From scratch or improvement: what to choose for a bot or service

The from-scratch vs improvement choice should follow risk and timeline - not team habit. Sometimes rewriting one part is cheaper than patching a weak foundation for a year.

Choose improvement if

• The architecture is clear and runs locally
• Payments and access already work reliably
• You need focused features, not a product-model change

In that case improvement gives speed: you reuse a proven shell and do not pay to rebuild everything.

Choose from scratch if

• The code is unsafe to change
• There are no tests, logs or release control
• You need a different architecture for realtime, multi-exchange or a cabinet

From scratch does not always mean longer. On a fragile fork the timeline often disappears into unpredictable bugs and sales stoppages.

A practical compromise

Sometimes you keep the working shell and rewrite only the bottleneck: billing, collectors or admin. That way you do not pay twice for everything at once and get a controllable product sooner.