Fifteen Years of Declaring AVE Dead Did Not Work

Advertising value equivalency (AVE) is a method that prices press coverage at what the same space would have cost to buy as advertising. The alternatives measure outcomes instead: the quality of reach, message pull-through, sentiment and share of voice.

For fifteen years, the global standard for communications measurement said that advertising value equivalency is not the value of communications. It said so in 2010, when practitioners from 33 countries met in Barcelona and AMEC published the principles that took the city’s name. It said so again in the 2015 revision, and again in 2020.

In June 2025 it stopped saying that, and started saying something else.

Principle 5 now reads: “Invalid measures such as advertising value equivalents (AVEs) should not be used. Instead measure and evaluate the contribution of communication by its outcome and impact.”

The previous wording was “AVEs are not the value of communication.” Read those two sentences next to each other, because the difference between them is the most useful thing in the document. The first is a description. It states a fact about the world, and a reader can agree with it completely and then carry on exactly as before, because a fact does not ask anything of you. The second is an instruction with a replacement attached.

The industry spent fifteen years describing the problem accurately. The standard changed because describing it accurately did not work.

Why being right was not enough

The original framing had two weaknesses that are only obvious in hindsight.

The first is the one above: agreement was free. Nobody in a conference session ever had to defend AVE, because the panel format only required them to nod along. Belief and behaviour were never connected, so belief changed and behaviour did not.

The second is more practical. “AVEs are not the value of communication” tells you what to stop doing without telling you what to start doing. If your chief executive asks what the agency retainer bought this quarter, and the only answer you have been given is that the number you used to provide is invalid, you have been handed a vacuum. Vacuums get filled by whatever is nearest, and what is nearest is the number you already know how to produce.

This is why the 2025 wording matters more than a style edit. Adding “instead measure and evaluate the contribution of communication by its outcome and impact” is an acknowledgement that the refusal was never sufficient on its own.

AVE survives on demand, not on belief

The thing that keeps AVE alive is not that anyone thinks it is accurate. Ask the person who put it in the deck and they will usually tell you, unprompted, that it is nonsense.

It survives because it answers the shape of the question it gets asked.

When someone holding a budget asks what the spend achieved, they are asking in a particular format: a figure, in currency, comparable to the other figures in front of them. AVE is the only output the profession has ever produced in that format. It is wrong in a shape that fits the conversation, and a wrong answer in the right shape beats a right answer in the wrong shape almost every time, at least until somebody checks.

Which means this is a demand-side problem. It will not be solved by another article explaining that AVE is invalid, including this one. It gets solved when the person receiving the report stops asking for a single figure, and that only happens when they have been given something better to ask for.

What the number is actually made of

It is worth walking the calculation, because the failure is not a matter of opinion and it collapses under one question.

Take the physical size of the coverage, or its duration in broadcast. Multiply by the outlet’s advertising rate card for equivalent space. Then, commonly, apply a multiplier to account for earned coverage being more credible than paid.

The rate card is the only auditable component. Everything after it fails for a distinct reason.

The multiplier has no source. It is typically somewhere between three and eight, and if you ask where the figure came from, the answer is that it is the multiplier the agency uses. Two agencies can value identical coverage differently by a factor of two, with no procedure for establishing which is correct, because neither is measuring anything external.

The premise contradicts itself. The entire argument for earned coverage is that audiences trust it more precisely because it was not bought. AVE then values it by pricing it as though it had been bought. The metric prices away the exact property it exists to claim credit for.

Nothing in it can be falsified. There is no observation that would show an AVE figure to be wrong, because it does not predict anything. A metric that cannot be wrong is not a measurement. It is a presentation.

Rate cards are list prices. Almost nobody pays them. A number anchored to a price that is not transacted is not a market value even on its own terms.

That is four independent failures in one calculation, any one of which is fatal. This is why the discussion has been settled at standards-body level since 2010 and why the only real question left is what goes in its place.

The four numbers that survive scrutiny

The replacement is not another single figure. Anyone offering you one has reproduced the original mistake with better branding. It is a small set, each answering one question, each traceable to something a sceptical person could go and check.

Share of voice against a named competitor set. What proportion of category conversation mentions you rather than a specific, named list of rivals, over a fixed window. It survives because it is comparative rather than absolute: it does not need to assert a value in currency, only a position relative to the companies you actually compete with for attention. The discipline is fixing the competitor list in advance and leaving it fixed, because a set that gets quietly revised whenever the result is unflattering is worth nothing. The full method and its traps matter more than the formula.

Sentiment trend against your own historical baseline. Not sentiment as a standalone percentage, which means very little, but movement relative to what is normal for your brand. A company that runs at 30 percent positive in an inherently contentious category is not in crisis at 28. A company that normally runs at 70 is in real trouble at 55. Stating the baseline alongside the figure is what makes the number readable by someone who does not work in your category.

Message pull-through. Of the coverage that ran, how much carried the argument you were actually trying to land, as opposed to mentioning you while telling a different story. This is the most labour-intensive of the four because somebody has to read the coverage rather than count it. It is also the closest to the real job, because coverage that names you while framing you badly is not a win in any honest accounting.

Correlation with a business signal, stated as correlation. Branded search demand, direct traffic, or category interest tracked across the same window as major coverage. The discipline lives entirely in the verb. Showing that branded search rose in the week of a placement is a legitimate, useful observation. Claiming the placement generated a specific number of leads, without the attribution infrastructure to support it, is AVE wearing a different suit: a confident figure with nothing underneath it.

The conversation this actually requires

Most writing on this subject stops at the framework, which is convenient, because the framework is the easy part. The hard part is the meeting where you produce four numbers for someone who wanted one.

A few things make that meeting go better.

Bring the replacement to the same conversation where you retire the old number. Removing AVE and promising a better approach next quarter reads as a retreat. Removing it while putting a share of voice trend and a sentiment baseline on the table is a swap, and swaps are easier to approve than subtractions.

Say what each number is for, in one sentence each. Most resistance is not to rigour, it is to being handed a dashboard with no stated purpose. “This one tells us whether we are gaining ground on the three companies we lose deals to” is a sentence that survives being repeated to a board by someone who was not in the room.

Let the numbers earn it over two or three quarters. The strongest argument is not the one you make in the meeting. It is the share of voice trend that correctly flagged a competitor’s push a month before sales felt it, or the sentiment baseline that caught a problem while it was still small. Once a set of numbers has predicted something real, appetite for a single unexplainable figure tends to fade without further argument.

Concede the genuine loss. The honest position is that you are giving up comparability with whatever was reported before, and that earned coverage does not reduce to one currency figure. Saying so plainly buys more credibility than pretending the transition is costless, and it pre-empts the obvious objection before somebody else raises it.

The only test that matters

Any measurement framework can be made to look complete once. The real test is whether it holds the second time, when the person who received it comes back a week later, picks one number, and asks how it was built.

AVE fails that immediately. The rate card is defensible and everything above it is an assertion, which is why the follow-up question has no good answer and why the metric has spent fifteen years being rejected by the bodies that set the standards while remaining in circulation anyway.

The four numbers above survive it, because each one resolves to something real: a mention that exists and can be read, a sentiment score against a history you can display, a piece of coverage somebody assessed against the message, a traffic line with a date on it.

That is a slower thing to build than a large figure in a slide, and it is the only version of this that still means something the second time it is questioned. Which, as the 2025 revision quietly concedes, is a lesson the industry took fifteen years to write down.

Common questions

What should replace AVE in PR reporting?

A small set of numbers that each answer one question and can each be audited: share of voice against a named competitor set, sentiment trend against your own historical baseline, message pull-through in the coverage that actually ran, and correlation with a business signal stated as correlation rather than cause. None of them requires an arbitrary multiplier.

What do the Barcelona Principles say about AVE?

Principle 5 has rejected AVEs since 2010. The wording changed in Barcelona Principles 4.0, published by AMEC in June 2025, from 'AVEs are not the value of communication' to 'Invalid measures such as advertising value equivalents (AVEs) should not be used. Instead measure and evaluate the contribution of communication by its outcome and impact.' The first version described a fact. The second gives an instruction and names a replacement.

Why is AVE considered a discredited metric?

It multiplies the size of a piece of coverage by the outlet's advertising rate card, then often applies a multiplier for earned credibility. The rate card is the only auditable part. The multiplier has no external standard and varies by whoever is presenting it, and the premise contradicts itself: earned coverage is considered valuable because it is not paid for, so pricing it as though it were paid space prices away the exact quality being claimed.

Why do PR teams still use AVE if everyone says it is wrong?

Because it answers the shape of the question it gets asked. When someone with a budget asks what the spend achieved, AVE produces a single currency figure that looks like an answer in the same units as everything else they review. It survives on demand from the person receiving the report rather than on belief from the person writing it.

Is there a defensible dollar-value alternative to AVE?

Not in the same form, and that is the honest answer rather than an evasive one. Earned coverage does not convert cleanly into one currency figure, and every attempt to force it there reproduces the original error in new packaging. The workable path is a small number of defensible metrics that each answer something specific.

How do you tell a CFO that PR cannot be reduced to one number?

Not by refusing, which reads as evasion. Bring the replacement to the same meeting, show which question each number answers, and let a share of voice trend or a sentiment baseline earn credibility by correctly calling something over a couple of quarters. Appetite for a single unexplainable figure fades once a defensible set has predicted something real.