An iterative process is useful when the output of one iteration multiplies the value of the next iteration. So, kicking a small prototype to market gives you insiight into how it's used, which mulltipies the vaue of the next prototype. That's why iterating quickly is important, even small amounts of incremental knowledge have a large effect, because they're being multiplied so frequently.

But if you slow down for any reason, like unpredicted production delays or the need to wait for market feedback, then iteration becomes a liabillity because all your eggs are in one basket.

Parallel processes make more sense when:

• there can be unpredictible production delays, slowing down iterations
• there's a lead-time or delay in getting feedback
• there are a range of options to narrow down quickly

Parallel process require:

• different functions and roles on the team to prototype widely, and communicate the trade-offs between them to everyone else
• a decision-making attitude of elimination, rather than selection
• strong honour of commitments when agreeing to eliminate options