ORIGINAL EDITORIAL SCORE METHOD / BMI 1.2.0

A great brand has to be made well and remembered correctly.

Craft measures the quality of the work. Memory measures whether its identity can be recalled, rebuilt and used without drifting. The Brand Memory Index combines both because one should never hide the weakness of the other.

This page documents the six-attribute rubric behind the original editorial previews. The Living Brand Index adds evidence-led review cases across eight findings, with each publication stating whether its earlier score still applies. The methods are not silently converted into one another.

How a living reading changes ↗

WHAT GOES INTO THE SCORE

Two lenses. Three attributes in each.

The original editorial previews use these six questions. Their weights come from the recorded scoring contract; they are not proof of validated audience memory.

01 / CRAFT

Is the work genuinely good?

Craft measures the quality, originality and rightness of the brand system itself.

EXE

Execution

40%

How precisely and beautifully the system is made.

TESTED THROUGH: Real touchpoints, a signed human review and a canon anchored model read.
DOES NOT COUNT ON ITS OWN: Scale, prestige and polished guidelines do not count on their own.

DST

Distinctiveness

35%

How far from category default, how ownable.

TESTED THROUGH: Blind attribution, competitor confusion, remnant recognition and verbal distinction.
DOES NOT COUNT ON ITS OWN: Fame and a familiar logo do not automatically make a system distinctive.

FIT

Fit

25%

Does this work belong to this company, this category, this moment.

TESTED THROUGH: Positioning, category context, cultural fit and verbal tone.
DOES NOT COUNT ON ITS OWN: Looking current or familiar is not the same as belonging.

02 / MEMORY

Can the identity survive being used?

Memory measures whether people and machines can retain, rebuild and extend the system without drift.

PRM

Primitives

40%

Rules, not artifacts: can the system be specified.

TESTED THROUGH: Primitive recovery, construction legibility and reconstruction from a compact grammar.
DOES NOT COUNT ON ITS OWN: Supplying files is not enough if the reusable rules cannot be recovered.

ELA

Elasticity

35%

Identity under extension, personalisation, moods.

TESTED THROUGH: Variation, personalisation, channel, mood and evolution stress tests.
DOES NOT COUNT ON ITS OWN: Repeating one safe template everywhere is not elasticity.

DUR

Durability

25%

What the machines actually retain, verbally and visually.

TESTED THROUGH: Unaided model recall, cross model coherence, reproduction and forgetting tests.
DOES NOT COUNT ON ITS OWN: Fame, age and simplicity are hypotheses, not evidence of recall.

FROM WORK TO SCORE

A score is earned, not generated.

The machine helps run the tests. It does not get the final word. Evidence, fixed anchors and editorial review are all part of the published result.

01Observe
Freeze the real brand system, its source material and the surfaces people actually encounter.
02Test
Run blind, recall, reconstruction and stress tests against the same evidence pack.
03Calibrate
Compare the six original attributes against declared anchors. Published preview scores remain editorial judgements while calibration is open.
04Sign off
The TBMC Taste Council reviews the evidence and the published receipt binds the result to the scoring version.

WHY CALIBRATION MATTERS

High scores have to be rare enough to mean something.

Our first machine draft put 93 of 100 provisional brands at 80 or above. We rejected it. The model was rewarding fame, reach and good documentation instead of demonstrated performance in the tests.

93 / 100DRAFT SCORES AT 80 OR ABOVE

84.5REJECTED MACHINE MEAN

Machine ratings stay provisional until the evidence holds against the anchors and a Taste Council editor signs the review.

THE SHARED ANCHOR LADDERSAME STANDARD / EVERY BRAND

95
Canon-grade: would not embarrass the anchor set
85
Excellent: conviction and coherence throughout; flaws are preferences
75
Strong professional work; competent everywhere, chosen in places
65
Adequate: consensus design; nothing wrong, nothing owned
50
Weak: borrowed aesthetics, unresolved system
40
Broken: the elements argue with each other

HOW TO READ THE RESULT

The score is a ranking. The band is the verdict.

The public scale stops at 99. Each band describes how strongly the brand system is made and how reliably it remains itself.

90 TO 99Indelible icon

Known by heart.

80 TO 89Committed signature

Committed to memory.

70 TO 79Retained mark

Held

60 TO 69Fading sketch

Recognised

0 TO 59Forgotten draft

The machines draw a blank.