One score out of a hundred, built from 412 sources across search, social and review sites, recomputed every six hours — with alerts that fire when it moves, so nobody has to remember to go looking.
Reputation problems are cheap to fix early and expensive to fix late, and the gap between the two is usually measured in weeks of nobody noticing. Most brands find out from a customer, which is already too late.
So this is built to be ignored. You do not log in daily. The index sits in the background, and the only time it asks for your attention is when something has actually moved.
| Month | Index |
|---|---|
| Jan | 54 |
| Feb | 56 |
| Mar | 55 |
| Apr | 59 |
| May | 58 |
| Jun | 61 |
| Jul | 64 |
| Aug | 62 |
| Sep | 66 |
| Oct | 69 |
| Nov | 70 |
| Dec | 72 |
Polarity, not volume. Bars grow out from the centre line, so a quarter that got louder in both directions does not read as progress.
| Quarter | Str. neg | Negative | Neutral | Positive | Str. pos | Net |
|---|---|---|---|---|---|---|
| Q1 25 | 11 | 19 | 40 | 22 | 8 | +0 |
| Q2 25 | 9 | 18 | 41 | 24 | 8 | +5 |
| Q3 25 | 8 | 16 | 40 | 26 | 10 | +12 |
| Q4 25 | 7 | 15 | 39 | 27 | 12 | +17 |
| Q1 26 | 6 | 13 | 38 | 28 | 15 | +24 |
| Q2 26 | 5 | 12 | 36 | 30 | 17 | +30 |
We poll 412 sources on a six-hour cycle — search results, social, news, review sites, forums and video.
Each mention is scored for polarity and checked against a sample we grade by hand, so the classifier stays honest as language drifts.
Scores are weighted by the traffic each surface actually receives and rolled into one number out of a hundred, plus three sub-indices.
Four thresholds are armed by default. Crossing one sends the alert with the mentions that caused it, not just the number.
| Source | Weight index |
|---|---|
| Search results | 96 |
| News & press | 81 |
| Social posts | 88 |
| Review sites | 67 |
| Forums | 59 |
| Video | 44 |
Severity is paired with a label, never carried by colour alone. Sparklines show the last ten readings against each threshold.
Monitored and unmonitored businesses get roughly the same number of problems. The difference is how long each one sits there unanswered.
| Quarter | Monitored | Unmonitored |
|---|---|---|
| Q1 | 54 | 54 |
| Q2 | 57 | 55 |
| Q3 | 59 | 53 |
| Q4 | 63 | 56 |
| Q5 | 62 | 54 |
| Q6 | 67 | 57 |
| Q7 | 70 | 56 |
| Q8 | 74 | 58 |
Mostly about how the number is built, which is the right thing to be sceptical about.
Context decides it, which is why the meter shows bands rather than a target. Above 60 is healthy for most consumer-service businesses. What matters more than the level is the direction and the size of any single move — a six-point drop in a week is worth more of your attention than sitting at 64 for a year.
Every mention is classified for polarity, weighted by the traffic its surface receives, and rolled into three sub-indices — search, social and review sites — which combine into the composite. Weights are fixed at setup and held, so a change in the number is a change in reality rather than a change in method.
Automatic, with a hand-graded sample every cycle so we can see when it drifts. Sarcasm and industry jargon are where it struggles, which is why alerts ship with the underlying mentions attached — you are meant to read them, not trust a number blindly.
It should create less. The default is silence: four thresholds, and nothing arrives unless one is crossed. Most clients hear from it a handful of times a quarter, which is roughly the number of times something genuinely needed a response.
Yes to both. Clients usually add the two or three forums or directories that matter in their specific market, and tighten the review-rating threshold if reviews are their main channel.
Yes, though the source list is smaller and the index moves in bigger steps, because fewer mentions means each one carries more weight. We adjust the alert thresholds accordingly so you are not woken up by normal variation.
A program review is a 45-minute read of your current numbers. You leave with the three charts above, filled in with your data, whether or not we work together.