Unprofitable growth & stacked S-curves

The stacked S-curve is often used to describe technological growth.
Each technology has an early period of making no progress, an inflection period where it makes a lot of progress, and then a leveling off once it has been saturated.
Even though each individual technology eventually levels off - another technology comes in to continue growth.
notion image
There’s an interesting version of this framework for understanding unprofitable VC-backed companies, and how spending capital on growth can work.
The framework: unbundle businesses into their component segments (ie: by geography, customer, or whatever makes sense), and place them along an S-curve.
This graph looks a lot messier, as companies have many segments running simultaneously.
notion image
Successful businesses thus filter to the bets (S-curves) that have higher long-run EV (ie: higher net revenue) and get there faster (ie: shorter build-up time). They ideally have a healthy pipeline of many S-curves building up.
The following diagram takes a snapshot of various segments at a given point in time:
notion image
Consider a hypothetical ridesharing company, Ultra. They lost $X/year, an average of $Y per ride. But, consider disaggregating Ultra by city.
Starting out in a new city is expensive, you need to develop network effects, usually by initially subsidizing users and drivers.
notion image
For a more quantitatively backed example, consider Palantir - they sell software to government entities to help them manage their data.
They split their customer pipeline into acquire/expand/scale phases, based on how deeply they’re working with a customer. Here are some margin numbers from 2019 (via Byrne Hobart).
notion image
The source of unprofitability for many later-stage VC-backed businesses, then is:
  • Expanding more quickly involves beginning more S-curves, and temporarily increases the ratio of unprofitable to profitable segments.
  • This improves growth in the long-run, as you’re mostly paying down the cost to get into a segment.
The causality of fundraising may be backwards from what you expect, ie: companies are unprofitable because they know they can raise, not vice versa.
The rate of beginning new bets is a growth vs margin lever, and companies turn it all the way towards growth.
Misc notes:
Other bets can be modeled as S-curves
A new feature or product vertical is unprofitable for a while (ie: takes many engineering & product hours, results in little new revenue), and then may be very profitable later on.
You could model making a less ambitious product as starting with fewer S curves.
Many S-curves are just not worth going down
  • A failure of capital-intensive growth might be overeagerness to begin S-curves that ultimately still end up in the unprofitable territory.
    • You might delude yourself into believing you’re investing in good S-curves, to eventually find out that you’re not.
    • Conversely: a failure of capital-light growth might be failing to capitalize on compelling sectors early enough to capture the market.
  • The availability of new S-curves is very contextual. Are these new potential opportunities within your strengths or not?
Other mechanisms for profit-growth tradeoff
Stacked S-curves are one mechanism to trade-off profit for growth, but there are others. Classically, one mechanism is to capture very little value early on, and give most of it to your users.
eg: A ridesharing company may offer cheaper rides earlier, and take a lower cut, for more growth and then increasing it later on.
 

Unprofitable growth & stacked S-curves

The stacked S-curve is often used to describe technological growth.
Each technology has an early period of making no progress, an inflection period where it makes a lot of progress, and then a leveling off once it has been saturated.
Even though each individual technology eventually levels off - another technology comes in to continue growth.
notion image
There’s an interesting version of this framework for understanding unprofitable VC-backed companies, and how spending capital on growth can work.
The framework: unbundle businesses into their component segments (ie: by geography, customer, or whatever makes sense), and place them along an S-curve.
This graph looks a lot messier, as companies have many segments running simultaneously.
notion image
Successful businesses thus filter to the bets (S-curves) that have higher long-run EV (ie: higher net revenue) and get there faster (ie: shorter build-up time). They ideally have a healthy pipeline of many S-curves building up.
The following diagram takes a snapshot of various segments at a given point in time:
notion image
Consider a hypothetical ridesharing company, Ultra. They lost $X/year, an average of $Y per ride. But, consider disaggregating Ultra by city.
Starting out in a new city is expensive, you need to develop network effects, usually by initially subsidizing users and drivers.
notion image
For a more quantitatively backed example, consider Palantir - they sell software to government entities to help them manage their data.
They split their customer pipeline into acquire/expand/scale phases, based on how deeply they’re working with a customer. Here are some margin numbers from 2019 (via Byrne Hobart).
notion image
The source of unprofitability for many later-stage VC-backed businesses, then is:
  • Expanding more quickly involves beginning more S-curves, and temporarily increases the ratio of unprofitable to profitable segments.
  • This improves growth in the long-run, as you’re mostly paying down the cost to get into a segment.
The causality of fundraising may be backwards from what you expect, ie: companies are unprofitable because they know they can raise, not vice versa.
The rate of beginning new bets is a growth vs margin lever, and companies turn it all the way towards growth.
Misc notes:
Other bets can be modeled as S-curves
A new feature or product vertical is unprofitable for a while (ie: takes many engineering & product hours, results in little new revenue), and then may be very profitable later on.
You could model making a less ambitious product as starting with fewer S curves.
Many S-curves are just not worth going down
  • A failure of capital-intensive growth might be overeagerness to begin S-curves that ultimately still end up in the unprofitable territory.
    • You might delude yourself into believing you’re investing in good S-curves, to eventually find out that you’re not.
    • Conversely: a failure of capital-light growth might be failing to capitalize on compelling sectors early enough to capture the market.
  • The availability of new S-curves is very contextual. Are these new potential opportunities within your strengths or not?
Other mechanisms for profit-growth tradeoff
Stacked S-curves are one mechanism to trade-off profit for growth, but there are others. Classically, one mechanism is to capture very little value early on, and give most of it to your users.
eg: A ridesharing company may offer cheaper rides earlier, and take a lower cut, for more growth and then increasing it later on.