MRS scores every article on four pillars, rolls them into the narratives the business must win, and benchmarks against the peers contesting them. Volume of coverage enters nowhere.
A composite score out of 100 for how earned media coverage is shaping a company's reputation. Every article scores on four pillars. A company's pillar score is the mean of its article scores, which is why more coverage cannot raise it.
How the coverage treats this company, not the overall mood of the document. A story can be broadly negative and still treat one company well, and most sentiment tools cannot tell the difference.
Whether the company is the subject of the story or one of six names in a round-up. A headline about you and a passing mention are not the same event.
Whether the outlet matters to the stakeholder this reporting serves. A tier-1 business title on your priority narrative outweighs ten tier-3 mentions. Reach never feeds this pillar.
How far it traveled, ranked against every other article in the set, never as a raw number. This is the smallest weight on purpose.
Reputation Intelligence teaches the framework: score each article on source authority, prominence, salience and sentiment. The four pillars are the calibrated version of the same model. Nothing was dropped.
| In the book | Pillar here | What changed |
|---|---|---|
| Source authority | Authority | Same idea. The book's voice bonus, a spokesperson or expert quoted, also counts here, along with whether the company is pictured. |
| Prominence and salience | Distinction | The same two factors, combined into one pillar: is the company in the headline or the lead, and is the story about it or about a competitor. |
| Sentiment | Sentiment | In the book every multiplier is positive, so a negative story still adds to a score, only less. Here a negative story lowers it. |
| Reach, "context, not scored" | Amplification | The one real addition. Counted as a percentile against every other article in the set, never as a raw number, and at the smallest weight. |
| Key message bonus | Not a pillar | Used in client work, where the messages are known. Never scored on the public index. |
Why add Amplification. Leaving reach out ignores whether anyone saw the story; letting it lead is how a passing mention in a huge round-up outranks a headline that matters. At 15 percent, reach can nudge a score but never decide one. The book's version is in the MRS Scoring Guide.
Pillars are how an article is scored. Drivers are the narratives a business must win. These five are the standard set, used on the public index so 100 companies can be compared on the same basis.
Governance, conduct, ethics, safety, compliance. The driver that decides a bad day.
Products, research, technology leadership, first-to-market.
Results, growth, guidance, operational execution.
Environmental and social impact, community, workforce.
Executive visibility, strategy, credibility of the people at the top.
Four fictional companies, each with the same five driver scores in every view. Only the weights change. In client work the weights come out of the PBI session, because they say what the business is trying to win. Pick a priority set, or move the sliders and set your own.
Illustrative companies and scores. The bars show each driver's weighted contribution to the score. On licensed coverage read by a person, drivers separate like this. On open data they move together, which is why the public index barely moves when you reweight it.
The step most composite scores skip is the third one. Averaging everything into a single figure hides the driver that decides your bad day.
Each article-company pair is tagged for salience, sentiment, prominence, imagery and voice. Reach becomes a percentile against every other article in the set.
Those tags produce four sub-scores. Every pillar is put on the same 0 to 100 scale so they can be compared and blended.
Pillars roll into the narratives the business must win, identified in the PBI session. This is the step that keeps a bad driver from being hidden by a good average.
Drivers blend into one number, weighted to that company's actual priorities, and ranked by distance from the sector median, not by raw score.
A company's pillar score is the mean of its article scores. Twenty good articles and two hundred good articles produce the same number. That is the main structural break from share-of-voice reporting.
A headline that is really about you outranks a huge-reach story where you are a passing mention. The report shows the high-reach stories the method demoted and why, so you can see the correction being made.
One article republished forty times is one article, carried as a reach multiplier. Unfolded syndication can move a net sentiment figure by ten points, which is a number someone would act on wrongly.
The company must be the subject of the claim, not merely present somewhere on the page. This is what keeps an unrelated tax story or a property sale out of your reputation set.
Rank is distance from your own sector's median, not raw score, so a company is not penalised for operating in a sector that reads negative.
Estimated figures carry a tilde. Human-verified and auto-scored stay visually distinct, and the count of each appears on the report, every period.
Runs on open data. Sentiment and Amplification are computed; Authority and Distinction are flagged approximations. 100 companies, six sectors, monthly, no gate.
It is labeled Public Signal everywhere it appears, never bare MRS, because calling an approximation by the full name would be the exact thing this method exists to avoid.
Runs on your licensed coverage. All four pillars computed, human verification across the priority set defined in PBI, and driver weights set to your board's actual priorities.
The two read different coverage and will produce different numbers for the same company. Every client report says so up front, so you never have to find it out.