LLenny's Podcast
← All episodes
Eeke de Milliano (Retool, Stripe)02 February 2023

How to foster innovation and big thinking

7Frameworks
15Insights

Frameworks in this episode

Innovation3 steps

Build for Your Best User, Not Your Worst

In early product development, design for the user who gets it instantly, not the edge-case abuser

Read →
Strategy4 steps

Charter, Goals, Roadmap: The Three Documents

Every level of the org needs three docs — mission/strategy, goals, roadmap — written top-down

Read →
Leadership3 steps

Minimum Viable Process (MVP)

Introduce the least process that reduces variance, and give your best people escape hatches

Read →
Innovation3 steps

New Products as Internal Startups

Incubate new products like funded startups: tiny teams, prove ROI before funding, keep them separate

Read →
Strategy3 steps

The 70/20/10 Investment Split

Allocate product building time 70% core, 20% strategic, 10% bets to balance stability and innovation

Read →
Leadership3 steps

The Product Talent Portfolio

Build a team of complementary spikes instead of clones of yourself, and hire specifically for your gaps

Read →
Innovation4 steps

The Three Innovation Blockers

Diagnose why a team plays it safe by checking the three specific things that quietly kill big thinking

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Myth Buster· 1

Myth Buster15:30

Pricing Isn't a Trap-Door Decision (But Titles Are)

Eeke uses Amazon's one-door vs two-door framing to challenge how teams treat certain decisions. People assume pricing is irreversible, but you can grandfather existing users and change pricing for future ones, so you can move faster than you think. Titles, by contrast, genuinely are a trap door because once you give someone a title you can't take it away, which is why Stripe was so rigorous about not having them.

  • A "trap door" (one-door) decision is one you can't come back from; a two-door decision is reversible
  • People wrongly treat pricing as a trap door
  • You can grandfather existing users and change pricing only for future users
  • Titles are a true trap door: once given, they can't be taken away

“is it a trapped or decision which I think is a an Amazon concept it's like the one if you make the decisions is it…”

Eeke de Milliano · 15:30

“a lot of people for example think pricing is a trapped or decision but actually like it's really easy to grandfather your existing users into…”

Eeke de Milliano · 16:00
#decision-making#pricing#leadership#stripe

Hot Take· 1

Hot Take39:30

Process Is Variance-Reducing, and That's the Hidden Cost

Eeke's core take on process: by definition it reduces variance, because you introduce it when you worry your org's variance is too high and want people to meet a standard. The cost is that while you raise the floor and bring folks up to the average, you also drag your highest performers and most creative thinkers down to that average. That tension is why the pull toward more process grows as companies get bigger.

  • Process exists to reduce variance and bring people to a common standard
  • The cost is averaging down your best, most creative people
  • Those top performers often don't need process to do their best work
  • Companies add more process as they grow because managing high-variance talent gets harder

“process by definition is variance reducing like you're introducing it because you worry that the variance in your org is too high”

Eeke de Milliano · 39:30

“you're also bringing other folks down to the average and oftentimes the folks you're bringing down are your highest performers your most creative thinkers”

Eeke de Milliano · 40:00
#process#leadership#org-design#talent

Explainer· 5

Explainer07:00

Why Merchants Eat the Cost of Stolen-Card Fraud

Eeke explains the counterintuitive economics behind Stripe Radar: when a fraudster buys goods with a stolen credit card, it is the merchant, not the card owner, who ultimately loses the money. Stripe built machine-learning models to predict fraudulent payments and then built a product layer so customers could understand blocked payments and write their own rules.

  • If a customer pays with a stolen card, the merchant is ultimately on the hook for those funds
  • Fraudsters follow the money wherever payments are processed
  • Stripe trained ML models to predict when a payment was fraudulent
  • The product challenge was explaining the model to humans and letting them interact with it

“if a merchant processes a payment from a customer and that customer used a stolen credit card the merchant is ultimately on the hook for…”

Eeke de Milliano · 07:00
#fraud#payments#machine-learning#stripe
Explainer14:00

Stripe's Real Edge: A Thousand Good Decisions, Not One Big Bet

Eeke pushes back on the idea that Stripe had one or two pivotal decisions that set its path. What it was actually good at was making a lot of good decisions all the time, big or small, driven by a cultural respect for rigorous thinking and a very strong writing culture. She argues you can't be a good writer without being a clear thinker.

  • There was no single pivotal decision; Stripe made many good decisions consistently
  • "Think rigorously" was a core value; nobody accepted "it's a best practice" as a reason
  • All communication ran through writing: business reviews, strategy memos, product reviews
  • Cultivating good writers produced good thinkers as a by-product

“was really good at was just making not just like one or two big decisions good decisions was like making a lot of really good…”

Eeke de Milliano · 14:30

“I don't think you can be a good writer unless you're a clear thinker”

Eeke de Milliano · 15:00
#culture#decision-making#writing#stripe
Explainer21:30

Three Reasons Teams Never Get to Innovate

Eeke reframes the innovation question: not why some teams are creative, but why most teams end up on boring incremental work despite nobody wanting to. She names three culprits: fear of failure (leaders want the upside of innovation but won't pay its cost), teams getting bogged down by urgent tech and product debt, and the simple fact that thinking big is genuinely hard when you're treading water day to day.

  • No one wants to work on boring incremental stuff, yet most teams do
  • Fear of failure: leaders want innovation's upside but won't accept the stumbles
  • Urgent work (tech debt, bugs, incidents) leaves no room for creative thinking
  • Thinking big is inherently hard and takes deliberate time most people don't have

“I think leaders want the upside of innovation but they're not really willing to deal with the cost of innovation which is like look if…”

Eeke de Milliano · 22:00
#innovation#leadership#product-management#tech-debt
Explainer35:00

How Retool Shipped Three New Products in One Year

Retool launched Workflows, Retool Mobile, and a virtual database in a single year with a ~300-person company. Eeke's playbook: start each initiative tiny (one or two people for the first six months), treat them like startups where Retool is the VC funding proven ROI, and deliberately keep the teams separate from the core org so they can move fast without getting bogged down in the realities of a product that already has market fit.

  • Each product started with one or two people for roughly the first six months
  • Teams got more funding only once it was clear there was something there
  • Retool acted as the VC; teams had to prove ROI in engagement or revenue to advance
  • Keeping new teams separate let Retool Mobile discover it had a different target customer

“we really treated them like startups where like rituals the VC and like ritual funds with like resources”

Eeke de Milliano · 36:00

“really deliberate about keeping these teams separate from the rest of the org especially early on”

Eeke de Milliano · 36:30
#product-management#innovation#org-design#retool
Explainer56:00

The Product Talent Portfolio: Hire People Who Don't Look Like You

Eeke's product talent portfolio idea warns against the temptation to build a team in your own image, because you value and can assess the skills you already have. The best teams balance complementary skill sets so the whole is stronger: pair homegrown PMs who deeply get the product and carry the culture with outside PMs who bring conventional rigor, and mix execution machines with visionaries. Every six months she charts the team's strengths and weaknesses and hires against the gaps.

  • Managers tend to hire in their own image because it's easier to assess familiar skills
  • Balance homegrown, culture-carrying PMs with experienced outside PMs for rigor
  • Mix execution-strong PMs with visionary PMs across all product pillars
  • Every six months she maps team strengths and weaknesses and hires for the weaknesses

“it's really tempting as a manager to build a team in your image because you understand their skill sets”

Eeke de Milliano · 56:30

“I always push the leads to sort of hire people who don't look like them”

Eeke de Milliano · 57:30
#hiring#team-building#product-management#leadership

Story· 1

Story24:30

"Crazy Ideas": Giving Teams Permission to Think Bigger

Because thinking big is hard when you're heads-down, Eeke says you have to explicitly give teams permission to think by building it into rituals. Every Retool team charter ends with a "think bigger" prompt, and each year leadership sends out a blank "crazy ideas" doc soliciting ideas that probably make no sense but could 10-100x the business. Retool Workflows started as one of those crazy ideas.

  • Team charters end with a "think bigger: with 20% more time, what would you do?" section
  • A yearly "crazy ideas" doc invites ideas with a 90% chance of making no sense
  • The 10% that land could make a 10-100x difference to the business
  • Retool Workflows (launched Nov 2022) came off the crazy ideas list

“you have to give teams permission to think”

Eeke de Milliano · 24:30

“crazy ideas are ideas that we shouldn't obviously do there's a 90 chance that they make no sense but in the 10 chance that they…”

Eeke de Milliano · 25:30
#innovation#rituals#product-management#retool

Tool· 1

Tool1:01:30

Favorite New Product: Rewind

In the lightning round, Eeke recommends Rewind, an app that records everything you do on your computer and makes it searchable. She says it fits how she now works, where instead of hunting through browser tabs she just searches directly for the thing she needs.

  • Rewind records everything you do on your computer and makes it easy to search
  • It fits a workflow of searching directly rather than navigating to tabs
  • Eeke was running it at the time of recording and calls it incredible

“it basically records everything that you're doing on your computer and then makes it really easy for you to search it”

Eeke de Milliano · 1:01:30
#tools#productivity#software#lightning-round

Takeaway· 6

Takeaway05:00

Her #1 Advice for Aspiring PMs: Go Get a Sales Job

Eeke joined Stripe in 2013 as its first account executive, not as a PM. Because Stripe's volume was inbound, the role was really about talking to customers, understanding their pain points, and figuring out how the product could serve them, which is the core of the PM job. She now tells anyone wanting to move into product management to first go sell.

  • Eeke was Stripe's first account executive when it was ~50 people with no PMs
  • The sales job was mostly inbound: helping customers map Stripe onto their business flows
  • Talking to customers and understanding pain points is what PMs actually do
  • Her standing advice for breaking into PM is to go get a sales job first

“when people ask me hey like what's your advice for going into product management I'm like well you should go get a sales job I…”

Eeke de Milliano · 05:00
#product-management#career#sales#stripe
Takeaway22:30

To Make Failure Safe, Put It on Stage

Eeke argues the only real way to reduce fear of failure is to shine light on it and normalize it. The tactic lives in the follow-ups: have people publicly write up and present what they learned, using channels usually reserved for wins. At Retool a "shipped" email list broadcasts learnings, and at Airbnb there was even an award for the biggest failure.

  • Normalizing failure means publicly discussing it, not just tolerating it
  • Rename "postmortem" to "retrospective" to frame it as learning
  • Have people write learnings and send them to the whole org or present at all-hands
  • Airbnb ran a short-lived award for the biggest failure

“you have to start Shining Light on failure that's that's really the only way like you have to start normalizing it a little bit”

Eeke de Milliano · 22:30

“at Airbnb there was I think for a period there's an award for the biggest failure”

Eeke de Milliano · 28:30
#innovation#failure#culture#leadership
Takeaway41:30

MVP = Minimum Viable Process (Always Ship an Escape Hatch)

Eeke connects standardization to a broader worry that we're asking humans to be more robot-like and formulaic, which quenches creativity. Her answer when process is necessary: make it the minimum viable process and always give people escape hatches. She now writes at the top of templates that if the template doesn't fit what you're explaining, don't use it, it's just the floor, go higher if you can.

  • Formulaic, mass-produced rules quench creativity and cut off the highs
  • If you introduce process, be sure you accept its cost and provide escape hatches
  • She calls it the MVP: minimum viable process
  • Templates now say: this is the bar, break out of it if it doesn't work for you

“so I've started referring to this as like the the MVP the minimum viable process”

Eeke de Milliano · 42:00

“be really really sure that you're okay with that cost and give folks Escape hatches”

Eeke de Milliano · 42:00
#process#creativity#templates#leadership
Takeaway45:30

Build for Your Best User, Not Your Worst

One of Eeke's core mental models: it's easy to fixate on abuse and all the ways a product could be used badly, then contort the product to guard against those worst users, who should only ever be a fraction of your base. Instead, especially in early product development, build for the user who jumps in and gets it immediately. When a team worried about future abuse, founder Anthony reframed it as an amazing problem to have.

  • Designing around worst-case users warps the product in weird ways
  • Worst users should be a small fraction, so don't build primarily for them
  • Onboarding should target the user who gets it immediately, not the ill-suited one
  • Founder Anthony reframed worrying about abuse at scale as "an amazing problem to have"

“build for your best user not your worst user”

Eeke de Milliano · 45:30

“wouldn't that be an amazing problem to have”

Eeke de Milliano · 47:00
#product-management#mental-models#onboarding#retool
Takeaway47:30

Build the Scooter, Not the Axle

Borrowed from Retool's head of design Ryan, this mental model reframes MVPs: if you're building toward a car, don't ship a lone axle and wheels, ship a scooter first, then a bicycle, motorcycle, and finally the car. The point is to find the thin slice that delivers complete value to the customer end-to-end. Retool Mobile applied this by launching only for companies with field workers doing inventory management.

  • An MVP should be a usable whole (scooter), not a non-functional component (axle)
  • Progress from scooter to bicycle to motorcycle to car
  • Find the slice that gets a customer to complete value on a smaller thing first
  • Retool Mobile shipped a deliberately narrow slice: field workers doing inventory management

“build the scooter not the axle”

Eeke de Milliano · 47:30
#product-management#mental-models#mvp#retool
Takeaway48:30

The 70/20/10 Split for Where Product Time Goes

Eeke reduces a lot of product management to funnels and portfolios, and offers her investment heuristic: 70% of building time to your core product with market fit, 20% to strategic initiatives that aren't core but you know you have to do, and 10% to bets. Maintenance, tech debt, and bugs live squarely in the 70%; no more than 70% of resources should go to the core.

  • 70% to the core product that already has product-market fit
  • 20% to strategic initiatives that are non-core but necessary
  • 10% to bets
  • Maintenance, tech debt and bugs fall inside the 70%, not a separate bucket

“70 of your building time should really be going to your core product that has product Market fit”

Eeke de Milliano · 48:30
#product-management#prioritization#mental-models#roadmap