Journalists need numbers. Numbers give a story authority, they make a headline, & they are easy to verify. Yet almost every business sits on operational data that nobody has ever thought to analyse, while paying for PR campaigns built on opinion.
Data-led PR is the most dependable way to earn genuine coverage & genuine links, & it does not require a research department. It requires a defensible method & the discipline to publish findings honestly, including the inconvenient ones.
Find the data you already hold
Start with an inventory rather than an idea. Walk through the systems your business runs & ask what each one records over time.
- Enquiry & sales records. What people ask for, when they ask, how that mix changes seasonally & regionally.
- Pricing & supplier data. Cost movements you observe across a category, which is often more current than any published index.
- Operational records. Delivery times, failure rates, repair causes, usage patterns. Anything you measure to run the business.
- Support & service logs. The problems customers actually experience, categorised, over years.
- Website & search data. What people search for on your own site is a direct record of unmet demand.
The filter is simple: would this pattern interest someone who is not your customer? A rising trend in a category people care about qualifies. Your internal efficiency metrics do not.
Anonymise & aggregate before you analyse. If a finding could identify an individual customer, it is not publishable regardless of how interesting it is.
When you need to generate data instead
Two reliable routes when internal data will not carry a story.
A survey. Ask a defined population a small number of well-constructed questions. The quality bar is in the design, not the sample size. Avoid leading questions, avoid double-barrelled questions, offer a genuine neutral option, & randomise answer order where relevant. State the sample size, the population, the fieldwork dates & the method whenever you publish a figure.
Public data analysis. Government publications, regulatory filings, open datasets & published statistics are frequently released in formats nobody finds usable. Structuring them, comparing across years or regions, & explaining what changed is a genuine editorial contribution. The data is public; the analysis is yours.
A third route, often overlooked: systematic observation. Checking a defined list of websites, prices, or services against a consistent set of criteria produces original data even though every individual fact was publicly visible.
Design the method before you look at results
This is the step that separates campaigns that survive scrutiny from ones that get quietly ignored.
Write down, in advance: what question you are answering, what you will measure, how you will define each category, what time period you will cover, & what would count as a finding. Fixing this before you see the data prevents the very human tendency to keep slicing until something looks dramatic.
Then be conservative with claims. Correlation is not cause, & a journalist's editor will catch an overreach faster than you expect. A modest, well-supported finding gets published. An exciting, poorly supported one gets one round of questions & then silence.
Write it up for two audiences at once
Your findings need to serve a journalist writing to a deadline & a reader who arrives from search two years later. Build one page that does both.
- Headline finding in a single sentence, with the number.
- Three to five supporting findings, each stated plainly with its figure.
- Context explaining why this matters & what changed relative to previous periods or other regions.
- A named quote from a real person at your company with a real title, offering interpretation rather than promotion.
- Full methodology, including sample, dates, definitions & limitations. State the limitations yourself; it builds credibility rather than undermining it.
- The underlying data in a downloadable, usable format.
- Charts that are readable, correctly labelled, & free to reuse with attribution.
That last point matters more than it seems. A chart a publication can drop straight into an article, correctly credited to a page on your site, is one of the most efficient link-earning devices available.
Present the visuals properly
Bad charts sink good data. Keep to a few rules: start numeric axes at zero unless you have a stated reason not to, label directly rather than relying on a legend where possible, use one colour with emphasis rather than eight competing ones, & put the finding in the chart title instead of a generic description.
Produce each chart at a size that survives being embedded in an article, & include the source line inside the image itself so attribution travels with it.
Pitch the finding, not the campaign
When you approach journalists, lead with the number. The subject line should contain the finding. The first sentence should contain the finding with its context. Only then mention who produced it.
Offer the things that reduce their work: the full dataset, a spokesperson available at short notice, high-resolution charts, & a regional breakdown if you have one. Regional cuts are particularly valuable because they let a national story be localised, which multiplies the outlets that can use it.
The approach here is the same one covered in our guide to digital PR, applied to material that is much easier to place.
Make it an annual asset
The strongest version of this is not a one-off campaign. It is a recurring study.
Repeat the same analysis each year with the same method. The second edition can report change over time, which is a story in itself & is easier to place than the first. By the third edition you have a reference series that journalists return to without being pitched, & a page that accumulates citations continuously.
Consistency of method is what makes this work. Changing the questions between editions destroys comparability & with it the entire value of the series.
What to measure
Track coverage secured & the relevance of the outlets, referring domains earned by the source page, the ongoing search visibility of that page, & any lift in branded search. Then track how much of that persists twelve months later, because durability is the real argument for data-led work over announcement-led work.
If you want help identifying what in your business is publishable, look at how we combine digital PR & content marketing, or start a conversation.