Ask five tools how your brand is doing and you get five dashboards: a star average here, a mention count there, a sentiment pie chart somewhere else. Each is accurate. None answers the question a founder or an agency account director actually asks, which is simpler: is our reputation getting better or worse, and should I do something about it this week?
A brand health score exists to answer that question in one number. This article explains what goes into a serious score — the four ingredient families, and why each is weighted the way it is — how to read the number without being fooled by it, and the specific levers that move it. It is also the metric at the center of PigPR Console, so where useful we will show how the console surfaces it.
What a brand health score actually is
A brand health score is a composite index, typically 0 to 100, that summarizes the state of a brand's public reputation across the channels where that reputation is formed: review platforms, news coverage, social threads, forums. Like a credit score, it is not a raw measurement of any single thing. It is a weighted synthesis of several measurements, computed on a rolling window, designed so that day-to-day noise moves it a little and real shifts in public perception move it a lot.
Two design decisions separate a useful score from a vanity number. First, it must be directionally honest: if it goes up while your reviews are getting worse, the formula is wrong. Second, it must be decomposable: when the number moves, you can open it up and see exactly which ingredient moved it. A score you cannot decompose is a horoscope.
How the score is computed
Formulas vary by vendor, but every credible brand health score draws from the same four ingredient families. Here is what each contributes and why it earns its weight.
Sentiment mix
The foundation. Every ingested mention — a Google review, a news article, a Reddit thread — is classified positive, neutral, mixed, or negative, and the score reflects the ratio, not just the average. Ratios matter because averages hide barbells: a business with fifty 5-star and fifty 1-star reviews averages three stars, which describes nobody's actual experience. Sentiment mix also tracks the mixed bucket separately, because mixed mentions are the leading indicator: they are where tomorrow's positives and negatives are decided.
Volume and velocity
A brand nobody discusses has no reputation to score. Volume — mentions per period — provides the confidence floor: a sentiment ratio computed from six mentions is a rumor; computed from six hundred, it is a signal. Velocity is the first derivative: how quickly volume is changing. A sudden burst of negative mentions should dent the score faster than the same mentions spread over a month, because bursts are how crises announce themselves. This is the same math that powers spike detection and the 6-hour rule.
Source weighting
Not all mentions are equal, and a serious score refuses to pretend they are. A detailed Google review from a verified local customer carries more weight for a local business than a one-line forum gripe, because it sits where buying decisions happen. A story in a high-circulation news outlet outweighs a scraped blog mention. Weighting schemes usually combine reach (how many people will see it), intent proximity (how close the audience is to a purchase), and permanence (reviews persist; tweets decay).
Response behavior
The ingredient most scores omit — and the one that makes the metric actionable. Public perception is shaped not only by what customers say but by how the brand answers. Response rate (what share of reviews get a reply), response time (how fast), and response quality signals feed the score as a multiplier on the negative side: unanswered negatives cost more than answered ones. This is deliberate. It makes the score reflect something the business controls completely, so the number always comes with a lever attached.
Why one number beats five dashboards
The case for a composite is not laziness; it is communication. An agency reporting to a client, or a GM reporting to an owner, needs a metric that fits in the subject line of a weekly email: 76, up 4 points, driven by response coverage. The dashboards still exist — they are the second click, the evidence behind the delta. But the score is the shared language that lets a whole team notice a trend at the same time and argue about causes instead of arguing about which chart to look at.
It also disciplines the team against the two classic reputation mistakes: celebrating a good month that was just quiet, and panicking over a loud week that was just one angry thread. Volume-weighting handles the first; velocity smoothing handles the second.
How to move your score
Because the score is decomposable, improving it is a menu, not a mystery. In rough order of how fast each lever acts:
- Close the response gap. Replying to unanswered negative and mixed reviews moves the response-behavior component within days. It is the only lever fully under your control tonight. Use the response framework and a same-day SLA.
- Fix the mixed bucket. Mixed reviews name specific, fixable frictions — slow checkout, confusing parking, one brusque staff interaction. Each fix converts future mixed mentions into positives and is permanent, unlike any marketing spend.
- Earn review velocity honestly. A steady cadence of fresh reviews outweighs a one-time blast. Ask at the moment of peak satisfaction, make the ask personal, and never gate or incentivize — both violate platform policy and distort the signal you are trying to improve.
- Contain negatives before they cluster. Alerts on rating dips and mention spikes let you respond while a problem is still one review, not six. Clustered unanswered negatives are the single most expensive pattern in the formula.
- Feed the positive sources. News and community coverage have the highest source weights. Sponsor the thing, take the local press call, publish the useful guide — then make sure your monitoring actually ingests those channels, or the score never sees the credit.
What a good score looks like
Benchmarks depend on category and volume, so treat these bands as orientation rather than grades: 85 and above is a dominant reputation — protect it with alerting, because falls from height are fast. 70 to 85 is healthy with visible headroom, usually won or lost on response coverage and mixed-review fixes. 50 to 70 is contested: the sentiment mix is fine but the operation is leaving points on the table. Below 50 is a repair project measured in quarters, not weeks — start with responses and the top two recurring complaints, and read the local business playbook before spending on anything else.
The bottom line
A brand health score is not the reputation. It is the instrument that lets you manage the reputation — honest about direction, specific about causes, and always attached to a lever you can pull.
Pick the score that decomposes, watch it weekly, and when it moves, ask which ingredient moved first. That habit — one number, one question, one lever — is what reputation management looks like when it grows up.