Your PR Machine Has Five Blockers: Score Yours in 45 Minutes

The piece was good. The data was yours. Nobody picked it up. Meanwhile a competitor with a thinner product keeps turning up in the newsletters your buyers read. That asymmetry is almost never about luck or budget, and it is rarely about the writing either. It is about a repeatable system that earns outside coverage, which they have and you do not, and five specific faults that keep it from getting built.

This audit suits B2B marketing leads and founders who need outside parties talking about them, for buyers and increasingly for the assistants buyers ask first. Budget forty-five minutes. If an agency already reports your coverage and branded search every month, you can skip it.

Scoring, and why the score is worth your morning

Give each fault 0 if it is genuinely fine, 1 if it is partly true, 2 if it is a real obstacle. Ten points available.

The reason to spend the time at all: Ahrefs ranked signals against visibility in Google's AI Overviews across 75,000 brands and found branded web mentions on top at 0.664, with branded anchors at 0.527, branded search volume at 0.392 and backlinks trailing at 0.218. Louise Linehan, the author, is careful to note that every factor shows a moderate to very weak relationship and that none of it demonstrates cause (Ahrefs). So this is a reason to work on being discussed, not a promise about what being discussed will buy you. We take the same careful line in our claim check on that study.

What journalists said they want in 2026

A fresh data point sharpens three of the five faults below. Muck Rack's State of Journalism 2026 surveyed working journalists, with its own report page describing the base as responses from nearly 1,100 of them (Muck Rack). Note that press coverage of the same report cites a base closer to 900, so the exact sample is unclear and the findings below are reported second-hand.

As summarised by PR trade coverage, 88% of journalists delete a pitch that misses their beat, while 86% say pitches do inspire stories, which places the failure on relevance rather than on outreach itself. More useful still: 78% define a genuinely relevant pitch as one that affects their audience or community, not merely one that lands inside their coverage area (The PR Net, reporting on Muck Rack's State of Journalism 2026).

The operational preferences in the same summary are specific enough to act on this week. Journalists prefer a one-to-one email, under 200 words, arriving before noon, with at most one timely follow-up (The PR Net). That is a format brief, and most B2B pitches fail it on length alone.

Blocker one: nothing in the pipeline a journalist could use

Symptom: your pitches announce a release or a post.

Test, five minutes: write down the last five things you could have pitched. Tag each as original data, a named expert view, a customer account or a usable tool. Nothing in those four categories means a 2.

Fix: commit to one piece of first-party data per quarter. Siege Media's B2B PR guide argues for original research, case studies and white papers precisely because other people want to cite them. A survey of a hundred of your own customers is a legitimate start, provided the method is described honestly.

What it will not solve: a dull finding in a beautiful report. Before designing anything, ask whether a journalist would quote the result.

Blocker two: a list built for volume

Symptom: one send, two hundred names, a reply rate you would rather not say out loud.

Test, five minutes: take your most recent list and count how many recipients published something on your topic in the past ninety days. Fewer than half is a 2.

Fix: cut to ten names and read three recent pieces by each. The relevance figures above make the arithmetic unavoidable: if the large majority of journalists delete off-beat pitches, a two-hundred-name blast is mostly a deletion exercise with a reputational cost attached (The PR Net).

What it will not solve: a writer who has never covered your area. Courtesy does not create a beat.

Blocker three: the opening line is about you

Symptom: your company's name appears before the reader's problem does.

Test, five minutes: read your first two sentences aloud. Does the journalist's audience show up before your brand? If not, score 2.

Fix: lead with the finding and who it affects. This is where the 78% definition of relevance earns its keep: the pitch has to name the consequence for that publication's readers, not just confirm that the subject sits in the right section (The PR Net). Siege's guide also cites the BuzzStream State of Digital PR report, in which 90.2% of respondents name beat alignment as a leading personalisation tactic and 87.9% name using the journalist's name (Siege Media). Those are practitioners describing their own habits, which is not the same as evidence of what works.

What it will not solve: a subject nobody is writing about. Change the angle, not the salutation.

Blocker four: the report counts links and nothing else

Symptom: backlinks are tracked monthly. Coverage that names you without linking is invisible.

Test, five minutes: search your brand across news and blog results for the last ninety days and count mentions with links and without. If only the linked ones reach your report, score 2.

Fix: add unlinked mentions, branded search volume and referral visits to the same sheet. Keep the link reporting, because Ahrefs' own write-up quotes earlier work by Ryan Law concluding that unlinked mentions do very little for classic search rankings (Ahrefs). The argument is for a wider report, not a replacement one, since classic search and AI answers may well reward different things. One channel most B2B teams never log at all is covered in our piece on YouTube mentions as a visibility signal.

What it will not solve: attribution. You still will not know which mention moved a deal. Ask the buyer.

Blocker five: no named human can comment today

Symptom: a reporter asks for an expert and the internal thread takes two days.

Test, five minutes: name the two people who could give ten minutes of usable comment on your subject this afternoon. If you cannot, score 2.

Fix: nominate two, write each a three-line biography and a list of five subjects they will speak to, and brief them on your own data first. Siege's guide makes expert commentary and exclusive insight the basis of a journalist's trust (Siege Media). Turning one expert's view into something citable is a craft of its own, set out in our five steps from a single expert interview, and the same insistence on real people runs through our community strategy guide.

What it will not solve: an expert with nothing new to say.

Reading your total

• 0 to 3: the foundations hold. Raise frequency and keep the data cadence.
• 4 to 7: repair your two highest-scoring faults before you brief an agency.
• 8 to 10: one data asset and a ten-name list. Nothing else, for one quarter.

Those bands are our judgement about where effort pays, not a finding from any study.

Build the asset before you buy the list

The common sequence is backwards. Teams buy media database access, assemble a target list, then look for something to say. We would invert it: produce the asset first, even a small one, and let the asset decide the list. A hundred-customer survey with an honest method gives you a reason to contact a specific ten people, and the finding tells you which ten.

The evidence supports the ordering. If the large majority of journalists bin pitches that miss their beat, and relevance in their own definition means affecting their readers, then the binding constraint is the thing you are offering, not the number of addresses you can reach (The PR Net). A database makes a weak offer travel further. It does not make it land. And the brand-mention correlation that makes this work attractive in the first place is a correlation, with the author of the study saying so explicitly (Ahrefs), which argues for cheap repeated bets rather than one expensive launch.

The order reverses for a team with original research and no coverage. If you already ship original research every quarter and still get no coverage, the asset is not the constraint and the list, the timing or the pitch itself is. At that point paid tooling and an outside specialist are a reasonable purchase, because you have something worth distributing. Teams that buy the tooling first usually discover they have automated the sending of a pitch that was never going to be answered.

A deliberately dull weekly rhythm

A system beats a push. For two people, this fits around other work. Early in the week, find one question journalists in your space are actively asking and choose it. Draft a short note, under two hundred words, aimed at three named writers, and send it in the morning. Hold the rest of that day and the next open for fast replies, because speed is most of the advantage a small team has. Close the week by logging mentions, branded searches and referral visits in one sheet, with a line on what you will change. Once a quarter, ship the data asset. The routine is boring by design. A name that journalists and assistants recognise is built by repetition, not by one good week.

The order of operations, if you only read this far

Asset first, because pitching without one burns goodwill you cannot buy back. Then the ten-name list, then the first two sentences of the pitch. Mention tracking third, because it is what keeps the whole effort honest when nobody has written about you for six weeks.

FAQs about building a PR machine

What makes something a PR machine rather than a campaign?

Repeatability. A machine has a standing supply of things worth covering, a researched and short target list, a pitch routine that runs whether or not anyone is excited, and a measurement sheet that catches coverage with and without links.

Does outside coverage help visibility in AI answers?

Possibly. Ahrefs found branded web mentions tracked AI Overview visibility more closely than any other signal it measured across 75,000 brands, while stating that the relationship is a correlation and not a demonstrated cause.

How long should a pitch be, and when should it go out?

Trade coverage of Muck Rack's 2026 journalist survey reports a preference for a one-to-one email under 200 words, sent before noon, with one timely follow-up at most. Treat that as a format brief rather than a guarantee of a reply.

Is it reasonable to use AI tools for outreach?

For background research and a first draft, yes. The personalised parts, the reason this story matters to this publication's readers, need a person who has actually read the writer's recent work. Everything that goes out still needs human review.