The cheaper the first prototype becomes, the easier it is to accumulate cost around a product with no validated value. STOP is therefore a valid outcome, not a failure.
Prototype · pilots
App Factory: validate the idea before more build cost
AI makes it possible to start building almost immediately — which also makes it easy to build a weak idea before anyone knows whether the market needs it. In App Factory every stage ends in a real decision: build, pilot, reshape or stop.
Case at a glanceDefine the gates between stages and the criteria for BUILD, PILOT, RESHAPE or STOP.
Several ideas have already been stopped after research; one product reached distribution setup. The next stage is real-device and audience validation.
01Problem→02Idea check→03Pilot→04Test→05Decision
A fast prototype does not automatically justify the next stage. Each additional spend needs new evidence.
The cheaper building becomes, the more important it is to know when not to build
AI dramatically lowers the cost of the first prototype but does not prove product value. That is why BUILD is not the only valid outcome: PILOT, RESHAPE and STOP are equally legitimate stage decisions. Further development has to be justified by new evidence, not by sunk effort.
Key finding0201 02 03 04 05
Decision before the next stage
After each stage, the product can continue, enter a pilot, change direction or stop.
Build
There is enough evidence to create a limited working version.
Pilot
The product needs a controlled real-world test.
Validate further
More evidence is needed before the next decision.
Reshape
The problem still looks real, but the current solution needs to change.
Stop
There is not enough evidence to justify more build work.
What must be clear before the next stage
A fast prototype only proves that something can be built. The next step needs evidence that the problem is real and worth more investment.
What has already happened
The process has already been used for real decisions across several product ideas.
A product reached distribution setup
At least one product reached distribution setup and release preparation.
Several ideas were stopped
Research showed insufficient evidence to continue some initially attractive concepts.
Further checks
Some products still need hands-on checks on real devices and browsers.
How the factory itself was simplified
Internal machinery was removed when maintaining it stopped helping product validation.
Before
The internal setup had grown large enough to compete with product work.
After
Only the tools needed for the current product check were kept.
What is checked before release
Before release, the workflow checks user value, technical readiness, security, privacy, physical product behavior, rollback readiness and distribution requirements.
Context & measurement
- Context
- Personal research project
- Scale / period
- Prototype · pilots · product validation in progress
Next step
Complete hands-on validation of the current products and make the next product decision.
AI products
Facing a similar challenge?
Tell me where SEO is blocked by product, engineering or data.