Pulse Methodology

Two kinds of ranking

Official charts are reproduced without alteration: the original order is preserved, the chart owner is named, the measurement period is shown, and nothing is blended with another chart.

Viranova indices combine several credible signals into a Pulse Score from 0 to 100 under the fixed weightings below. Each signal is converted to a comparable 0–100 scale before weighting. Viranova's own audience data is always a minority signal, so a ranking never mainly reflects what Viranova already chose to cover.

The Pulse 50 score

Each platform observation is scored on chart strength, momentum and staying power, then combined by canonical entity across the signal families actually observed — one versioned methodology, published in full.

The flagship board currently draws from 21 active source adapters. Each platform observation is scored first; observations resolving to the same canonical entity are then combined under the eight-family signal network. A family without a real observation for that entity and scoring day is excluded and disclosed — never estimated as zero — and the observed weights renormalise.

Inside each platform observation

• Chart strength — 62%. Position maps to 0–100 with the top-weighted curve 100 × ((depth + 1 − position) ÷ depth)^0.75.
• Momentum — 26%. Movement is measured against the previous verified collection of the same source. A proven first arrival scores 72; when no earlier collection exists, movement stays neutral at 50.
• Staying power — 12%. Collection days seen are weighted by how much of the seven-day window that source has actually covered. A young source is pulled toward neutral; Netflix uses its published weeks-in-Top-10 field against a ceiling of eight.
• Ties resolve by published rule: stronger consumption, then the deeper chart, then chart position, then title.
• Cross-platform comparison is the index's editorial judgement, made comparable by scoring every entry with the same formula on its own chart. Platform orders are never altered inside their own charts.

The signal network

• Consumption — 32%. Verified streaming, listening or viewing rankings.
• Video — 10%. Verified long-form and short-form video performance.
• Social — 18%. Verified social engagement, sharing and acceleration.
• Discovery — 10%. Verified discovery behaviour such as Shazam or equivalent.
• Search — 8%. Google Trends South Africa and approved search signals.
• Conversation — 10%. Verified discussion velocity across authorised sources.
• Broadcast — 7%. Licensed radio and broadcast performance.
• Viranova — 5%. Volume-gated first-party Viranova audience behaviour.

Search requires a conservative normalised match to a credited entity name or multi-word title. Viranova's first-party family activates only above 500 eligible sitewide events and remains capped at 5%. Social, conversation and broadcast contribute nothing until an authorised source is actually active.

Score, coverage and confidence are three different numbers

• Pulse Score (0–100) measures performance on the evidence that exists — signal families with real observations, renormalised under the published network weights (Consumption 32, Social 18, Video 10, Discovery 10, Conversation 10, Search 8, Broadcast 7, Viranova 5).
• Signal coverage (%) measures how much of the vertical's applicable signal universe those observations span. It is disclosed on every entry and is never blended into the score.
• Confidence measures the freshness and integrity of the observations themselves against each source's own cadence limit.

Entity identity

Every ranked thing resolves to a canonical Viranova Pulse identity with per-platform source mappings. Automatic resolution is deterministic — exact platform identifiers first, then normalised metadata — and ambiguous cases queue for an editor rather than merging. Fuzzy text similarity never merges identities; manual merges, splits and aliases are recorded and auditable. The same canonical graph connects people and entities to Viranova stories, profiles and events, but those editorial relationships become ranking evidence only where the published first-party rule explicitly allows it.

Source overlap and double counting

Sources measuring substantially the same behaviour share an overlap group, and scoring admits at most one member of a group per signal family — an aggregate industry chart plus its component platforms can never count the same listening more than once. When an authorised aggregate is live it anchors the family and component platforms serve movement and context only.

Time modes

Pulse computes three real temporal products: Now (movement between the two most recent verified collections), Week (the primary Pulse 50, 7-day window) and Month (30-day window, which holds until at least 14 verified collection days exist). Each has its own window; none is a relabelling of another.

Social integrity

When authorised social sources activate, signals showing manipulation patterns — impossible engagement bursts, abnormal ratios, purchased-engagement markers — are discounted or excluded and the exclusion is recorded in the scoring trace. Detection thresholds are deliberately not published.

Automatic approval, failing closed

Editions are approved by the engine itself when — and only when — the integrity gate passes: verified snapshots present and fresh, valid ranks and titles, snapshot-backed evidence on every scored entry, scores in range, no retired wording. Any failure holds the edition for DATA INTEGRITY and alerts operations; degraded output is never published. Corrections to published editions remain append-only.

Methodology versioning

Weights and thresholds never change silently: every edition and every score trace records the methodology version it was computed under (this document describes v3.2), and archived editions keep displaying the version that computed them.

Signal roadmap

Source-gated products remain visibly held until their authorised families clear their versioned product floor. The two weighted experimental indices below use their original, stricter rule: they remain held while more than 20% of required weight is unavailable. A roadmap source contributes nothing before activation.

Breakthrough 10 weights

Breakthrough measures acceleration, not total popularity — a song can lead it without being №1 on Soundwave 10.

• Official SA chart-position growth — 35% (Coming soon — coming soon)
• YouTube chart or view momentum — 20%
• Apple Music chart growth — 15%
• Shazam or discovery growth — 10%
• Google search growth in South Africa — 10%
• Verified social-content momentum — 5% (Coming soon — coming soon)
• Viranova audience interest — 5%

Culture Current 10 weights

• South African search momentum — 25%
• Verified social momentum — 20% (Coming soon — coming soon)
• Coverage across approved publications — 20% (Coming soon — coming soon)
• Event, streaming or consumption evidence — 15%
• Viranova audience signal — 10%
• Sustained growth across the period — 10%

Accuracy rules

• After the first collection day, an overall-board entry needs at least 2 observed dimensions: scored families, multiple platforms, or three verified chart days. Cross-Platform 20 is stricter: at least two platforms and two scored families.
• Live product coverage floors are versioned by product: 30% for the current overall, music, screen and podcast boards; 55–60% for source-gated products. Below its floor, a product holds and says why.
• A failed source is recorded as failed. It never silently becomes a zero.
• Every ranking run carries an internal confidence score: 90–100 verified, 75–89 publishable with editor review, below 75 held.
• Search interest is treated as attention, never as approval. Spikes driven by scandal, death or other negative news are flagged, not celebrated.
• No fabricated figures. No ranking based only on an AI answer. No changing official chart orders. No retroactive changes without a correction record. Sponsorship can never affect a position; evidence may be submitted, placement may not be bought.

South African classification

An entry counts as South African when a credited name matches a verified Viranova artist profile or eligibility record — nationality is never guessed from a name. Entries that cannot be verified are excluded and the exclusion is disclosed. Clean and explicit versions, spelling variants and re-issues resolve to one entity; remixes only merge with originals where the source chart already merges them.

Sources · Corrections · Back to Pulse