Fabio Manganiello on Nostr: npub19j7hd…3e8r4 and the problem lies at the root of the business model. Products ...
npub19j7hdrzu4v5cndyl4ry2eelq6tdn75s04z5hld8w3cu63lugqrsq53e8r4 (npub19j7…e8r4) and the problem lies at the root of the business model. Products nowadays are in constant MVP mode - “produce something that we can show in a nice demo to the stakeholders by the end of the quarter, deliver it in prod, and we’ll figure out later how to scale it up”.
Tech debt is relegated to a dusty corner of the backlog that seldom gets prioritized.
Testing things E2E in different environments (and provisioning a good developer experience, so other engineers can easily replicate and provision things and have test data to play with) becomes a nice to have that almost nobody ever gets.
Infrastructure-as-code, containers, pods etc. were supposed to make infrastructure easy to replicate, but in the hands of short-termist managers they’ve ended up making things worse - by deploying brittle configurations without a whole view on the whole architecture and its replicability, and eventually they have only exacerbated the “it works on my machine/test it in production” problem.
Making things better through proper engineering/automation solutions gets relegated to the post-MVP phase (which usually never happens), and until then PMs operate under the extremely misguided assumption that adding a new wiki page / shared doc/spreadsheet that a human must read in order to understand the process is cheaper than spending a couple of days/weeks actually making the process better.
Even proper user testing is often neglected, unless something is really a crucial user-facing feature, as there usually isn’t enough time for it in permanently understaffed teams that have the responsibility of developing and deploying everything in unrealistically strict timelines.
I’ve seen this pattern in enough companies to believe by now that it’s not a company-specific issue - it’s an industry-wide problem that affects any business once it goes past the MVP phase.
Tech debt is relegated to a dusty corner of the backlog that seldom gets prioritized.
Testing things E2E in different environments (and provisioning a good developer experience, so other engineers can easily replicate and provision things and have test data to play with) becomes a nice to have that almost nobody ever gets.
Infrastructure-as-code, containers, pods etc. were supposed to make infrastructure easy to replicate, but in the hands of short-termist managers they’ve ended up making things worse - by deploying brittle configurations without a whole view on the whole architecture and its replicability, and eventually they have only exacerbated the “it works on my machine/test it in production” problem.
Making things better through proper engineering/automation solutions gets relegated to the post-MVP phase (which usually never happens), and until then PMs operate under the extremely misguided assumption that adding a new wiki page / shared doc/spreadsheet that a human must read in order to understand the process is cheaper than spending a couple of days/weeks actually making the process better.
Even proper user testing is often neglected, unless something is really a crucial user-facing feature, as there usually isn’t enough time for it in permanently understaffed teams that have the responsibility of developing and deploying everything in unrealistically strict timelines.
I’ve seen this pattern in enough companies to believe by now that it’s not a company-specific issue - it’s an industry-wide problem that affects any business once it goes past the MVP phase.