Digital PR

What digital PR is

Digital PR is earning coverage on other people’s sites: pitching stories and data to journalists, responding to their queries, then getting mentioned and linked as a result. That part is settled. What has changed is where the coverage ends up.

The box asks about AI. The advice written about digital PR does not.

A guide from one PR software vendor comes closest of the six, and mentions AI only as an industry challenge and as a feature of the product it sells. Another measures the work by press coverage, site traffic and leads, and stops there.

One definition is worth quoting because it is the one Google’s overview leans on. Semrush describes digital PR as building visibility and reputation through media coverage, high-quality backlinks and positive mentions across digital channels. That sentence contains three outputs, and by the end of this page two of them will have a second audience nobody in the list is counting.

How the work is run

The same four tactics recur everywhere. That overlap is a good sign, not a lazy one. This is a settled discipline with a small number of moves, and the useful question is which of them anyone has measured.

Four moves, and they are not equally evidenced.

The four are not equally evidenced. Do not read them as a menu. Media outreach is the one everybody names, including Google’s summary; reactive PR is close behind. One agency publishes a campaign result, more than 300 pickups and 40 percent organic traffic growth, though its page does not say which of these tactics produced it. Newsjacking and influencer work appear in two guides with nothing measured beside them. That does not make them ineffective. It does mean the evidence stops at somebody recommending them. There is one survey figure nearby that speaks to why journalists engage at all, and it needs two caveats before it is useful. Semrush cites Cision’s State of the Media survey for the finding that 61 percent of journalists want to receive original research and reports from PR professionals. The first caveat is provenance: the number reaches this page through Semrush, not from Cision. The second is what it can support. It records what journalists say they want to receive. That is a preference, not a measurement of what gets published, so it explains why a data-led pitch is worth trying and not how often one lands.

What digital PR reporting misses

Digital PR reporting stops at press pickups, backlinks and referral traffic. Setting those metrics side by side makes it easy to see which surface none of them watches.

Press, links, traffic. Nothing else appears.

That is why your invoice is counted in links. Roger Montti at Search Engine Journal recorded Google’s John Mueller saying unlinked mentions are not used for PageRank or for understanding the link graph.

Six digital PR guides, read 4 September 2026Whether your own reporting has the same shape

What each guide tells you to measure. The last row is a contract, not a recommendation and is marked so.

The shape is what they share. Every measure is something that happens on a page: a story published, a link placed, a visitor arriving. That is not a limitation anyone chose; it is what a PR programme has always been able to count. Whether the newer surface was measurable when each list was written is not something these dates settle: two of them were updated within the last year. One entry in that figure deserves a second look. A tools recommendation names Google Alerts for monitoring, and Alerts reports new results that show up in Google Search. So it can tell you when a page mentioning you turns up as a new result in Google Search, and nothing about anything that is not a page. What it does and does not reach is set out in Google Alerts for brand mentions.

Where the coverage ends up now

Earned coverage now feeds something it did not five years ago. The largest sample in this area looks at what AI systems quote when they answer a question.

Most of what these systems cite is earned.

Read carefully, that is a fact about the systems. It says nothing about your programme. It says the material these engines lean on is overwhelmingly the kind digital PR produces, and almost none of it is the kind you can buy. It does not say that running a campaign will get you cited, and nobody has tested that.

A second measurement sits alongside it and carries the same caution. Louise Linehan at Ahrefs looked across 75,000 brands. Brand web mentions correlated around 0.66 with visibility in AI answers and backlinks 0.218, and the authors state plainly that correlation is not causation. That pair supports a direction, not a dial, and the fuller reading is in brand mentions.

One agency on both sides

It would be easy to read the gap as the industry being behind, and one agency makes that reading harder. That agency is not behind at all. It sells the newer work and does not describe it where a buyer would look. That is a different problem from not knowing about it.

One firm is not behind at all.

Six published digital PR guides, and one agency’s own navigation. Read 4 September 2026Which of the two your own PR reporting looks like

What gets described against what the same firms sell.

Why the two pages differ is not something this page can establish, and the reasons could be several. What is visible is only the difference itself, and where it sits: the explainer is the page that gets read, and the service list is not.

The second error is the opposite one, and it is the easier one to fall into after reading the number above. The citation figures do not license a claim that PR now works because of AI. They describe what got cited. Treating a composition as a mechanism is the same mistake as reading a correlation as a lever, and both are easy to make with numbers this clean.

What digital PR costs

Most agencies quote on request. One publishes its rate card, which is enough to see what a month of this work costs and what the money buys.

The unit is a link.

Digital PR Ltd publishes what that unit buys. A minimum of five links per three-month payment, three to five stories, and every link earned, not bought. Numbers you can hold a quote against.

How the practitioners define their own work

1. 1The definition is bolded on the page: a tool used by brands to evolve their business and build an online presence. Nothing in it names a link.
2. 2The contents list does name one, under measuring ROI, which is where the link comes back as the countable part.
3. 3An agency and a definition can disagree without either being wrong. The agency sells a deliverable, and a deliverable has to be countable.

Read the second block on that page twice. The agency states plainly that it never pays for links and that all of them are earned in national and regional outlets. Set against the citation study, that alignment is accidental and not a strategy. Paid links, the one thing the contract refuses to buy, make up three tenths of one percent of what answer engines cite. A deliverable counted in links and stories is checkable, which is unusual enough to say out loud. It also means the contract is written against the usual measurement: links and stories, both counted. What sits outside it is anything that happens to that coverage afterwards, and nothing in the stated unit accounts for where it gets read. If the second surface matters to you, it has to be asked for, because nothing in the standard unit accounts for it.

Earned coverage that carries a link to your site, with no payment anywhere in it, is our link building service.

What to do with this

Buy digital PR for what it has always done: coverage on sites you do not own and did not pay to appear in, with links and mentions as the by-product. You are paying for the work of earning it, which is a different thing from paying for the placement. That case stands on its own and does not need an AI argument.

Same work, one more thing to count.

Then add one line to the reporting. Alongside pickups, links and referral traffic, ask whether the coverage shows up when an answer engine is asked about your category. A study of 25 million cited links says that is the kind of material these systems draw on. Nobody has tested whether commissioning coverage causes the citation. Observe it; do not buy it as a promise. The procedure for observing it is in how to measure AI visibility.