Testing recipes is hard!
Testing of recipes is often time bound, requires precision but is also subject to analytical machines/labs and sensory panels.
Sensory panels are expensive to setup and are often the last step in validating a recipe before taking to scale up trials, factory manufacturing and eventually rollout.
Turing helps with making Testing easier and faster and more efficient.
Informal Testing
Informal Testing is one such unique capability in the Turing platform that allows for rapid iterative testing that reduces need on expensive sensory panels and leverages the formulators unique expertise to making fast progress.
Core Workflow
Informal Testing is designed inline with the core Turing workflow
1. Setup project - Identify inputs, outcomes, goals goals priority and constraints
2. Calibrate - Ensure all goals are defined, apporpriately prioritized and all requires constraints are in place
3. Generate Initiative - Recieve a list of suggested recipes that maximizes learning potential
4. Run tests and Upload Test Results - Download recipes, make outcome measurements and upload measurements
5. View Progress, Finalize Project or Generate another round of recipes to test. -Use analytical tools to decide if goals have been met or testing needs to continue
How it works?
Formulators can determine that for a round of testing, whether they will use Formal or Informal Testing methodology. Informal Testing could be done by the formulator, their peers or a combination. Formal Testing often requires testing by an independent, highly trained sensory panel (which are harder to budget for , to setup and train and execute)
1. For each outcome defined in the project and part of the test plan, determine testing strategy (Informal or Formal)
2. If Formal, ensure that the sensory panel is trained on the rating scale required for the outcomes as defined in the project
3. If Informal, ensure that the informal testers are aware of the acceptable values for Informal Testing
The above is a sample of what your test results data would look like.
1. For each outcome, determine the "Turing_m_type" as either "formal" or "informal"
2. For each outcome, define the "Turing_confidence" as either "high" or "low"
3. For each outcome, if the "Turing_m_type" is defined as formal, ensure results are recorded according the measurement scale defined in the project scope
4. If the "Turing_m_type" is defined as informal, the acceptable values for the outcome measurement are
"ok", "very low", "high", "very high"
5. Note that any other values apart from the ones defined above or spelling mistakes will cause the test results to be rejected during upload or cause inaccuracies in the processing and subsequent results.
6. As you can see in the example, you can mix informal and formal testing,
7. Turing also supports Missing Outcomes i.e you can choose NOT measure a particular outcome in a round of testing and leave it blank