Craft
How is it made?
Methodology / BRI 1.0
Three questions, drawn to scale. How a brand is made, how it behaves and what it carries forward. Live from day one, sourced, signed and open to new evidence.
How is it made?
How does it behave?
What does it carry forward?
01
Each bar is the primitive’s share of its lens.
Is there a true, clear idea the system is built from?
How precisely are the mark, type, colour and motion made?
How far is it from the category default?
Does the language have a character worth keeping?
Do the parts read as one system?
Does it extend across products, campaigns and cultures without drift?
Does it have designed moods for seasons, places and moments?
Can anyone, agents included, make a surface nobody has drawn yet?
Does the core hold in every hand while the expression flexes? Drift and sameness both lose points.
What can people and machines rebuild without the guidelines?
Are the remembered parts uniquely theirs?
What equity from its past does it keep alive?
Does recall hold across models and over time?
02
Two illustrative brands. A simple average barely separates them. The BRI, a weighted geometric mean, sees that one is beautiful and does not hold.
Balanced
82.9BRI
82.9Simple average
Beautiful, does not hold
77.9BRI
79.5Simple average
The formula
Weights are fixed and published. Any change ships as a new BRI version that restates every score and keeps the old one.
Try the scorer03
Two hundred cells, one BRI. The signed editor carries half. The other half comes from evidence Brand Recall does not control.
Target mix / unavailable signals are carried by the editor, never scored as zero.
04
On reveal day Behaviour cannot be observed, so it starts as a Forecast. Drag the evidence and watch it hand over.
B = w·F + (1 − w)·O, w = 5 / (5 + n)
Scored evidence0
Confidence±6
Forecast 100%
Observed 0% / forecast weight 1.00
Demonstration / confidence is a count heuristic
05
Only through editions, weekly or on a material event, each with a reason. One ordinary edition moves BRI by 4 points at most. Evidence fades with an 18-month half-life, so a brand has to keep behaving. The proposed machine panel re-runs monthly.
06
The five bands, with every retained recall placed on them.
60 to 69
70 to 79
80 to 89
90 to 99
Lost: below 60 / five historical BMI editions / McLaren sample / BRI restatements pending
07 / The Public Read
Counts after 100 ratings / one per account per edition / affiliated raters shown, not counted
07 / The Memory Test Later
Pick the real mark / rebuild the palette / name the brand from one primitive
08
A signed edition keeps the read, its sources and the reason for every change. The old score remains inspectable. A new edition cannot rewrite what we said before.
Designed moods are different from drift. A recognisable core can carry a wide range of expressions; elasticity and consistency are scored separately.
Evidence is labelled by its supplier. A studio response is published beside the independent read, and the editor owns the decision.
One rating per eligible account per edition. Affiliated and held accounts do not count. The Public Read stays separate and cannot directly change a published BRI.
TBMC clients are disclosed on their recall, or not recalled. Scores are independent editorial opinion. Corrections and conflicts remain visible.
How it works / try it
Score the primitives yourself. The BRI is a weighted geometric mean, so balance wins. A beautiful identity that cannot hold together in the world pays a price.
Start from
This is a toy. To score for real, rate a brand on its recall. Your read joins the Public Read.
Rate McLaren