THE FINAL REFRACTION OF A PROPOSAL
One bid enters.
A decision system receives it.
A good proposal can be celebrated in one evaluation room and struggle in another—even when the published rules remain the same. This journey adds the complexity slowly: first the readers, then their different duties, then the scoring mechanism, then the controls that make an answer more resilient.
01 RECEIVE02 READ03 SCORE04 MODERATE05 EXPLAIN
01 · THE DECISION CHAIN
The proposal does not meet one reader.
Select a person in the chain. The answer stays still; the duty, question and evidence need change.
Not every person scores every criterion. The chain shows handling, advice, scoring, assurance and moderation roles that may exist; the published procurement defines the actual authorised panel.
02 · THE SAME FICTIONAL BID · FOUR ROOMS
Change the condition.
Watch provisional readings separate.
These are disclosed teaching states—not predictions and not data about a real procurement. Their purpose is to make one simple truth visible: evidence location, risk context and professional duty can change the effort required to support a score.
INDIVIDUAL REASONSMODERATIONONE JUSTIFIABLE RECORD
03 · THE WINNER-FLIP MACHINE
Now move the mechanism—not the bids.
Three fictional bids remain fixed. Change only price weighting, minimum quality and panel method. The podium explains exactly why it moves.
What sits beneath this simple machine?
The retained dataset contains 137 curated mechanism families combined with 25 parameter variants: 3,425 reproducible teaching states. The simplified machine above exposes the three variables most visitors need first; the deeper instrument tests winner stability across the disclosed mechanism family.
04 · CONTROL WHAT CAN BE CONTROLLED
Prediction is fragile.
Bid resilience can be designed.
Do not enter the wrong room
Test mandatory conditions, strategic fit, margin, evidence and delivery capacity before writing absorbs the team.
Make every requirement locatable
Separate compliance, scoring, contract and submission instructions; assign an owner and release condition.
Place evidence beside the claim
Name scope, recency, relevance and consequence so confidence does not depend on generous interpretation.
Read through every authorised duty
Ask what service, technical, commercial, security, user and moderation readers must each be able to defend.
Test the score before submission
Use independent review against the published criterion, not preference, reputation or wordsmithing alone.
Make the result alter the system
Convert assessment summaries, clarifications and delivery evidence into changed thresholds, evidence and qualification rules.
NEXT · THE ARGUMENT
Move from scoring mechanics to the words themselves.
Listen to the buyer, write the response, test its visible features and revise the friction out of it.