Iteration vs Incrementalism vs En Masse

Iteration = Build=>Measure=>Learn=>Decide (Letting the ideal Emerge based on a pre-defined outcome)

Incrementalism = Design=>DecomposeIntoShippableParts=>BuildFirstPart=>BuildSecondPart=>BuildThirdPart… so on until design is complete (Converging on a pre-defined ideal)

En masse = Design=>BuildEntireDesign=>Ship (binary… pre-defined ideal and done)

Iteration = Great for lean, agile, learning-focused cultures where the value hypothesis is only validated by outcome data

Incrementalism = Great for teams that believe their pre-validation is so good they don’t need to measure to see if it actually worked… but they do want to deliver this “predetermined value” continuously

En masse = Great for teams that believe their pre-validation techniques completely and care more about efficiency than continuous value delivery. It is all about the Big Bang approach.

All three are fine approaches based on your context.

However, most SaaS companies will find exponential success by striving for iteration.

I worry about the over-confidence in the pre-validation of the other two techniques.

I’d rather approach everything with a mindset of “I’m not sure, let’s find out”.

—-

Appendix

I recently ran across this article that has an accompanying image that shows iteration vs incrementalism really well.

Rule of thumb (when to use what)

• If dealing in days (< 9 biz day appetite/estimation)… likely use En Masse
• If dealing in weeks (< 5 week appetite/estimation)… likely use Incrementalism
• If dealing in months… MUST use Iteration

However…

• … in all cases we should have a desired outcome
• … in all cases we should proceed with humility and a curious “I wonder” posture aka Emergent
• … therefore in all cases there should be an iteration-based outcome at the root… there may be En Masse and Incremental tactics we use within the iteration… but if we aren’t iterating, in SaaS, we are leaving precious learning on the table, and learning should be pursued at all times.