Chapter 26
PR and reviews
PR's job is no longer coverage. It is getting into the sources the machine quotes, with something only you have.
What changed
PR is usually judged by its clippings: how many pieces ran, how many people they reached, and what the same space would have cost as advertising. That last figure is the advertising value equivalent, or AVE. The industry's own measurement body gave up on it long ago. AMEC's 2020 Barcelona Principles are blunt: "AVEs are not the value of communication" (AMEC). In my experience, it still turns up in reports.
Reviews get run as a score to protect. Someone watches the star rating, chases a few happy customers when it dips, and leaves the bad reviews unanswered.
Neither habit asks what the buyer, or now the machine, read at the moment of checking.
What it is for
PR and reviews have one job once the goal is being selected. They put other people's words about you into the places buyers and machines check.
That makes this the main build for Believed. Judgements about you are only accepted from others, and this build is how you earn them.
It serves three more stages. It builds Known in the machine's memory: news mentions went with AI prominence in Chapter 13's exploratory study. It serves Found, because the sources cited for a buying question decide who gets named. And it supplies the reason at Preferred, in someone else's voice.
It depends on Kept. Reviews come from customers whose promises were kept. PR can't fix a product that lets people down; it'll only spread the news faster. And it doesn't do the work of Answered or Bought. Your facts, prices and terms belong on your own site.
The anatomy of a good one
It runs as two tracks on their own clocks. Reviews turn every week; PR turns once a quarter. For most small firms, reviews are the Monday task, so start there.
Reviews, asked at the right moment [Believed]. Ask when the job is done, at the first win. Never gate: the US Federal Trade Commission says, "Don't ask for reviews only from customers you think will leave positive ones" (FTC).
A public reply to every review [Believed · Kept]. The next buyer reads it. Davide Proserpio and Georgios Zervas found that when hotels began responding, their ratings rose by 0.12 stars on average, and they received 12% more reviews (Proserpio and Zervas). Hotels got fewer but longer negative reviews, once unhappy guests knew someone would read them.
The questions buyers ask, and the sources cited [Found]. Ask your buying questions in the assistants your buyers use, many times each, and read the citations. Those are your target list. A trade newsletter may matter more than a national paper.
Something only you have [Preferred]. Proprietary data, a finding or a clear point of view. Cision surveyed 1,899 journalists, who named original research, expert access and embargoed information as what they value (Cision). The finding should carry your reason, so it travels whenever the number is quoted.
The right outlets [Believed]. Take the finding to the sources your map turned up, starting with those cited most.
Ask again [Found]. Rerun the same questions each quarter and check whether you now appear.
The mistake most teams make
They pitch announcements to big lists: a new hire, a new office, an award. Nobody else needs to quote any of it. The same Cision survey found 72% of journalists say fewer than a quarter of pitches they receive are relevant.
The second mistake is on the reviews side: leaving bad reviews unanswered, or filtering them. That's now law, not just manners. The FTC's rule on fake reviews took effect on 21 October 2024 (FTC), and in the UK hiding negative reviews is banned too (CMA). Answer the bad ones first, because the next buyer reads them first. Own what was wrong and say what changed.
And don't write your own Wikipedia entry. Wikipedia is among the sources assistants cite most, and paid advocates are "very strongly discouraged" from editing their own articles (Wikipedia). Earn the independent coverage its editors can cite, and let them write it.
How you know it's working
Retire coverage counts, AVE and the star rating as goals. Steer by the share of your buying questions where you appear in at least one cited source; the number of top-cited sources you're in, quarter on quarter; whether your finding is quoted with your reason attached; reviews in the last 90 days; and the share of reviews answered, with the time taken on the negative ones.
An example
Fieldline, the invented scheduling-software firm, found a G2 review saying the travel savings "never showed up after the pilot". Nobody had replied. The customer team answered that week: the customer hadn't switched on route grouping, here were the drive-time figures since the fix, and here was a named contact. A month later, the finance question returned the review and the reply together.
The PR track went less well. The trade magazine ran Fieldline's drive-time finding as a feature, but two quarters on the assistants still cited an industry survey instead. The weekly track is faster. The quarterly one may yet fail.
The full playbook. The PR and Reviews Playbook, with the source map and PR target list, the data-story pitch and the review response guide, is free at rocmode.com/playbooks.
Whose evidence this is. Cision sells PR and media software. AMEC is the measurement industry's own body. The FTC and CMA are regulators. Peer-reviewed: Proserpio and Zervas, whose study used hotels, so other categories are untested.