Every distribution report leads with a big number. "Potential audience: 84 million." "Impressions: 312,000." Those numbers are not measurements. They're the sum of the monthly traffic of every site that carried your release, which has roughly the same relationship to your results as the population of a city has to your dinner party attendance.
Here's what to measure instead, in the order it's worth your time.
First: decide what the release was for
You can't measure a release that had no job. Before distribution, write one sentence: this release exists so that ___. The honest options are usually:
- Create a dated public record of a fact (funding, acquisition, certification, leadership change).
- Put a claim into the index where search and AI systems can find it.
- Give a journalist something citable to accompany direct outreach.
- Produce third-party URLs for credibility surface.
- Drive traffic to a specific page.
Only the last one is a traffic goal. Most releases aren't, which is why judging them by clicks produces despair.
The metrics worth tracking
1. Verified placements (not placement count)
Open the report. Count the URLs that are live, contain your release intact, and are on domains you'd be comfortable naming. That's your real number, and it's always smaller than the headline.
Then sort them into three buckets: recognizable outlets, legitimate trade or regional sites, and automated long-tail domains. The first two buckets are usable in an "as featured in" context. The third is index presence and nothing more.
Do this in the first week. Many placements are removed within 30 to 90 days, so archive screenshots while they're live. Here's what a real placement report contains.
2. Referral traffic to the linked page
In your analytics, filter to the linked destination page and look at referral sources for the 14 days after distribution.
Set expectations honestly: referral traffic from syndicated placements is typically small. Tens of sessions, not thousands. Automated republishing doesn't generate readers. If you see meaningful referral traffic, it came from a specific site where a human found it. That site is worth identifying, because it is a publication that actually reaches your audience.
Track it with a campaign parameter on the link so you can isolate it cleanly.
3. Branded search volume
This is the best available proxy for whether anything landed. In Search Console, compare impressions and clicks for queries containing your company name in the two weeks before versus the two weeks after.
A visible bump means people encountered your name somewhere and went looking. That's the mechanism press actually works through: indirect, delayed, name-driven. A flat line means the release didn't reach anyone who cared, which is useful information about either the news or the targeting.
4. Direct and organic traffic to the announcement page
Same window, same comparison. People who see your name in a syndicated story often don't click the link. They search or type your URL. Direct traffic lift is real signal, and it's invisible if you only look at referrals.
5. AI citation and retrieval presence
Newer, and increasingly the most interesting one. A few weeks after distribution, query the major AI answer systems with the questions your news should answer: "who makes X," "what did Acme Robotics announce," "what companies offer Y in Ohio."
You're checking two things: whether your fact appears, and whether it's attributed correctly. Press releases are unusually well-shaped for retrieval because they're declarative, dated, and factual, but only if you wrote them that way. Retrieval doesn't care about nofollow. It cares about whether a clean sentence states the fact.
Do this as a repeatable check with a fixed set of prompts so you're comparing like to like over time.
6. Journalist response
If you paired distribution with direct outreach, which you should, count replies rather than opens. Three thoughtful replies from relevant reporters is a better outcome than 400 placements, and it tells you your angle works.
7. Assisted conversions
In your analytics attribution reports, look at whether the announcement page or the release referral appears anywhere in conversion paths, even as a first or middle touch. Press rarely closes anything. It frequently appears early in paths that close months later.
This is the metric that justifies the spend to a CFO, with the caveat that attribution windows lie in both directions.
The metrics to ignore
Potential audience / impressions / reach. Sum of site traffic. Means nothing about your item.
Total placement count, unqualified. 400 placements where 380 are automated aggregators is a smaller result than 40 placements on trade sites your buyers read.
"Pickup rate." Undefined term used differently by every vendor.
Social shares of the wire page. Nobody shares wire pages.
Ad value equivalency. Multiplying column inches by a rate card is a practice PR measurement bodies have discouraged for years. It produces impressive numbers and no information.
Anything that goes up no matter what you publish. If the number would be the same for a great release and a pointless one, it isn't measuring your release.
A simple measurement routine
- Before: write the one-sentence purpose. Record baseline branded search, direct traffic, and your standing AI prompt results.
- Day 1–7: verify placements, archive screenshots, note any real publications in the list.
- Day 14: pull referral, direct, and organic traffic to the linked page. Compare branded search to baseline.
- Day 30: re-run the AI prompt set. Check which placements are still live.
- Quarterly: look at assisted conversions and whether branded search held its new level or decayed.
Five data points, an hour of work total, and you'll know more than any dashboard will tell you.
The honest framing
Most individual press releases produce small, diffuse, hard-to-attribute effects. That's not a failure of measurement. It's the actual nature of the tactic. The value shows up as a pattern across quarters: a company with a consistent record of dated announcements looks different to search engines, AI systems, journalists, and customers than one with nothing.
Measure for that pattern. Judge individual releases on whether they did the specific job you wrote down, and be suspicious of any number that only flatters you.
Worth reading alongside do press releases still work and press release ROI.
See a real placement report before you spend anything. [Here's how it works](/how-it-works).