Methodology: how every score is calculated

The formulas behind FindRivals: weighted scores, price index, value-adjusted price band, market share and HHI, sentiment, keyword opportunity, threat score, Five Forces and ICE.

FindRivals calculates everything in your browser with deterministic formulas. The same inputs always give the same outputs. This page documents each one so you can check, challenge or reproduce any number.

Weighted score (0-100)

Each business is scored 0-10 on each criterion. Weights are normalised to sum to 100%.

Weighted score = 10 × Σ(weightᵢ × scoreᵢ) ÷ Σ(weights of scored criteria)

Criteria without a score are left out and the remaining weights re-normalised, so a missing score does not count as zero. Businesses are ranked by weighted score.

Price comparison

  • Recurring prices convert to a monthly equivalent: yearly ÷ 12, quarterly ÷ 3, weekly × 52 ÷ 12, daily × 365 ÷ 12.
  • Per-unit and one-time prices are kept separate. The comparison uses whichever basis most competitors publish, unless you choose one.
  • Currencies convert at live rates from open.er-api.com (ExchangeRate-API) when available, otherwise at a dated built-in table.
  • Entry price is a business's lowest price on the chosen basis.
  • Price index = entry price ÷ median competitor entry price × 100. Your own business is excluded from the median. Below 80 = budget, 80-120 = mid-market, above 120 = premium.
  • Value-adjusted band: your weighted score's percentile among competitors places you within the range between the 25th and 75th percentile competitor prices (±15 percentile points). It is guidance, not a demand forecast.

Feature coverage and gaps

Features are marked yes (1), partly (0.5) or no (0) and weighted by importance (critical 3, important 2, nice to have 1).

Coverage = Σ(importance × mark) ÷ Σ(importance) × 100

Adoption is the average mark among competitors. Classification: table stakes ≥ 70% adoption; table-stakes gap when you lack a table-stakes feature; gap ≥ 40% adoption and you lack it; differentiator when you have it and ≤ 30% of competitors do; white space when ≤ 20% of competitors have an important or critical feature you also lack.

Market share and concentration

Share = business value ÷ total of tracked businesses, on one chosen basis (revenue in one currency, reviews, visits, followers, employees, listings or outlets).

HHI = Σ(shareᵢ in %)². Levels follow the 2023 US Merger Guidelines: > 1,800 highly concentrated, 1,000-1,800 moderately concentrated, < 1,000 unconcentrated. CR4 = sum of the four largest shares.

Digital share of voice

For each metric (visits, domain authority, backlinks, followers, reviews), share = business value ÷ total. The composite is the average share across metrics with data.

Review sentiment

Reviews are split by blank lines or line breaks. Each review is scored with a lexicon of about 250 positive and negative words (−4 to +4), with negation within three words reversing polarity (× −0.8) and intensifiers such as "very" amplifying it. The review score is divided by √(words + 4). Reviews above +0.25 are positive, below −0.25 negative. The index maps the average to 0-100 (50 = neutral). Themes use keyword groups for ten topics; a theme's tone is the average score of reviews mentioning it.

Keyword gap

Keywords are normalised to lower case. Status: unique (only you), shared, missing, or missing (strong) when at least half of competitors (minimum two) target it. Opportunity = 25 × log₁₀(volume + 1) × (0.5 + share of competitors targeting it), with 50 used when volume is unknown.

Threat score (0-100)

Threat = 40% strength + 25% overlap + 20% momentum + 15% reach

  • Strength = weighted score (unscored = 50).
  • Overlap = 60% competitor type (direct 1.0, indirect 0.65, substitute 0.45, aspirational 0.3) + 40% cosine similarity of feature sets.
  • Momentum: declining 0.15, stable 0.45, growing 0.75, surging 1.0.
  • Reach = average of reviews, visits and followers relative to the largest business.

Levels: ≥ 75 critical, 60-74 high, 40-59 moderate, < 40 low.

Five Forces

Each force is the average of five 1-5 ratings. ≥ 3.67 high, 2.34-3.66 medium, below that low. Attractiveness = (5 − average of all forces) ÷ 4 × 100.

Positioning white space

Positions are normalised to a unit square. FindRivals checks a 9 × 9 grid and marks the point farthest from every business when that distance exceeds 0.18.

Auto-SWOT and actions

SWOT entries come from: criteria at least one point above or below the competitor average; competitors averaging under 6 on a criterion; feature classifications; price position relative to rank; cheaper and higher-scoring rivals; high or critical threats; surging momentum; market concentration; strong Five Forces; review themes with two or more mentions; keyword gaps. Actions are scored with ICE (Impact × Confidence × Ease, each 1-10); ease ≥ 8 goes in the 30-day horizon, 6-7 in 60 days, the rest in 90 days.

What FindRivals does not do

  • It does not scrape websites or collect private data.
  • It does not estimate traffic, rankings or revenue itself; those come from you or the tools you use.
  • AI features (when available) are labelled, optional and never used for calculations.