Kate LaVail’s recent PRNEWS article regarding analytics (AI and Communications Analytics: A Match Made in Heaven) got a lot right about the current state of measurement for PR. Dashboards, as she explains, are now a staple ingredient in virtually any type of client engagement. She also rightly discusses the increasing value of first-party data. Brands are simply better equipped to reach their priority audiences when they are less dependent on cookies.
What may be a surprise, however, is that her statement that, “it’s hard to put a price on communications, because there isn’t a direct, 1:1 relationship with sales” is no longer true. That code has been cracked, and it’s not all due to AI computing power.
Tracking Individuals Versus Populations
The problem of measuring a PR-to-sales correlation has largely been due to the failure, heretofore, of the methods. We can blame a fixation on tracking tools for warping our thinking. A generation of analysts have been trained on the idea that if we just track individuals closely enough through cookies, or apps spying on your phone, or samba units spying on your TV, that consumer behavior can be deduced. That method fails for advertising, and it doubly fails for earned media.
Why? Because PR’s unit of study (true for marketing as well) is NOT the individual. It is populations of people. Unlike an ad, PR’s objective isn't to persuade one person and repeat that effort as many times as necessary to achieve audience movement. Instead, PR works to sway entire populations with memorable messages that get amplified by other parties. By using a method to analyze populations, not individuals, and by looking for permanent changes, not impulses, impact can be seen in the sales data, visualized and fully valued. And that value is enormous. Our 24-month analysis of PR campaigns for a B2C client proved that as much as $27.10 in sales were generated by every $1.00 invested in PR.
Enter Earned Lift Models
Here's how: Bayesian analysis can determine not only what has happened with a planned PR campaign, but what would have happened in its absence. This kind of experiment-based observation measures the difference of differences in large data sets, allowing analysts to correlate PR launches to product sales lift. Today, ELMs—“Earned Lift Models”—as well as Large Language Models (LLMs) are combining to bring the era of AVEs—Ad Value Equivalents—to an end. If you’re ready to shout, “Death to AVE!,” then you’re ready to embrace ELMs.
Designed by economists and data scientists, Hahn’s patent-pending Earned Lift Model helps make sense of a brand’s previous and future earned media placement investments in terms that matter to your business’ bottom line ROI. The process calculates the expected business impact of your brand’s media mentions by examining earned media hits and observable changes in customer behaviors, and ultimately sales. Simply put, transactions before and after a media mention are measured.
The saying, “advertising is what you pay for, but publicity is what you pray for…” reminds us that earned coverage is vastly more valuable than paid placement. ELMs uncover this value by measuring the sales impact of every earned mention for up to five prior years then correlating types of mentions with known sales data. Even better, a well-constructed ELM is capable of making strategic predictions to improve decision-making and campaign outcomes. That’s a much better starting point for constructing PR plans versus the shiny object approach which can be simply gambles in a nicely designed slide deck.
Tim Weinheimer is the general manager of Hahn Labs, a marketing and PR measurement company based in Austin, Texas. Michael Griebe is Hahn Labs’ chief data officer.
[Editor's Note: The writer’s views do not necessarily reflect those of PRNEWS. We invite opposing essays from readers.]