imperative to declarative, working back from data SLAs.
She wanted everyone to own it, because you can't copy and paste an algorithm and people who don't understand it can't anticipate edge cases. The idea came from an ant farm scene in Gödel, Escher, Bach.
She wants the whole company using the product and living in the community. Over 30% of dbt Labs employees have contributed to its own data transformation project, across product, data, marketing and engineering. All employees are in the 50,000-member dbt Slack.
When an experience misses the mark, dozens of people jump in to fix it. Many staff came from the community, which gives them real ownership.
She single-threads the team on one main priority. The process change with the biggest payoff was doing fewer things, with one mission and everyone growing in the same direction.
She hires for growth mindset and an authentic reason for joining. Her interview question is when the last time was that you taught yourself something new. She also asks why dbt Labs, because that answer matters most when things get tough.
She works inside values that put creating value before capturing it. dbt Labs' values include transparency wins, humility, and work done well being its own end. Board decks are shared, hard conversations happen in the open, and the team never feels successful, only obliged to serve its community.
The mission is to help analytics engineers spread organizational knowledge through data.
M&A, competition, pricing, and investing | Julia Schottenstein (dbt Labs)
Julia Schottenstein explains how to prepare for M&A, handle competition, set pricing and open source boundaries, and ship fast, using dbt Labs as the example. dbt is used by 20,000 companies every week, and she led the acquisition of Transform.
Schottenstein invested in technical startups at NEA before moving to product. At NEA she backed early-stage companies building for technical audiences: dev tools, infrastructure and data. She calls a move from VC into product unusual, since most people go the other way.
She sees investing and product as the same interests in different proportions: product goes much deeper on technology and gives markets less focus.
She first heard of dbt in 2019, when users called it an identity. Users described dbt as part of who they were, not just a tool. That strength of feeling gave her strong conviction about the company.
She lost the dbt investment, then asked to join the company. In 2020 founder Tristan Handy had a term sheet from another fund but also offered her a shot, and Sequoia won the deal. She asked to put in nearly 20% of her liquid net worth personally, and the board vetoed it.
She did not tell herself she had dodged a bullet. She saw dbt as a runaway train she wanted to board, and called Tristan a few months later to ask to join.
She carries three VC habits into product work. She invests time in a network of operators at companies slightly ahead of dbt Labs and asks how they handled open source, pricing and acquisitions. She is a T-shaped generalist who knows a little finance, business and product and goes deep where she specializes.
She also thinks in power laws and makes bets that bend the trajectory, though a PM can't have 50% of their work go to zero the way a fund can.
She judges early companies on people, market, product and distribution. The questions are whether you trust the founder to lead, whether the market is growing with room for a new entrant, whether users show real spark, and whether there is a distribution edge. The same four tests apply whether she invests or joins.
With dbt, Tristan could paint a compelling future and also dig into daily analytics engineering work, and the chaos of cloud warehouses was the opening because dbt added order.
She thinks distribution is arguably more important than product. Getting to market is very hard, so she looks at a company's edge in ecosystem or distribution and at how it will sell.
No company scores 10 out of 10 on all four tests. If you join, she says to ask what you bring that covers the weakest dimension.
Distribution strengths differ between bottoms-up and enterprise companies. dbt started horizontal, and people could begin without ever talking to sales. For enterprise companies she checks whether the team can land a complex sale, has a background in the space and has a network.
Not every company should be product-led. Some sell top-down.
She reads product-market fit from users who can't stop talking about it. Users who bring the product up constantly and want to share it with teammates and people at other companies are the clearest sign. That love does much of the go-to-market work because users become evangelists.
dbt came from founders feeling the pain as consultants. The founders ran Fishtown Analytics, a consulting firm, for almost two years and used dbt to serve clients. Whenever they hit a paper cut or a slow workflow, they built the fix into dbt.
The timing helped: cloud data warehouses were exploding in 2019, the year Snowflake went from a $4 billion to a $12 billion company.
Simplicity was dbt's strength, and openness made it spread. Anyone who knows SQL can do work once held by data engineers, with no sales conversation needed. Users shared it inside and across companies, and 20,000 companies now use dbt every week.
Revenue is reinvested in the community and product, so the flywheel spins faster, and the large user base draws partners to build for dbt.
She keeps the open standard free and charges for state and collaboration. The open source core is where you describe business logic. The proprietary cloud product covers stateful interactions and cross-team or structural collaboration.
Most lost dbt Cloud deals go to dbt open source, and the team is happy to lose to itself.
She ships good enough and treats tech debt as a sign of success. The first dbt Cloud scheduler was a naive for loop over a jobs table, with no distributed design of Go workers and RabbitMQ because there were no users yet. It has been rebuilt several times and now serves 8,000 companies and 10 million runs per month.
She repeats two sayings against perfectionism: worse is better, and tech debt is a champagne problem. The team was a little embarrassed by the first version, but it got the job done.
She says to plan for M&A while you still don't need it. A viable path to staying independent gives you a real alternative in every conversation: do nothing and stay the course. The best preparation is a strong offense in building the company.
Most companies have no path to staying independent forever, so she says they must plan for M&A.
She advises hurting two or three likely buyers while staying friendly. Find the area where you hold an advantage and make yourself impossible to ignore. Transform pressured dbt Labs with a strong semantic layer product while presenting itself as a partner, and dbt Labs later bought it.
"I would inflict pain on that potential buyer make it impossible for them to not notice you."
She bought Transform because it had what dbt Labs lacked. Transform built only the semantic layer, which lets companies define metrics once. dbt Labs had distribution and ecosystem but was behind on that product.
Transform's founders came from Airbnb, famous in data for its Minerva metrics layer, and Transform had already built much of the dbt integration, so the work after the deal was easier.
She says never shut the door on a possible buyer. Founders often refuse to talk to an incumbent or future buyer out of a too-competitive stance. M&A is about creating plan Bs, and you don't yet know if you can go the distance alone.
She builds buyer relationships long before a sale. Corporate development teams exist to meet every possible target, so take the meeting, say you aren't ready to sell, and ask for an intro to a product lead or GM who could sponsor a deal. A startup met Airbnb's head of product at a random South by Southwest party, and a year later it led to an acquisition.
Airbnb was never an obvious buyer, which is why she says to meet everyone.
When you must sell, she says to call a list of about a dozen. No more than a dozen companies will be a very good fit. Screen out those with disqualifying criteria, then call the rest.
With runway left, she narrows to two or three buyers and uses the inflict-pain approach instead.
Out of runway, she says to be plain and not clever. A short note works: the product was interesting, the plan didn't pan out, we want to keep the team together, we are running a process. Buyers mostly acquire the team, so have a data room and your people ready.
Everyone already reads "evaluating strategic alternatives" as a sale.
With runway, she never mentions M&A and talks partnerships. M&A is a dirty word if you plan to stay independent. Frame conversations as collaboration and knowledge sharing, and use "strategic partnership" as the code word every buyer understands.
She tells founders to lean on investors and not fear disappointing them. Investors know about 50% of early-stage portfolios return nothing, and would rather see you land at a great company that helps them later. Ask your VC for intros to possible buyers.
She adds that sometimes you should quit, though she hopes founders keep finding a way forward.
She expects the M&A market to revive soon. The market peaked in November 2021, and founders now accept those valuations won't return. Companies running out of cash or options will bring better assets to market.
She can't name a date but thinks "we're pretty close".
She starts willingness-to-pay talks before the product is built. You don't choose whether to have a pricing conversation, only when. Having it late means sales is pitching to buyers who aren't excited enough to pay.
She credits Madhavan Ramanujam for the idea, and says zero interest rates let investors fund stars and usage so many companies never planned revenue. dbt Labs doesn't claim to be perfect at pricing.
dbt charges a small fraction of the value it creates, by design. Customers value dbt at 20 to 35% of their cloud data warehouse spend, and the price is a very small fraction of that.
One core value says the company cares more about value creation than value capture.
She ran dbt Labs' first pricing change while the stakes were low. It was the first pricing change in the company's history and taught the team its customers' price elasticity. The goal was to make pricing catch up with how people valued the tool.
Spreadsheets modeled the impact, but only customer conversations showed appetite. Conversion and churn were tracked closely and the team was largely happy with the result.
She asks relative-value and price-band questions, since people won't name a price. Customers compared dbt's worth to their cloud warehouse, then said which price is a no-brainer, which is fair or comfortable, and which is too expensive. Product and product marketing ran dozens of these conversations.
She set three pillars for competing: hold vision, grow the pie, lean into strengths. Most competitor criticism is noise, so keep your eyes ahead. Grow the market with partners and leave space for the ecosystem, holding ground only on the transformation standard and the semantic standard.
Transform's founder Nick Handel led the exercise of writing these down. The market grows as companies move from reporting and BI to ML, which trains on clean transformed data.
She had engineers act out an algorithm with rope so the team owned it. At an offsite, each engineer was a graph node and the rope was the edges, and they walked the new algorithm step by step. The change moved dbt DAG runs from imperative to declarative, working back from data SLAs.
She wanted everyone to own it, because you can't copy and paste an algorithm and people who don't understand it can't anticipate edge cases. The idea came from an ant farm scene in Gödel, Escher, Bach.
She wants the whole company using the product and living in the community. Over 30% of dbt Labs employees have contributed to its own data transformation project, across product, data, marketing and engineering. All employees are in the 50,000-member dbt Slack.
When an experience misses the mark, dozens of people jump in to fix it. Many staff came from the community, which gives them real ownership.
She single-threads the team on one main priority. The process change with the biggest payoff was doing fewer things, with one mission and everyone growing in the same direction.
She hires for growth mindset and an authentic reason for joining. Her interview question is when the last time was that you taught yourself something new. She also asks why dbt Labs, because that answer matters most when things get tough.
She works inside values that put creating value before capturing it. dbt Labs' values include transparency wins, humility, and work done well being its own end. Board decks are shared, hard conversations happen in the open, and the team never feels successful, only obliged to serve its community.
The mission is to help analytics engineers spread organizational knowledge through data.