The Amount of Work To Replace a Tech Stack
Why are most smaller tech companies most likely to use better technology than far larger, more financed competitors?
This is one of my favourite charts. Note — it’s quite long.
Some of these products/services rely on such a colossal amount of pre-existing infrastructure to operate, that to consider replacing them would translate to building the company again. I think this touches on why smaller firms, like start-ups, are often at the cutting edge of tech because they’re not bound to anything initially.
But as any start-up grows, it indulges in an amount of technical debt to maintain operational proficiency which inevitably evolves into a form of “backwardness”. Eventually a question will be asked — do we update the tech stack for modernity?
I was thinking about how to formalise the amount of work required to replace a tech stack. Why do firms like Facebook still use a highly-modified version of PHP or the US nuclear missile system still rely on floppy disks?
I think there are a few key components:
- The more employees, the harder it is to detach the company from the pre-existing infrastructure. Changes to the architecture of operations become exponentially harder when more people are involved, even more so when they’re overseas, different languages etc.
- The overall reward can be defined by monthly active users. 0 MAU suggests no reward in changing the tech stack in the first place, whereas say 100k MAU means upgrading a system in a way that doesn’t affect a large volume of traffic.
- Because the work outputted by a company is defined by its team, then the company needs to be defined by the employees’ relationships to each other which is n-squared, when n is the number of employees.
This can be formalised as:
The amount of work, in a scalar value, required to replace a tech stack (R) is equivalent to the monthly-active-users (MAU) divided by the number of employees squared (n²).
That is, R = MAU/n².
Some examples:
- R(Netflix) = 182 million/9,40⁰² = 2.01
- R(Slack) = 15 million/2,04⁵² = 3.57
- R(Facebook) = 2.6 billion/58,604² = 7.57
- R(Spotify) = 345 million/4,405² = 17.77
- R(Gmail) = 1.8 billion/10,00⁰² = 18.00
- R(TikTok) = 1 billion/6,500² = 23.66
- R(ClubHouse) = 600k/12² = 4,166.67
The value for R is a scalar value that’s indicative of the amount of work required to replace the entire tech stack. Intuitively, the amount of work becomes less with more employees or less active users.
Thanks for reading,
Bryan
