4 ms·

Instacart Is Generating Profits?

24 points by rafaelc 11y ago

vannevar 11y ago
There's a lot of discussion in the article about AI and improved efficiencies, but I wonder how much is due to real improvement and how much is simply the result of increasing prices and cutting pay, both of which are also cited.
- pbreit 11y ago
  Yeah, it seems like they trotted out the "data scientist" to provide some smoke an mirrors. 20% efficiency increase vs 50% labor cost improvement?
tryitnow 11y ago
Interesting. I'd love to see their financial data. The problem with private companies is that they can say pretty much whatever they want and there's no way to do a "smell test" by looking at audited GAAP numbers. It does make sense though that they should realize economies of scale (which is probably the driving force in their margin improvements despite all the talk of 'date science').
pbreit 11y ago
I thought Primack was smarter than this. Seems like he might have been played here. Even at $7/order (apparently its best market), I'd be surprised if that covers variable costs, even in a market denser than Atalanta.
draw_down 11y ago
Before this I didn't know that they charged subscription fees. I have toyed with the idea of using Instacart, but I wouldn't pay a subscription. I don't really see the point of a subscription fee for a retail service. I give you the money, you give me the thing, end of transaction, right?
- jasonpeacock 11y ago
  You don't have to be a subscriber to use Instacart.
- tholmes 11y ago
  Last time I did the math you have to order pretty regularly for the subscription to break even. I use instacart about once a month and just pay the delivery fee.
- pkroll 11y ago
  You can just use it and pay a fee every time, or you can subscribe and get Free Delivery... of course they still charge more than most of the stores for the groceries themselves. I subscribed last year for both myself and so I could send my mom groceries, and barely, barely broke even: 25 orders. Given the price increases since then, you'd still need 25 orders to break even.
rw2 11y ago
This is not suprising. There are grocery delivery startups like freshdirect in NYC that has much lower delivery cost and they are doing delivery from outside of the city instead of inside. Granted, they have daily windows, instead of realtime delivery. Long term, with enough customers, it's easy to cut down costs with a variety of different ways.
Dwolb 11y ago
Instacart are thinking very well about their business operations well but I still am apprehensive on their unit economics. The numbers given are $5.99 for a single delivery fee or $149 for an annual subscription. In Atlanta they're making $6.96 per order and in Chicago they're making $4.29. Let's say a shopper makes $15/hr and can pick 5 orders per hour with an average basket size of $100. The delivery driver also makes $15/hr and can deliver all 5 orders with the basket size of $100. Total labor cost is $30, or $6 per delivery and total revenue is 5*$5.99 or $29.95 for a slightly negative gross profit. So then is it reasonable to pick and checkout more than 5 orders per hour or 1 every 12 minutes in a random, non-optimized grocery store layout in the US? Is it reasonable that a grocery store would pay out 6% off topline revenue to Instacart when grocery industry profits are already in the single digits? Could it be the average order revenue is substantially higher than $5.99 if annual subscribers use their service fewer than once every two weeks? Or is their cost basis truly lower than fully burdened employee costs of $15/hr? It just seems like the economics are pretty sensitive to operational metrics so it'd be great to get an official view into the business.
- cehrnrooth 11y ago
  Location density of the deliveries has got to have a massive impact on their effectiveness. I'm highly skeptical of their ability gather groceries and deliver to 5 separate locations within 1 hour. Even in a dense area where it's 5 minutes between each location (plus parking, plus entry / exit and delivery) that's 25 minutes minimum. The longer the time, the lower the $/Hr. The further away the locations are, the higher the transportation costs are for the driver, further cutting into their actual take home pay (not including, miles, depreciation, insurance, etc...). Instead of making deliveries after the order (and trying to bundle orders and deliveries), I would think a dispatch model where one set of workers go out in the morning and make the orders and deliver to a central location. Then have a second shift make deliveries, but that would make it difficult to maintain the independent contract status of their workers.
  - Dwolb 11y ago
    Agrees on the location density point - the route for the driver has got to be near perfect for their model to work. The dispatch system is an interesting thought. I wonder if it fully aligns with their delivery time promise. I think the dispatch system points to another 'lever' however - certain recurring orders can be lazily filled. i.e. when the shopper has no rush orders he/she can fill her free time filling orders which are upcoming, but not immediate
yueq 11y ago
The key issue really is Instacart doesn't have pricing power of the groceries. This will limit their potential to be profitable.

24 points by rafaelc 11y ago

There's a lot of discussion in the article about AI and improved efficiencies, but I wonder how much is due to real improvement and how much is simply the result of increasing prices and cutting pay, both of which are also cited.

pbreit 11y ago
Yeah, it seems like they trotted out the "data scientist" to provide some smoke an mirrors. 20% efficiency increase vs 50% labor cost improvement?

tryitnow 11y ago

Interesting. I'd love to see their financial data. The problem with private companies is that they can say pretty much whatever they want and there's no way to do a "smell test" by looking at audited GAAP numbers. It does make sense though that they should realize economies of scale (which is probably the driving force in their margin improvements despite all the talk of 'date science').

pbreit 11y ago

I thought Primack was smarter than this. Seems like he might have been played here. Even at $7/order (apparently its best market), I'd be surprised if that covers variable costs, even in a market denser than Atalanta.

draw_down 11y ago

Before this I didn't know that they charged subscription fees. I have toyed with the idea of using Instacart, but I wouldn't pay a subscription. I don't really see the point of a subscription fee for a retail service. I give you the money, you give me the thing, end of transaction, right?

jasonpeacock 11y ago
You don't have to be a subscriber to use Instacart.
tholmes 11y ago
Last time I did the math you have to order pretty regularly for the subscription to break even. I use instacart about once a month and just pay the delivery fee.
pkroll 11y ago
You can just use it and pay a fee every time, or you can subscribe and get Free Delivery... of course they still charge more than most of the stores for the groceries themselves. I subscribed last year for both myself and so I could send my mom groceries, and barely, barely broke even: 25 orders. Given the price increases since then, you'd still need 25 orders to break even.

rw2 11y ago

This is not suprising. There are grocery delivery startups like freshdirect in NYC that has much lower delivery cost and they are doing delivery from outside of the city instead of inside. Granted, they have daily windows, instead of realtime delivery. Long term, with enough customers, it's easy to cut down costs with a variety of different ways.

Dwolb 11y ago

Instacart are thinking very well about their business operations well but I still am apprehensive on their unit economics. The numbers given are $5.99 for a single delivery fee or $149 for an annual subscription. In Atlanta they're making $6.96 per order and in Chicago they're making $4.29. Let's say a shopper makes $15/hr and can pick 5 orders per hour with an average basket size of $100. The delivery driver also makes $15/hr and can deliver all 5 orders with the basket size of $100. Total labor cost is $30, or $6 per delivery and total revenue is 5*$5.99 or $29.95 for a slightly negative gross profit. So then is it reasonable to pick and checkout more than 5 orders per hour or 1 every 12 minutes in a random, non-optimized grocery store layout in the US? Is it reasonable that a grocery store would pay out 6% off topline revenue to Instacart when grocery industry profits are already in the single digits? Could it be the average order revenue is substantially higher than $5.99 if annual subscribers use their service fewer than once every two weeks? Or is their cost basis truly lower than fully burdened employee costs of $15/hr? It just seems like the economics are pretty sensitive to operational metrics so it'd be great to get an official view into the business.

cehrnrooth 11y ago
Location density of the deliveries has got to have a massive impact on their effectiveness. I'm highly skeptical of their ability gather groceries and deliver to 5 separate locations within 1 hour. Even in a dense area where it's 5 minutes between each location (plus parking, plus entry / exit and delivery) that's 25 minutes minimum. The longer the time, the lower the $/Hr. The further away the locations are, the higher the transportation costs are for the driver, further cutting into their actual take home pay (not including, miles, depreciation, insurance, etc...). Instead of making deliveries after the order (and trying to bundle orders and deliveries), I would think a dispatch model where one set of workers go out in the morning and make the orders and deliver to a central location. Then have a second shift make deliveries, but that would make it difficult to maintain the independent contract status of their workers.
- Dwolb 11y ago
  Agrees on the location density point - the route for the driver has got to be near perfect for their model to work. The dispatch system is an interesting thought. I wonder if it fully aligns with their delivery time promise. I think the dispatch system points to another 'lever' however - certain recurring orders can be lazily filled. i.e. when the shopper has no rush orders he/she can fill her free time filling orders which are upcoming, but not immediate

yueq 11y ago

The key issue really is Instacart doesn't have pricing power of the groceries. This will limit their potential to be profitable.