Industries / Finance and business services / Advertising and marketing / Public relations
Measurement
Measurement is working out what PR activity achieved, from counting coverage to showing changes in what people know, think or do.
6
Milestones achieved, from 1997 to 2025.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1997
Outputs versus outcomes
Walter K. Lindenmann, director of research at Ketchum Public Relations, set out an "effectiveness yardstick" with three levels: outputs (for example, the number of placements), outgrowths, and outcomes measured with before-and-after polls and experiments. AMEC's later evaluation framework (2016, below) uses a similar ladder of outputs, out-takes and outcomes.
Credit
Walter K. Lindenmann, Ketchum
2010
Ad value rejected
At a summit in Barcelona organized by the measurement trade body AMEC and the Institute for Public Relations, 200 delegates from 33 countries "voted overwhelmingly" for seven principles. One says "Advertising Value Equivalents (AVEs) do not measure the value of PR and do not inform future activity." An AVE prices coverage at what the same space would have cost as an advertisement.
Credit
AMEC; Institute for Public Relations
2016
A planning framework
AMEC launched its Integrated Evaluation Framework, a free online tool that takes communicators through seven steps. Users enter objectives, inputs, activity, outputs, out-takes, outcomes and impact, and download the result as a PDF.
Credit
AMEC
2020
Principles revised
AMEC launched Barcelona Principles 3.0 after an international consultation. The first principle says "Setting goals is an absolute prerequisite," and the fifth says "AVEs are not the value of communication."
Credit
AMEC
2023
AVEs persist
The World PR Report 2022–23 of ICCO, the international body of PR agency associations, found that "61% of organisations globally still use AVEs," and 48% in the UK. The communications consultant Stephen Waddington, reporting the figures, notes the sample was "relatively small" and that the major monitoring vendors keep AVEs "quietly embedded" in their products.
LessonA vote against a measure did not remove it from use while the tools kept offering it.
Credit
ICCO; Stephen Waddington
2025
Version 4.0
AMEC unveiled Barcelona Principles 4.0 at its global summit in Vienna, aligned with its evaluation framework and planning tools and addressing "the rise of AI." The fifth principle now reads "Invalid measures such as advertising value equivalents (AVEs) should not be used," and the seventh calls for ethical governance of AI in research and evaluation. In Muck Rack's survey that year, 67% of PR professionals said producing measurable results is the best way to raise PR's standing with leadership.
communicatemagazine.combusiness-school.open.ac.ukmedia.muckrack.com
Credit
AMEC (team lead Richard Bagnall)
Watch for
Earned media in AI answers
Muck Rack's Generative Pulse studies count the links that ChatGPT, Claude, Gemini and Perplexity cite. The December 2025 edition analysed more than 1,000,000 links and found about 94% non-paid, 82% in a broad "earned" group that includes encyclopedias, government, academic and third-party company sites, and just under 25% from journalism. Press releases rose from 1.2% to about 6% of citations between July and December 2025. Muck Rack's own authors later wrote that "we certainly wouldn't tell a communications leader that our research means media relations drives 84% of AI visibility. It doesn't."
Achieved whenan outside party publishes its prompts, code and category labels and finds similar shares.
Claimed by
Muck Rack, December 2025 to September 2026
Open
01
An open AI-citation count
An open repeat of the AI-citation studies: a published set of prompts run on four assistants, every cited link labelled into categories by two people, and the shares reported both in Muck Rack's categories and with "earned" split into journalism, encyclopedias, government, academic and third-party company content.
The briefCredit · unclaimed02
Outcomes, not ad space
A free tool that takes a list of coverage (dates and links) and an organization's own figures (website visits, sign-ups, enquiries), compares the days after each story with matched days before it, and writes a short report in AMEC's outputs, out-takes and outcomes format as a replacement for AVE.
The briefCredit · unclaimed
Open 01 · the brief
An open AI-citation count
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open repeat of the AI-citation studies: a published set of prompts run on four assistants, every cited link labelled into categories by two people, and the shares reported both in Muck Rack's categories and with "earned" split into journalism, encyclopedias, government, academic and third-party company content.
Why it is still open
Muck Rack's reports describe their categories but do not publish the prompts or code, and the company's own authors say the "earned" share should not be read as the share that media relations drives. Coverage of the same data has led with "84%" earned and with "nearly 25%" journalism. Searches found no open repeat.
Built with what exists
the assistants' APIs with web search on; a large language model, or Gemini Decision 1 (see the Note "A thousand questions, one pass"), to sort links into categories in bulk, with people checking a sample; Label Studio for the human labels.
First step
Write 200 prompts in two industries and label their cited links by hand to settle the categories.
Done looks like
5,000 prompts across ten industries, run twice a month apart on four assistants, with category shares, the two labellers' rate of agreement, and all prompts, code and labels published.
Open 02 · the brief
Outcomes, not ad space
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A free tool that takes a list of coverage (dates and links) and an organization's own figures (website visits, sign-ups, enquiries), compares the days after each story with matched days before it, and writes a short report in AMEC's outputs, out-takes and outcomes format as a replacement for AVE.
Why it is still open
In ICCO's survey 61% of organizations still used AVEs, and Waddington writes that the main vendors keep them embedded. AMEC's framework records what users enter but does not calculate results. Searches found vendor tools that count shares or views of coverage, but no free tool tested on campaigns whose results are already known.
Built with what exists
exports from web analytics such as Google Analytics or the open-source Plausible; the Wikipedia pageviews API and Google Trends for public interest; a standard before-and-after time-series method in Python; a large language model to draft the report text from the numbers.
First step
Run the before-and-after comparison by hand on one past campaign with its own web data.
Done looks like
Tested on 20 past campaigns from volunteer teams whose results are known, with the share of campaigns where the tool's estimate points the same way as the known result published, and ten teams using it instead of AVE for one quarter.