Skip to content

How the Growth Index is computed

Every grade is built from one series: a venue’s month-over-month growth in its clean onchain stablecoin book. Seven steps turn that series into a published grade.

1
Input

The clean stablecoin book, complete months only

For each live venue we take its attributed onchain stablecoin book, aggregated by calendar month. Only onchain evidence is used; third-party registry or exchange labels never enter the score. The month in progress is always excluded.

2
Growth series

Month-over-month growth, trailing 12 months

The book is converted to month-over-month growth rates over the trailing 12 complete months. From here on, size is gone: every later step works only on these growth rates, so a small venue and a large one are measured on the same footing.

3
Momentum · 60%

How fast is it growing?

Momentum starts from the medianof a venue’s trailing-12 growth rates — never the mean, so a single spike month cannot fake growth. That median is then peer-ranked across the field, so a good month in a market-wide boom is not over-rewarded: a venue is scored against how the rest of the field grew over the same window, not against an absolute target.

4
Consistency · 40%

How steady is that growth?

Consistency is the MAD (median absolute deviation) of the same trailing-12 growth series. MAD, rather than standard deviation, treats a one-off explosion as an outlier and penalizes chronic choppiness instead. It is detrended, so steady growth reads as consistent rather than volatile.

5
Confidence

Shrink toward neutral until the window is full

Confidence is set by how much of the trailing-12 window a venue can fill. A venue with a full 12 months is trusted completely; a younger venue is shrunk toward neutral (50) until it fills the window. That is not a penalty. A young venue is held closer to the middle until it has shown enough, which is why an explosive newcomer climbs gradually rather than topping the board on its first big month.

An older venue is not shrunk for history we cannot see: the window never uses data older than 12 months. Full vs Partial coverage is a transparency label only and does not move the grade.

6
Score & grade

60% Momentum · 40% Consistency

The two pillars combine into a 0–100 score, the grade of record, and map to the bands below. Growth is weighted above steadiness, so a flat-but-steady venue cannot top the board on stability alone.

Grade
Score
Reads as
AAA90–100Fast, steady growth
AA80–89Strong growth, mostly steady
A70–79Healthy growth; speed or steadiness lagging
BBB60–69Moderate growth, with a watch-item
BB50–59Flat or choppy growth
B40–49Slow and uneven growth
CCC25–39Shrinking and unsteady
D0–24Sustained decline
7
Publication

A full month behind, every month

The score is recomputed as of each complete month, so both pillars update monthly. September’s standing publishes in October; a venue added after a month has closed joins from the next cycle. That lag is the price of never publishing an unfinished number.

Each month-over-month change in the grade (▲ / ▼) is tagged with its primary driver, Momentum or Consistency, and written up in the monthly report. Because Consistency moves slowly, the grade is deliberately sticky and does not whipsaw on one noisy month.

Provisional

A venue with fewer than 4 complete months is graded but flagged Provisional.

Public-layer only

Terminal / validation venues are graded internally but never appear publicly until promoted.

Migration freeze

A venue in a known infrastructure migration is frozen at its last grade, never auto-downgraded. A coverage gap is not a decline.

What is published

Public method, internal recipe

Everything above is published: the input, the steps, the pillars, the weights, the bands and the eligibility rules. The exact internal transforms that turn the growth series into pillar scores and the final score are proprietary and not published.

Methodological principles

Seven rules the Index holds to

Same inputs, same framework
Eligible venues are evaluated using the same Index architecture and timing rules.
Observed behaviour over labels
The Index uses onchain evidence, never unverified registry or CEX labels, as grade inputs.
Growth over size
Every input is a growth rate. Absolute scale never enters the score.
Median over mean
Median-based growth and MAD reduce the influence of isolated spikes and extreme observations.
Confidence is window-fill
A venue that can't yet fill the trailing-12 window is shrunk toward neutral; lifetime coverage does not move the grade.
Complete months only
The published standing always uses a completed month. Partial-month data is never used.
Coverage gaps are not declines
A known infrastructure migration can trigger a freeze rather than an automatic downgrade until the new rail is measured.
Scope

What the Index does not claim

A grade is a statement about observed onchain growth under this methodology. The Index does not:

  • rank companies by total size,
  • measure revenue, profitability, solvency, or valuation,
  • infer management quality or intent,
  • treat third-party labels as ground truth,
  • penalize a venue simply because its onchain history is shorter — thin history reduces confidence, not the grade,
  • treat a known coverage gap as a business decline while an infrastructure migration is being investigated.

AfriFlux does not sell grades, rankings, or Index changes. Commercial relationships do not determine Index outcomes. The Growth Index is not a credit rating, and not a recommendation to invest in, transact with, or otherwise evaluate a company. It is not a claim about a company’s solvency, financial condition, management, intent, or overall business quality.

Measure the growth. Explain the change.