Digital PR Built on Your Own Data, Not a Panel Survey

The last four campaigns run here landed coverage on between 9 and 34 domains each, and 6 to 21 of those coverage pieces carried a link. The rest named the brand without linking, which is the normal ratio and the reason a campaign has to be worth running for the coverage alone before it is worth running for the links. Where the story comes from: Data you already hold. Studios sit on matchmaking times, churn curves, patch-day concurrency, refund reason codes, controller drift rates. Sports and fitness businesses sit on activity records, injury patterns, race entry demographics, equipment failure rates. Publishers sit on reading behaviour nobody has ever put in public. One of those, cut in a way nobody has cut it, is a story. The first two weeks are spent in your warehouse with your data team, and I have ended engagements at the end of those two weeks because there was nothing in there and saying so was worth more than running a campaign anyway. What I refuse to build: A panel survey of 2,000 respondents presented as research. That is the default output of this category, journalists in Copenhagen, Stockholm and London recognise the headline construction on sight, and the people you want reading it are the people who ignore it. Also refused: index pages ranking cities or countries on a metric assembled from four public sources, correlation written up as a finding, and a fabricated statistic pegged to a news event two days late. How it goes out: The dataset gets a methodology page on your own domain with the numbers downloadable as CSV, because a journalist who can check the work files faster and cites the source. One exclusive goes under embargo to the outlet best placed to break it, 48 hours ahead. The wire follows where the story is Nordic, through Ritzau, TT or NTB, and then trade and specialist press in sequence. Every pitch is written for one journalist based on what that person published in the last six months. The list stays under thirty names. Not included: Crisis communication. Executive profile building. Award submissions. Paid syndication that guarantees pickup, because those placements carry no editorial decision behind them and are treated accordingly by everyone downstream. Media training beyond a single preparation session with whoever ends up quoted. Measurement: A coverage log with outlet, journalist, date, URL, whether the mention links, the rel attribute, and whether the linking page is indexed. Referring domains before and after, with links the campaign did not cause separated out rather than counted. Branded search in Search Console. Where it is relevant, I run the same five prompts monthly against the main AI answer engines and record whether your dataset is being cited when the topic comes up. Who this is not for: A company that cannot get legal sign-off to publish a figure about its own product inside three weeks. A team that wants a link count written into the contract. I will not sign one, because the number depends on whether the story is any good and on what else breaks that week, and a seller who signs it is planning to buy the shortfall. I also do not run this for betting operators, where the coverage that comes back is rarely the coverage anyone wanted. In the handover: The dataset, the pitch drafts, the journalist list with notes on who replied and how, and a written account of the stories I rejected and why. Clients tell me that last document is the most useful file in the folder, and it is the one nobody asks for at the start.