How Do You Track Where a PR Story Got Picked Up?

You track it by searching a distinctive phrase from the story itself — not your brand name — repeatedly over the days after it runs. A story spreads faster than anyone tells you it is spreading, and most of what you’re looking for is a near-identical copy of the same text on a domain you’ve never heard of. Your name is a weak handle for that. A verbatim sentence is a strong one.

This is a different job from the ongoing brand sweep described in how to find unlinked mentions of your brand. That one runs forever at a low cadence and asks “is anyone talking about us.” This one is event-driven: one story, one burst, a window of a few days, and a much more specific question — where did this exact piece end up, and did the copies carry a link.

Why a phrase beats a name

A syndicated copy usually keeps the body text intact and changes the furniture: new headline, new byline, house formatting, sometimes a trimmed intro. Your brand name may survive that, or it may not — it might have only ever appeared in a subheading the receiving outlet rewrote.

A distinctive interior sentence survives almost everything. Pick one that would be strange to write by accident — a quote from your spokesperson, an unusual noun pairing, the specific phrasing of a finding — and search it in quotes. That query finds copies your name never would.

If you write the release or the pitch yourself, this is worth designing for: one memorable, unusual sentence somewhere in the middle of the text gives you a tracking handle for months.

One sweep is not a census

However you run the search, you are seeing what one index has crawled and chosen to show. That is a sample of the web, not the web.

The major engines maintain separate indexes with separate crawling and ranking behaviour, and vendors sell them as separate products for that reason — the Bing side of Serply’s search API is documented explicitly as an alternative source whose different ranking algorithms may return different results than Google’s. In practice that means a smaller regional outlet or a trade publication can be sitting in one index and effectively invisible in the other. If a placement matters enough to report on, checking the same phrase against a second engine is a cheap way to find the pickups the first sweep missed.

Neither of them is complete, and it’s worth being honest about that in whatever you report: “these are the pickups we found” is true, and “this is where the story ran” is not.

Count editorial decisions, not URLs

Here is where PR reporting goes wrong most often, and it’s an honesty problem rather than a technical one.

When a story goes out over a wire or a syndication partner, it can appear on dozens of domains within a day. Those are not dozens of publications that read your story and decided it was worth their readers’ time. It is one decision — by the originating outlet, or by whoever bought the feed — and then a distribution mechanism. The copies are the same text, published automatically, usually by sites that would not have covered you otherwise.

Treating that as twenty separate wins inflates the number and, worse, tells you the wrong thing about what worked. The useful question after a campaign is which editors chose the story, and that number is usually small and always more informative. Sort your pickup list into three piles:

• Original coverage — an outlet wrote its own piece, with its own angle and framing. This is the pile that means something.
• Syndicated copies — the same text redistributed through a feed or a partnership. Real distribution, real audience, one decision.
• Aggregator and scraper copies — automated republication with no editorial involvement at all. This is not coverage. It’s plumbing noise that happens to contain your words.

The distinction is the same one behind editorial links versus placed links: what makes a link mean something is a human deciding to add it, and a distribution mechanism is not a human.

Most copies won’t carry a link

Expect this and you’ll be less annoyed by it. Receiving outlets strip outbound links routinely, or convert them to plain text, or replace them with a link back to the wire service rather than to you. The originating piece may link you; the copies frequently don’t.

That’s an editorial and technical habit of the receiving publication, not a judgment about you, and the reasons overlap heavily with the ones in why people mention you without linking. The original outlet is the one worth a polite note if the link is missing there. Chasing a hundred scraper copies for links is effort spent on the pile that was never editorial in the first place — and, since those copies are duplicates of one story, the odds that anyone treats them as a hundred independent endorsements are not good.

The line you don’t cross

There is an industry that will sell you the last section as a product: distribute a release with an optimised link in it to hundreds of sites, and report the resulting URLs as link building. Google’s spam policies name that pattern directly — links with optimised anchor text in articles or press releases distributed across other sites are listed as link spam, not as coverage.

Publishing genuine news and tracking where it lands is normal PR work. Buying mass republication for the links is the thing that gets sites penalised, and it’s the pattern behind unnatural links and manual action. The tracking routine in this post is for finding out what happened. It is not a distribution plan.

What to actually do

1. Before the story goes out, pick the distinctive sentence you’ll search for, and write it down with the launch date.
2. Sweep daily for the first three days, then a couple of times over the following fortnight — pickup trails off fast but doesn’t stop cleanly.
3. Run at least one sweep against a second search engine before you call the list final.
4. Sort every hit into original, syndicated, or scraped, and report the first category as the result — see how to earn backlinks without buying them for what to do with the handful of original pieces that named you and didn’t link.