DoorDash
A delivery app for restaurants, groceries, convenience, alcohol and retail, with the DashPass membership.
DoorDash built DashPass by testing a cheap precursor, accepting a money-losing first version, then tuning minimum order ($15), fee discount and eligible merchants until it was profitable at the August 2018 launch. It reached 35M+ members, and about 75% of grocery and retail orders come from members. Lesson: fix unit economics before launching a subscription, then keep price flat and improve benefits.
- 1Email first-time customersPrice complaints became the seed
- 2Sell a delivery pack10 deliveries for $20, built in weeks
- 3Ship a hacky MVP4-5 engineers, three or four weeks
- 4Run it as A/B testOrder frequency up almost 2%
- 5Fix economics, launch nationallyProfitable version launched August 2018
From Jimmy Liu on Lenny's Podcast
What was said about DoorDash
1How DashPass went from a money-losing test to a profitable launch
DashPass started from customers saying delivery was too expensive. Jimmy Liu emailed every first-time DoorDash customer in San Francisco, asking what went well and what went poorly. Affordability came up again and again, alongside missing restaurants and bad deliveries.
He called these emails probably the single highest signal used to grow the market, and later did the same for first grocery orders. Customers described delivery as a luxury they couldn't afford every day.
Don't launch a subscription until each extra order is profitable. DoorDash waited until delivery economics turned positive, then picked test markets by sorting submarkets by profitability. Two years earlier, San Francisco was losing $10 to $12 an order on about 1,500 deliveries a day.
When each order is profitable, customers, DoorDash, merchants and Dashers all want more orders. "We can't have a situation where people ordering more actually causes us to lose more money."
A cheap delivery pack tested the idea before the subscription existed. The team shipped a pack of 10 deliveries for $20, or $2 each, in a week or two in Dallas and one other Midwest submarket. It showed people ordered more with discounted delivery.
The pack was much faster to build than a subscription and became one more data point, with customer talks and finance modeling, to justify building DashPass.
A four-to-five engineer pod shipped a hacky MVP in weeks. Jimmy carved one pod out of his own team and built DashPass in three or four weeks: unlimited free delivery over $20 with the whole roughly 15% service fee removed. It lost money on purpose to test one simple value proposition.
The COO said it took the best customers, made them unprofitable and got them to place more unprofitable orders. Jimmy feared being fired but knew several levers could fix the math: "Just build the thing, test it, and get real data."
An A/B test proved members ordered more, and the lift compounded. Executives worried heavy buyers were simply the ones signing up, so DashPass ran behind an A/B test. Order frequency rose almost 2% in two to three weeks, growing from 0.5 to 1% up to 2%, against typical tests that gain about 0.1% and flatten.
At about $4 billion in run-rate gross order value, a one or two percent lift was worth a great deal.
Three levers turned DashPass profitable: minimum, fee discount, eligible merchants. The order minimum dropped from $20 to $15 and the service fee discount was cut to about 10 of the 15 points. Merchants were included based on ticket size, commission rate and order accuracy, and the profitable version launched nationally in August 2018.
Customers had complained the $20 minimum was hard to hit. Restaurants with frequent missing items were excluded because inaccurate orders raise support costs. The final terms and member economics are not given in numbers.
Customer evidence plus modeling won over the skeptics; results converted them. The COO, a former Amazon executive, called DashPass a dumb idea, and the head of product also doubted it. Jimmy had talked to more customers about affordability than anyone and had built credibility first; months after launch the COO called DashPass the way DoorDash wins affordability.
Jimmy suggests building credibility for roughly six months at a new company before betting on yourself. "If you don't bet big, you can't win big."
2Scaling DashPass: flat $9.99, upsells and new verticals
Once DashPass caught fire, DoorDash made it a company-wide effort. The team grew from four or five engineers to over 100, and in annual planning every team was asked what DashPass benefit it could build. Members showed higher order frequency, retention and LTV.
When DashPass hit 10 million members faster than expected, Tony Xu raised the goal to 20 million, then 100 million. It exited 2025 with more than 35 million members, about half of monthly users.
Price stayed at $9.99 because compounding beats a one-time revenue bump. DashPass has cost $9.99 a month since 2018 while benefits grew. Jimmy says a $1 or $2 hike could lift revenue 10 to 20% overnight but would slow growth over 5 to 10 years.
This is his judgment, with no test shown; he notes no one runs an A/B test for 10 years, so some bets need conviction. He treats price as one of the last levers, since a higher fee makes the product worse.
Retention comes from one big saving plus many small features. Members save about 80% on fees, a large improvement on the core problem rather than 10 or 20%. On top of that, many small changes to sign-up, billing and discovery each add a point or two.
Jimmy says the resulting retention is causal, not just correlation.
Audit where people discover the membership and push it on big baskets. Walk the flows, especially checkout, and ask what share of customers even see the upsell. DoorDash promotes DashPass more loudly when a single order would more than pay for the $10 monthly fee, and stays quieter on small orders.
Talking to customers about placement and wording makes the upsell a win-win, since it shows them how they could have saved.
Free delivery pulls members into grocery and retail. About 75% of DoorDash's grocery and retail orders come from DashPass members. Free delivery lowers the risk of trying a new category for people already in the habit.
One app that solves more jobs stays top of mind versus separate apps.
Merchants joined because DashPass brought large, profitable orders. High-commission, large orders grew merchants' business, and reliable volume matters to restaurants with big fixed costs. Jimmy calls the merchant pitch the overlooked part of DashPass's growth.
3The 2013 founding tests in Palo Alto
A 43-minute, $9 page tested whether anyone wanted delivery. The four Stanford founders built a static page with eight PDF menus, a Google Voice number that rang all their phones and a Square reader for payment. They suspected delivery from non-delivering restaurants might not exist because nobody wanted it.
Spending no money on marketing, they wanted organic demand and watched whether Tony Xu's bank account drained. It did not.
Doing every delivery themselves revealed what the business really was. With no logistics experience, the founders took orders and drove the food for roughly six months, at about 10 orders a day and a best day of 21. They found a delivery has about 20 steps, each with delays, and needs a consumer site, restaurant app, driver app and dispatch system.
Delivering by hand also showed the typical customer was a busy mom with young children. "You actually have no idea what are all the sources of delay in an order until you actually go and do the work."
Suburbs beat dense cities on both speed and demand. Deliveries were faster in Palo Alto than in San Francisco thanks to easy parking, single-family homes and no lobbies or elevators. Stanford's nearest restaurants were 2 miles away, so customers needed delivery more.
Palo Alto's hub-and-spoke layout matches most US towns. DoorDash went after suburbs with little competition while Postmates spread across cities and categories.
Restaurants came first because they offered the densest network. The founders compared restaurants, grocery, convenience and retail. About a million US restaurants versus a couple hundred thousand grocery stores made prepared meals the best start for a network that could later carry anything.
Only 20,000 to 25,000 restaurants delivered at the time, mostly pizza; existing services just faxed orders. Keith Rabois backed DoorDash after the founders' stat that 93% of US restaurants didn't deliver. The idea came from a one-person baker's binder of refused delivery orders, found while interviewing about 300 businesses.
The YC summer tested three unit-economics questions, not demo day. In summer 2013 the founders only asked whether consumers would pay $6, whether restaurants would partner for 15% and whether drivers could earn a workable wage.
Rabois makes the same point: validate the mechanics, such as deliveries per hour to break even and enough density, rather than asking people if they want the product.
Higher pay barely moved Uber X drivers, proving Dashers are different people. Fearing drivers only chased money, Xu offered 20 DoorDash and 20 Uber X drivers a guaranteed $25 an hour against their roughly $20. Only one of the 40 switched.
Today more than half of Dashers are women, the average Dasher works three to four hours a week, and 90% drive under 10 hours.
4The Stanford game refunds and re-earning trust daily
DoorDash refunded every late order despite weeks of cash left. After a Stanford football game in September 2013, every delivery ran at least an hour late. Nobody asked, but the founders refunded everyone at about 40% of a bank account holding two or three weeks of cash, then delivered cookies at 5 a.m.
"We'd rather die trying to be excellent than to live to be mediocre." Xu took from it that the scoreboard resets to zero each day: customers judge every order on selection, price, speed, accuracy and problem handling.
Xu works customer support daily to see what data misses. He reads long, 2,000-word emails from Dashers, debugs them himself in internal tools and then calls or emails back. He argues data should win prioritization, but improving a product means fixing the tails where power users and new users show up.
"The greatest killer of a business is usually silence." Jimmy Liu kept a similar habit of emailing first-time customers.
5Experiments as a learning system, and how new bets get funded
Run thousands of experiments; about 5% survive and compound. Because the physical world is undocumented and always changing, DoorDash runs tens of thousands of experiments and roughly 5% reach customers. Hacky tests become products, then engineered systems.
At millions of orders a day, one-in-a-million events happen often, so the answer is prevention plus fast emergency response. "It's always the data that you can't see that kills you."
New bets earn funding stage by stage, like internal venture capital. Anyone can propose an idea, and if an experiment shows customers want it, it enters planning with a small budget that grows only as it solves the problem. Xu modeled this on DoorDash's own scarce early years, when its best products were built.
He runs two operating systems: the core business as a big airplane getting a midair engine transplant, and new ventures as paper airplanes with their own measures, talent, timelines and incentives.
DoorDash built its own delivery robot when partners wouldn't. Partners focused on robotaxis and nobody wanted to build a small last-mile vehicle, so DoorDash started its own project in 2019. About six years later it shipped the Dot, which runs on roads, sidewalks and bike lanes in Phoenix and Scottsdale.
"We don't have a religion about whether or not we have to build the product or someone else has to build the product."
Local data and logistics aim to make DoorDash any merchant's first call. DoorDash structures physical-world data to flag out-of-stocks, underpriced items and bundles, runs pricing tests for merchants, and operates warehouses and fulfillment for Kroger and CVS.
The idea was seeded in 2014 when a California farm asked if DoorDash could run its trucks. Xu grounds it in needs that never change: selection, price, speed and convenience.
6Surviving 1,000 days of failed fundraising
Xu focused on controllable metrics while over 100 investors said no. DoorDash's Series C collapsed in 2016 as public markets fell, and it raised a fraction of what peers did for about three years. Xu showed every metric, including the falling cash balance, at all-hands and rallied 20 to 25 leaders on one goal: take share, get more profitable, don't run out of cash.
He stopped counting rejections after 50. Unable to outspend rivals, DoorDash had to win on retention and engagement. He still ignores the stock price, since he can't control it.
7Hiring for bias to action
Test candidates with real tasks and check references until one is negative. DoorDash's final interview gave candidates $20 and eight hours to acquire 100 customers, and engineers did deliveries with Xu in his Honda. Xu does 20 references on every senior hire.
Christopher Payne, later COO, drove deliveries for four hours with his son after one conversation, then wrote a 30,000-word email on why the logistics algorithm was flawed.
8How DoorDash overtook Grubhub
Supplying its own delivery fleet expanded selection beyond Grubhub's network. DoorDash changed the atomic unit of supply by delivering for any restaurant, including ones that had never delivered or had no agreement. Grubhub stopped being comprehensive, and DoorDash raised huge rounds to lock in national chains.
Grubhub had raised $80 million and $100 million in its IPO while challengers raised $100 million every six months. The pandemic turned DoorDash's margins positive for the first time. It runs a heavily managed marketplace that owns delivery logistics.
Grubhub assumed money-losing delivery would fail and didn't overreact. Grubhub expected negative-margin rivals to lose their subsidies, but they kept raising and won. Casey Winters says Grubhub could probably have acquired DoorDash multiple times early on.
Sarah Tavel adds that Grubhub, as a public company, was slow to fulfill delivery itself and couldn't access capital like private rivals. "The only rational reaction is overreaction."
DoorDash's Playbook for Scaling DashPass to 35M+ Members
DashPass grew to over 35 million members, about half of monthly users, and drives retention and expansion into grocery and retail. DoorDash beat better-funded rivals like Uber and GrubHub through culture, speed and execution, and was losing $12 a delivery in San Francisco in 2015. The first version removed the roughly 15% service fee at a $20 minimum, lost money, but lifted order frequency almost 2% in an A/B test. Tuning the minimum to $15, the fee discount to about 10 points and merchant eligibility made it profitable for the August 2018 national launch, and it has stayed $9.99 a month.
DashPass started from customers saying delivery was too expensive. Jimmy Liu emailed every first-time DoorDash customer in San Francisco, asking what went well and what went poorly. Affordability came up again and again, alongside missing restaurants and bad deliveries.
He called these emails probably the single highest signal used to grow the market, and later did the same for first grocery orders. Customers described delivery as a luxury they couldn't afford every day.
Don't launch a subscription until each extra order is profitable. DoorDash waited until delivery economics turned positive, then picked test markets by sorting submarkets by profitability. Two years earlier, San Francisco was losing $10 to $12 an order on about 1,500 deliveries a day.
When each order is profitable, customers, DoorDash, merchants and Dashers all want more orders. "We can't have a situation where people ordering more actually causes us to lose more money."
A cheap delivery pack tested the idea before the subscription existed. The team shipped a pack of 10 deliveries for $20, or $2 each, in a week or two in Dallas and one other Midwest submarket. It showed people ordered more with discounted delivery.
The pack was much faster to build than a subscription and became one more data point, with customer talks and finance modeling, to justify building DashPass.
A four-to-five engineer pod shipped a hacky MVP in weeks. Jimmy carved one pod out of his own team and built DashPass in three or four weeks: unlimited free delivery over $20 with the whole roughly 15% service fee removed. It lost money on purpose to test one simple value proposition.
The COO said it took the best customers, made them unprofitable and got them to place more unprofitable orders. Jimmy feared being fired but knew several levers could fix the math: "Just build the thing, test it, and get real data."
An A/B test proved members ordered more, and the lift compounded. Executives worried heavy buyers were simply the ones signing up, so DashPass ran behind an A/B test. Order frequency rose almost 2% in two to three weeks, growing from 0.5 to 1% up to 2%, against typical tests that gain about 0.1% and flatten.
At about $4 billion in run-rate gross order value, a one or two percent lift was worth a great deal.
Three levers turned DashPass profitable: minimum, fee discount, eligible merchants. The order minimum dropped from $20 to $15 and the service fee discount was cut to about 10 of the 15 points. Merchants were included based on ticket size, commission rate and order accuracy, and the profitable version launched nationally in August 2018.
Customers had complained the $20 minimum was hard to hit. Restaurants with frequent missing items were excluded because inaccurate orders raise support costs. The final terms and member economics are not given in numbers.
Customer evidence plus modeling won over the skeptics; results converted them. The COO, a former Amazon executive, called DashPass a dumb idea, and the head of product also doubted it. Jimmy had talked to more customers about affordability than anyone and had built credibility first; months after launch the COO called DashPass the way DoorDash wins affordability.
Jimmy suggests building credibility for roughly six months at a new company before betting on yourself. "If you don't bet big, you can't win big."
Once DashPass caught fire, DoorDash made it a company-wide effort. The team grew from four or five engineers to over 100, and in annual planning every team was asked what DashPass benefit it could build. Members showed higher order frequency, retention and LTV.
When DashPass hit 10 million members faster than expected, Tony Xu raised the goal to 20 million, then 100 million. It exited 2025 with more than 35 million members, about half of monthly users.
Price stayed at $9.99 because compounding beats a one-time revenue bump. DashPass has cost $9.99 a month since 2018 while benefits grew. Jimmy says a $1 or $2 hike could lift revenue 10 to 20% overnight but would slow growth over 5 to 10 years.
This is his judgment, with no test shown; he notes no one runs an A/B test for 10 years, so some bets need conviction. He treats price as one of the last levers, since a higher fee makes the product worse.
Retention comes from one big saving plus many small features. Members save about 80% on fees, a large improvement on the core problem rather than 10 or 20%. On top of that, many small changes to sign-up, billing and discovery each add a point or two.
Jimmy says the resulting retention is causal, not just correlation.
Audit where people discover the membership and push it on big baskets. Walk the flows, especially checkout, and ask what share of customers even see the upsell. DoorDash promotes DashPass more loudly when a single order would more than pay for the $10 monthly fee, and stays quieter on small orders.
Talking to customers about placement and wording makes the upsell a win-win, since it shows them how they could have saved.
Free delivery pulls members into grocery and retail. About 75% of DoorDash's grocery and retail orders come from DashPass members. Free delivery lowers the risk of trying a new category for people already in the habit.
One app that solves more jobs stays top of mind versus separate apps.
Merchants joined because DashPass brought large, profitable orders. High-commission, large orders grew merchants' business, and reliable volume matters to restaurants with big fixed costs. Jimmy calls the merchant pitch the overlooked part of DashPass's growth.
Xu works customer support daily to see what data misses. He reads long, 2,000-word emails from Dashers, debugs them himself in internal tools and then calls or emails back. He argues data should win prioritization, but improving a product means fixing the tails where power users and new users show up.
"The greatest killer of a business is usually silence." Jimmy Liu kept a similar habit of emailing first-time customers.
Tony Xu of DoorDash: Surviving 1,000 Days of Startup Hell
Started in 2013 as a 43-minute, $9 static page with PDF menus, run by four Stanford founders who made every delivery themselves. It found suburbs like Palo Alto faster to serve, runs tens of thousands of experiments with about 95% failing, and survived over 100 investor rejections in 2016. It now also offers fulfillment, merchant tools and its own Dot delivery vehicle.
Restaurants came first because they offered the densest network. The founders compared restaurants, grocery, convenience and retail. About a million US restaurants versus a couple hundred thousand grocery stores made prepared meals the best start for a network that could later carry anything.
Only 20,000 to 25,000 restaurants delivered at the time, mostly pizza; existing services just faxed orders. Keith Rabois backed DoorDash after the founders' stat that 93% of US restaurants didn't deliver. The idea came from a one-person baker's binder of refused delivery orders, found while interviewing about 300 businesses.
The YC summer tested three unit-economics questions, not demo day. In summer 2013 the founders only asked whether consumers would pay $6, whether restaurants would partner for 15% and whether drivers could earn a workable wage.
Rabois makes the same point: validate the mechanics, such as deliveries per hour to break even and enough density, rather than asking people if they want the product.
Test candidates with real tasks and check references until one is negative. DoorDash's final interview gave candidates $20 and eight hours to acquire 100 customers, and engineers did deliveries with Xu in his Honda. Xu does 20 references on every senior hire.
Christopher Payne, later COO, drove deliveries for four hours with his son after one conversation, then wrote a 30,000-word email on why the logistics algorithm was flawed.
Hard truths about building in the AI era | Keith Rabois (Khosla Ventures)
Rabois backed it after the founders' stat that 93% of US restaurants don't deliver, and Andrew Mason's line that phones need an 'I'm bored' and an 'I'm hungry' button made it click. Tony does 20 references on every senior hire.
A 43-minute, $9 page tested whether anyone wanted delivery. The four Stanford founders built a static page with eight PDF menus, a Google Voice number that rang all their phones and a Square reader for payment. They suspected delivery from non-delivering restaurants might not exist because nobody wanted it.
Spending no money on marketing, they wanted organic demand and watched whether Tony Xu's bank account drained. It did not.
Doing every delivery themselves revealed what the business really was. With no logistics experience, the founders took orders and drove the food for roughly six months, at about 10 orders a day and a best day of 21. They found a delivery has about 20 steps, each with delays, and needs a consumer site, restaurant app, driver app and dispatch system.
Delivering by hand also showed the typical customer was a busy mom with young children. "You actually have no idea what are all the sources of delay in an order until you actually go and do the work."
Suburbs beat dense cities on both speed and demand. Deliveries were faster in Palo Alto than in San Francisco thanks to easy parking, single-family homes and no lobbies or elevators. Stanford's nearest restaurants were 2 miles away, so customers needed delivery more.
Palo Alto's hub-and-spoke layout matches most US towns. DoorDash went after suburbs with little competition while Postmates spread across cities and categories.
Restaurants came first because they offered the densest network. The founders compared restaurants, grocery, convenience and retail. About a million US restaurants versus a couple hundred thousand grocery stores made prepared meals the best start for a network that could later carry anything.
Only 20,000 to 25,000 restaurants delivered at the time, mostly pizza; existing services just faxed orders. Keith Rabois backed DoorDash after the founders' stat that 93% of US restaurants didn't deliver. The idea came from a one-person baker's binder of refused delivery orders, found while interviewing about 300 businesses.
The YC summer tested three unit-economics questions, not demo day. In summer 2013 the founders only asked whether consumers would pay $6, whether restaurants would partner for 15% and whether drivers could earn a workable wage.
Rabois makes the same point: validate the mechanics, such as deliveries per hour to break even and enough density, rather than asking people if they want the product.
Higher pay barely moved Uber X drivers, proving Dashers are different people. Fearing drivers only chased money, Xu offered 20 DoorDash and 20 Uber X drivers a guaranteed $25 an hour against their roughly $20. Only one of the 40 switched.
Today more than half of Dashers are women, the average Dasher works three to four hours a week, and 90% drive under 10 hours.
DoorDash refunded every late order despite weeks of cash left. After a Stanford football game in September 2013, every delivery ran at least an hour late. Nobody asked, but the founders refunded everyone at about 40% of a bank account holding two or three weeks of cash, then delivered cookies at 5 a.m.
"We'd rather die trying to be excellent than to live to be mediocre." Xu took from it that the scoreboard resets to zero each day: customers judge every order on selection, price, speed, accuracy and problem handling.
Xu works customer support daily to see what data misses. He reads long, 2,000-word emails from Dashers, debugs them himself in internal tools and then calls or emails back. He argues data should win prioritization, but improving a product means fixing the tails where power users and new users show up.
"The greatest killer of a business is usually silence." Jimmy Liu kept a similar habit of emailing first-time customers.
Run thousands of experiments; about 5% survive and compound. Because the physical world is undocumented and always changing, DoorDash runs tens of thousands of experiments and roughly 5% reach customers. Hacky tests become products, then engineered systems.
At millions of orders a day, one-in-a-million events happen often, so the answer is prevention plus fast emergency response. "It's always the data that you can't see that kills you."
New bets earn funding stage by stage, like internal venture capital. Anyone can propose an idea, and if an experiment shows customers want it, it enters planning with a small budget that grows only as it solves the problem. Xu modeled this on DoorDash's own scarce early years, when its best products were built.
He runs two operating systems: the core business as a big airplane getting a midair engine transplant, and new ventures as paper airplanes with their own measures, talent, timelines and incentives.
DoorDash built its own delivery robot when partners wouldn't. Partners focused on robotaxis and nobody wanted to build a small last-mile vehicle, so DoorDash started its own project in 2019. About six years later it shipped the Dot, which runs on roads, sidewalks and bike lanes in Phoenix and Scottsdale.
"We don't have a religion about whether or not we have to build the product or someone else has to build the product."
Local data and logistics aim to make DoorDash any merchant's first call. DoorDash structures physical-world data to flag out-of-stocks, underpriced items and bundles, runs pricing tests for merchants, and operates warehouses and fulfillment for Kroger and CVS.
The idea was seeded in 2014 when a California farm asked if DoorDash could run its trucks. Xu grounds it in needs that never change: selection, price, speed and convenience.
Xu focused on controllable metrics while over 100 investors said no. DoorDash's Series C collapsed in 2016 as public markets fell, and it raised a fraction of what peers did for about three years. Xu showed every metric, including the falling cash balance, at all-hands and rallied 20 to 25 leaders on one goal: take share, get more profitable, don't run out of cash.
He stopped counting rejections after 50. Unable to outspend rivals, DoorDash had to win on retention and engagement. He still ignores the stock price, since he can't control it.
Test candidates with real tasks and check references until one is negative. DoorDash's final interview gave candidates $20 and eight hours to acquire 100 customers, and engineers did deliveries with Xu in his Honda. Xu does 20 references on every senior hire.
Christopher Payne, later COO, drove deliveries for four hours with his son after one conversation, then wrote a 30,000-word email on why the logistics algorithm was flawed.
The hierarchy of engagement | Sarah Tavel (Benchmark, Greylock, Pinterest)
It started narrow in the suburbs with little competition and lost money early on long drives, yet kept both sides happy enough to retain. It later raised heavily against Uber and overtook Grubhub.
Suburbs beat dense cities on both speed and demand. Deliveries were faster in Palo Alto than in San Francisco thanks to easy parking, single-family homes and no lobbies or elevators. Stanford's nearest restaurants were 2 miles away, so customers needed delivery more.
Palo Alto's hub-and-spoke layout matches most US towns. DoorDash went after suburbs with little competition while Postmates spread across cities and categories.
Supplying its own delivery fleet expanded selection beyond Grubhub's network. DoorDash changed the atomic unit of supply by delivering for any restaurant, including ones that had never delivered or had no agreement. Grubhub stopped being comprehensive, and DoorDash raised huge rounds to lock in national chains.
Grubhub had raised $80 million and $100 million in its IPO while challengers raised $100 million every six months. The pandemic turned DoorDash's margins positive for the first time. It runs a heavily managed marketplace that owns delivery logistics.
Grubhub assumed money-losing delivery would fail and didn't overreact. Grubhub expected negative-margin rivals to lose their subsidies, but they kept raising and won. Casey Winters says Grubhub could probably have acquired DoorDash multiple times early on.
Sarah Tavel adds that Grubhub, as a public company, was slow to fulfill delivery itself and couldn't access capital like private rivals. "The only rational reaction is overreaction."
Thinking beyond frameworks | Casey Winters (Pinterest, Eventbrite, Airbnb, Tinder, Reddit, Grubhub)
It started in less dense suburbs with restaurants that had never delivered and raised huge rounds to sign national chains. The pandemic turned its margins positive for the first time.
Suburbs beat dense cities on both speed and demand. Deliveries were faster in Palo Alto than in San Francisco thanks to easy parking, single-family homes and no lobbies or elevators. Stanford's nearest restaurants were 2 miles away, so customers needed delivery more.
Palo Alto's hub-and-spoke layout matches most US towns. DoorDash went after suburbs with little competition while Postmates spread across cities and categories.
Supplying its own delivery fleet expanded selection beyond Grubhub's network. DoorDash changed the atomic unit of supply by delivering for any restaurant, including ones that had never delivered or had no agreement. Grubhub stopped being comprehensive, and DoorDash raised huge rounds to lock in national chains.
Grubhub had raised $80 million and $100 million in its IPO while challengers raised $100 million every six months. The pandemic turned DoorDash's margins positive for the first time. It runs a heavily managed marketplace that owns delivery logistics.
Grubhub assumed money-losing delivery would fail and didn't overreact. Grubhub expected negative-margin rivals to lose their subsidies, but they kept raising and won. Casey Winters says Grubhub could probably have acquired DoorDash multiple times early on.
Sarah Tavel adds that Grubhub, as a public company, was slow to fulfill delivery itself and couldn't access capital like private rivals. "The only rational reaction is overreaction."
Developing a growth model + marketplace growth strategy | Dan Hockenmaier
It is a heavily managed marketplace that does work in the value chain, such as owning delivery logistics.
Supplying its own delivery fleet expanded selection beyond Grubhub's network. DoorDash changed the atomic unit of supply by delivering for any restaurant, including ones that had never delivered or had no agreement. Grubhub stopped being comprehensive, and DoorDash raised huge rounds to lock in national chains.
Grubhub had raised $80 million and $100 million in its IPO while challengers raised $100 million every six months. The pandemic turned DoorDash's margins positive for the first time. It runs a heavily managed marketplace that owns delivery logistics.
More in Food & Drink
Food & DrinkYelp
A local business review and discovery app for restaurants, services and shopping.
2 sources · 14 tips
Food & DrinkGrubHub
A food delivery marketplace that connected diners with restaurants and took payments.
4 sources · 9 tips
Food & DrinkMcDonald's
A global fast food chain built on franchised restaurants and company-owned real estate.
3 sources · 13 tips
Food & DrinkCoca-Cola
A global beverage company that sells cola syrup to franchised bottlers and builds its brand through advertising.
2 sources · 16 tips
Food & DrinkStarbucks
A global coffeehouse chain selling espresso drinks, beans and food, with a large mobile ordering app.
2 sources · 15 tips
Food & DrinkRaising Cane's
A chicken finger restaurant chain built on one craveable menu and company-owned locations.
2 sources · 17 tips
Food & DrinkWhole Foods Market
A natural and organic grocery chain founded in Austin, Texas, later bought by Amazon.
2 sources · 18 tips
Food & DrinkInstacart
A grocery delivery app with shoppers fulfilling orders.
6 sources · 9 tips