How do you measure the results of PR?

Most PR reporting fails because nobody wrote down where things stood beforehand. Take a baseline the week before launch.

1. Branded search volume. Impressions and clicks for your brand name in Google Search Console, last three months.
2. Referral traffic. Monthly sessions from referral sources in your analytics tool.
3. Referring domains. The count from Ahrefs, Semrush or a similar tool.
4. AI mentions. Ask ChatGPT, Perplexity and Google AI Mode five or six questions a buyer would ask about your category. Note whether you are named, and who is named instead. Save the answers.
5. Lead sources. How many enquiries a month, and what share have no clear source.
6. Tracking links. UTM-tagged URLs ready for every placement where you control the link.

Output: what you actually got

For each placement, record the title, URL, date, whether it is a distinct article or a syndicated copy, whether it is indexed, and whether the link is followed, nofollow or sponsored. Count distinct articles and syndicated copies separately. Two hundred copies of one release are one announcement.

Outcome: what changed

• Referral visits from each placement, and whether those visitors stayed. High bounce from one title suggests the angle and the audience did not match.
• Branded search month on month. A rise usually means coverage put your name in front of people who then looked you up.
• Rankings for the pages that received links, checked at 30, 60 and 90 days.
• AI mentions. Re-run the same questions from your baseline. Being named where you were not before is a durable result.

Business impact: what it was worth

• Enquiries with a PR source in your CRM.
• "How did you hear about us?" answers that mention the coverage or the publication.
• Deals where the coverage was used in the sale, such as in a proposal or a due diligence pack.

Pick an attribution rule before the campaign and keep it. First-touch suits awareness campaigns. Last-touch suits direct response. Switching rules to make a quiet month look better ruins the trend line.

Add a "how did you hear about us?" field to your enquiry form and ask the question on first calls. Some of PR's influence never shows up cleanly in analytics: someone reads an article on their phone, then searches your name on a laptop a week later.

This is also where systems help. An automation that tags every new enquiry with its source, logs the self-reported answer and posts a weekly summary saves someone from rebuilding the same spreadsheet each month. Our AI Automation work builds exactly that kind of reporting flow.

• Leading with impressions or "potential reach". Context, not a result.
• Counting syndicated copies as separate wins.
• Judging rankings after two weeks. Search engines take time to crawl and weigh new links.
• Ignoring unlinked mentions. A quote in a trade title with no link still shows up in branded search and in what AI engines know about you.
• No baseline. Without one, every number is just a number.

Hypothetical. Say you run a London B2B software firm and you run one campaign: a regional wire release plus three hand-picked placements on SaaS and business titles.

At 30 days, the output is clear: one wire release with syndicated copies, three distinct articles, all indexed, two with followed links. Referral visits are modest, mostly from one of the three titles. Branded search impressions are up on the baseline.

At 90 days, the page that received the followed links has moved up for two mid-volume terms. When you re-run your AI questions, one engine now names you in a category answer. Your CRM shows four enquiries that mention the coverage.

Verdict: the campaign worked on all three layers, one title clearly outperformed, and next quarter's budget shifts towards titles like it.

Every live link from a WebBox campaign lands in your portal with its date and title, and we separate syndicated copies from distinct placements so your output numbers are honest from day one. We ask for your baseline before launch, or help you take one, and we review at 30, 60 and 90 days against it.

We report what the coverage did, not how big it sounds. If one title drove the enquiries and another drove nothing, you will hear that, and the next campaign changes accordingly.