Message pull-through: meaning in PR reporting
Sascha Kirstein
Short definition
In PR reporting, message pull-through is the share of analysed coverage that carries a predefined key message with its meaning intact. If 44 of 80 relevant articles contain the message, its pull-through is 55 percent.
Also known as: Key message pull-through, Message penetration, Message delivery
When a PR report says "message pull-through", it answers one question: did the messages we wanted to get across actually appear in the coverage? The figure is a percentage per key message. It describes the content of articles, broadcasts or posts. It does not say whether anyone read, understood or believed them.
What message pull-through means in a report
Before a campaign, the team agrees on a small set of key messages. After publication, someone reviews each relevant piece of coverage and records which messages appear, word for word or as a faithful paraphrase. The report then shows, for each message, how much of the coverage carried it.
Other names for it are message penetration and message delivery. None of these terms has one binding formula. So every report should state how its number was calculated.
The formula
Message pull-through = relevant items containing the key message ÷ relevant items to which the message applies × 100
Two details make the number honest:
- Count per item, not per sentence. An article that repeats a message five times counts once.
- Use the right denominator. A message about one product does not belong in the base for coverage about another product.
For a campaign overview, a second figure is useful: the share of items carrying at least one key message. Label it "any-message pull-through" so nobody confuses it with the result for a single message.
A short example
A company launches a service with two key messages: "data stays in the EU" and "set up in one day". Of 80 relevant articles, 44 mention EU data storage and 20 mention the one-day setup. Pull-through is 55 percent for the first message and 25 percent for the second. So the second message rarely makes it into coverage. Check whether the press material gave journalists enough to work with.
How to quantify pull-through reliably
Pull-through looks simple and is easy to inflate. Three habits keep it reproducible:
- Write a codebook first. Define each message, what counts as a paraphrase and what does not.
- Keep the filter stable. Same period, same sources, same duplicate rules when you compare campaigns.
- Check distortions, not just presence. A message that appears in a critical or garbled form should be coded separately, not counted as a hit.
With a lot of coverage, AI tools can suggest matches. A person should still check borderline cases against the codebook. A near-miss paraphrase often changes the meaning.
What pull-through cannot show
The AMEC Integrated Evaluation Framework separates outputs, outtakes and outcomes. Pull-through sits at the output level. It shows that a message reached the coverage. Whether the audience noticed or remembered it needs surveys or recall tests. Whether attitudes or behaviour changed needs outcome data. Report pull-through next to reach and tone, and do not present it as proof of impact.
The full guide, with coding rules and a worked report, is Key message pull-through: definition, origins and calculation.
How aclipp helps
In aclipp you set up your key messages as a custom metric, with one option per message. aclipp assigns them to each clipping automatically during analysis. You can correct any assignment by hand. Dashboards and reports then show, per message, how much coverage carried it.
Frequently asked questions
It is the percentage of relevant coverage that contains a defined key message. A pull-through of 40 percent means that 40 of every 100 analysed items carried that message in recognisable form.
There is no universal benchmark. The rate depends on the message, the topic, the media mix and how strict the coding rules are. Compare a message with its own history under the same rules, not with a figure from another organisation.
The same logic works for AI answers: check whether answers about your brand carry your key messages correctly. Use a fixed set of prompts and providers, and report the result separately from media pull-through, because the units differ.