About the Brand Power Index

Brands are powered by people. The BPI is a global ranking system built from the voices of real humans — not marketing budgets, not stock tickers, not press releases. You rank the brands you love. You call out the ones that let you down. The world's most powerful brands earn their rank one vote at a time.

Two ranks, one story

Every brand in the index has two ranks. The BPI Rank is the ranking that matters. The Data Rank is kept alongside it as a contrast.

Primary

BPI Rank

100% powered by people. A confidence-scaled, weighted mean of every 1–10 vote, normalized 0–100 across eligible brands.

Secondary

Data Rank

The old-school view. Nine weighted categories covering financials, social, loyalty, PR, ads, growth, market share, community, and heritage.

When the two disagree, that's the interesting story. A brand with billions in revenue and a BPI rank of #300 tells you something real. A scrappy upstart at #5 in BPI with a Data Rank of #800 tells you something else. We show you both.

How the BPI Rank is built

The BPI Rank turns millions of individual 1–10 votes into a single 0–100 score. The full formula, plain and explicit:

// For each brand:
weightedMean = weighted mean of all 1–10 votes
weightedCount = sum of vote weights
signal = weightedMean × log(1 + weightedCount)
bpiScore = minMaxNormalize(signal, eligibleBrands) × 100
Where eligibleBrands = brands with weightedCount ≥ 50.

1. Each vote gets a weight

Before we average anything, every individual vote is multiplied by three factors: identity verification, industry familiarity, and recency. They combine multiplicatively.

Verification tier
New account (CAPTCHA only)0.5×
Phone-verified1.0×
Verified Voice (paid)2.0×
Industry familiarity
None0.3×
Casual0.7×
Engaged1.0×
Expert1.5×

On top of that, every vote is multiplied by a recency factor that halves every 180 days. A vote cast today is worth twice what that same vote will be worth six months from now. The ranking follows how people feel now.

2. Confidence matters as much as score

A 9.5/10 from 60 voters is not the same as an 8.5/10 from 5,000 voters. The BPI multiplies each brand's weighted mean by the natural log of its vote count, so high confidence can outweigh a slightly higher average. In practice:

BrandMeanVotesSignal
Brand A9.56039.1
Brand B8.55,00072.4

Brand B wins despite a lower mean, because 5,000 real voters is substantially more evidence than 60. The logkeeps it from running away — 50,000 votes aren't ten times as persuasive as 5,000.

3. Eligibility: 50 weighted votes

A brand needs at least 50 units of weighted-vote mass before it enters the BPI ranking. Below that, the brand shows a “Gathering data · N/50votes” badge and sits at the bottom of the BPI view. This stops a single enthusiastic user from handing a brand the #1 spot.

4. Normalize to 0–100

Once every eligible brand has a signal value, we min-max normalize them into a clean 0–100 BPI score. The brand with the highest signal gets 100; the lowest eligible brand gets 0; everyone else scales linearly in between. Ranking is just a descending sort of those scores.

Secondary methodology

How the Data Rank is built

The Data Rank is a composite score on a 0–1000scale built from ten weighted dimensions of brand power. Reach captures how many people know the brand; Connection measures how intensely the audience engages; Affinity, how much they care; Relevance, how central the brand is to its category. Financial Strength, Momentum, Presence, Activation, Community, and Heritage round out the picture. Together, they're a data-grounded estimate of the same thing the BPI Rank captures directly — the brand's power in the world.

Why Reach and Connection are separate

A niche brand with a small but rabidly active audience can post double-digit engagement rates; a global brand reaches billions at 2%. Collapsing both into one bucket would let the niche brand outrank the global one on sheer intensity. Splitting them — and capping Connection at a 5% ceiling so outlier engagement doesn't run unbounded — keeps the ranking honest across very different brand shapes.

Category weights

CategoryWeight
Reach18%
Connection8%
Financial Strength15%
Relevance15%
Affinity15%
Momentum10%
Presence7%
Activation6%
Community3%
Heritage3%

Weights sum to 100%.

How raw metrics are normalized

Different metric types call for different normalization. Four modes:

Log + min-maxFollowers, revenue, market cap, search volume, media mentions, ad spend, revenue-per-ad-dollar
Metrics that span orders of magnitude — a 100M-follower brand shouldn’t drown out a 1M-follower one purely by scale.
Linear min-maxRepeat purchase rate, growth YoY, crisis count (inverse), ad recall, ad conversion
Bounded metrics where every raw point should count equally across the dataset.
Capped linearEngagement rate (Connection)
Engagement has a natural saturation point. A 6% rate isn’t three times stronger than 2% — it’s just saturated. We cap credit at a 5% ceiling so outlier audiences don’t overwhelm the ranking.
Fixed-range mappingSocial sentiment, media sentiment, NPS
Sentiment lives on a known scale. We anchor to the scale rather than let extremes shift with the dataset.
Within-industry groupingMarket share % (Relevance)
Market share is meaningful only against peers — 3% in Luxury Fashion beats 3% in Consumer Packaged Goods.

Final composite

Each category score sits on a 0–100 scale. The category scores are combined using the weights above, then multiplied by 10to produce the 0–1000 Data Score. The Data Rank is a descending sort of those scores across the full index.

Data sources

Every data point in the Brand Power Index comes from publicly available, legally accessible sources. Transparency is a core principle — explore the methodology behind every score.

Reach
Platform follower counts and branded search volume — how many people know the brand exists.
Connection
Engagement rate across public profiles, with a diminishing-returns ceiling so outlier fandoms don’t dominate the global view.
Financial Strength
SEC/EDGAR filings, annual reports, and revenue-per-ad-dollar as an efficiency signal.
Relevance
Within-industry market share — 3% of Luxury Fashion means something different than 3% of Packaged Food.
Affinity
Published NPS benchmarks, repeat-purchase studies, and social sentiment — how much the audience actually cares.
Momentum
Year-over-year change in revenue, social footprint, and market share.
Presence
Media mention volume, press sentiment, and crisis events (inverse).
Activation
Public ad spend rankings, recall studies, and conversion benchmarks.
Community
Aggregate favorites from BPI members — brands people save and return to.
Heritage
Year founded, log-scaled as a durability signal.

Frequently asked questions

What is the BPI Rank, and how is it different from the Data Rank?

The BPI Rank is the primary, people-powered ranking — the weighted mean of every 1–10 vote, scaled by confidence (log of vote count), normalized 0–100 across eligible brands. The Data Rank is a secondary view built from public numbers (financials, social footprint, PR, ads, community, heritage). When the two disagree, that’s the story — the people see something the numbers don’t, or vice versa.

How much does one vote actually count?

Every voice matters, but not every voice carries the same weight. A brand-new account counts at 0.5×. A phone-verified human counts at 1×. A paid Verified Voice member counts at 2×. Your self-reported familiarity with an industry boosts your vote in that space. Recency matters too: votes decay on a 180-day half-life so the ranking reflects how people feel now, not five years ago.

Can I pay to vote more loudly?

No one can buy a higher rank for a brand. What $3/month Verified Voice gets you is a trust badge and a 2× multiplier on your own votes — because identity-verified voices are harder to fake than burner accounts. Everyone can vote for free, forever.

What stops brigading and pile-ons?

Sudden spikes on any single brand (e.g., 50+ votes in ten minutes) pause that brand’s BPI score for 48 hours while the votes are still recorded. Extreme votes (1s or 10s) require users to have voted on other brands in that industry first. Velocity caps per user. Coordinated rings get flagged to our moderation queue. We don’t hide opinions — we just make it harder for a dozen throwaway accounts to outweigh thousands of real ones.

Why does a brand sometimes show “Gathering data”?

A brand needs at least 50 units of weighted-vote mass before it enters the BPI ranking. Until then it shows a progress badge (“Gathering data · N/50 votes”) and sits at the bottom of the BPI view. This stops a single enthusiastic user from handing a brand the #1 spot.

Why don’t you use brand logos?

The BPI represents brands through typography, color, and your voice — not marketing assets. The focus is on what people actually feel.

Can my brand be added to the index?

The index covers over a thousand global brands. Don’t see one you want to rank? Hit “Add Brand” in the nav — we auto-detect the category and open it up for ranking immediately.

The People's Brand Report

Monthly

Which brands the world just fell in love with. Which ones got called out. Industry upsets, cult favorites, and the gap between what the people say and what the data says — in your inbox once a month.