How to Measure Crypto PR Results: A KPI Framework for Founders

Most crypto PR reports measure activity and present it as a result. A monthly document lists pieces placed, releases sent and pitches made, and the implicit message is that more of these numbers means the work is succeeding. None of them says whether anyone who matters actually saw the coverage, believed it, or acted on it. This framework replaces that report with eight metrics that measure what the activity actually produced, how to calculate each one, and what a genuinely strong or weak result looks like. It is written for a founder or comms lead who is already running PR, is receiving reports, and cannot tell from them whether the work is producing anything.

Why Do Most Crypto PR Reports Measure the Wrong Thing?

Crypto PR metrics fall into three categories, and most reports stop at the first one.

Input metrics measure activity: press releases sent, pitches made, journalists contacted, outreach emails delivered. These describe effort, not effect. A month with forty pitches and zero placements is not a productive month, whatever the input count implies.

Output metrics measure what that activity produced: pieces placed, reach, domain authority, share of voice. These are a genuine step forward from input metrics because they describe an outcome that exists independently of the agency's own account of its effort. A placement in a named publication with a verifiable domain authority score is a fact a client can check, not a claim the agency makes about itself.

Outcome metrics measure what that output changed in the business: inbound interest, investor awareness, community growth attributable to a specific period of coverage, listing-committee recognition. These are the metrics that connect PR spend to a business result, and they are also the hardest to measure with certainty, because attribution across multiple simultaneous marketing activities is never perfectly clean.

Most agency reports stop at output metrics, and many stop at input metrics dressed up as output metrics: a piece count presented without domain authority context, or an aggregate reach figure with no per-placement breakdown. This is convenient for the agency producing the report, because input and basic output metrics are the easiest numbers to make look good regardless of whether the underlying work moved anything. A report showing forty pieces this month looks like progress even if the pieces ran on sites with negligible traffic and zero editorial standing. The founder reading it has no way to tell the difference unless the report is built to show it.

The rest of this framework is built around forcing that visibility: metrics that cannot be inflated by activity volume alone, and a reporting structure that makes the gap between input and outcome explicit rather than hidden inside an aggregate number.

What Should a Crypto PR Report Actually Contain?

‍

Every row in this table should appear in a genuine PR report with its own number, not folded into a single combined score. A report that only gives a combined summary is hiding which of these eight is actually driving the result and which is padding it.

How Do You Set a PR Baseline Before a Campaign Starts?

A campaign without a baseline cannot be evaluated. It can only be described. If nobody recorded what things looked like in week zero, a report in month three showing "improvement" has nothing to improve against, and a founder has no way to distinguish a genuine gain from a number that was never measured before.

Five things should be recorded before any outreach begins.

Current share of voice. Run the same competitor-set calculation described in the framework above, using whatever coverage already exists for your project and your named competitors, before the new campaign starts. This is the number every later share-of-voice report gets compared against.

Existing coverage inventory with domain authority. List every piece of coverage that already exists about the project, with its domain authority score attached. This tells you what quality bar the existing footprint sits at, so a new placement can be judged against that bar rather than in isolation.

Current AI answer visibility. Run the manual AI citation check described later in this guide, before the campaign starts, and record the exact result: named or not, in what position, citing which sources. This is the single most commonly skipped baseline, because most teams do not think to check it until months into a campaign, by which point there is no pre-campaign number to compare against.

Inbound volume. Record how many unprompted enquiries, investor contacts or partnership requests the project was already receiving per month, before the campaign starts. A rise in inbound volume only means something if you know what the volume was beforehand.

Branded search volume. Pull the trailing three months of branded search data from Google Search Console or a keyword tool. This is the cleanest quantitative baseline in the whole framework, because it is not self-reported by anyone and it is directly comparable before and after a major placement.

Recording these five things takes a single working session in week zero. Skipping it means every subsequent report is a description of activity with no reference point, which is exactly the reporting failure this framework exists to fix.

How Do You Measure Share of Voice in Crypto Media?

Share of voice is the proportion of relevant category coverage that names your project, relative to a fixed set of named competitors, within a fixed set of publications, over a fixed period.

To calculate it, define the competitor set and the publication list before you start counting, not after, so the number cannot be adjusted retroactively to look better. Count every article in that publication list, over the period, that mentions your project or a competitor in the set. Your share of voice is your mention count divided by the total mention count across the whole set.

A strong result is a share of voice that rises over consecutive rolling 90-day windows against a stable competitor set. A weak result is a flat or declining share against the same set, even if your own raw placement count is rising, because that means competitors are gaining coverage faster than you are.

The most common way this metric is gamed is by quietly shrinking the competitor set or the publication list between reporting periods, so a static or declining share looks like an improving one against a narrower comparison. Ask explicitly, every time a share-of-voice number is presented, whether the competitor set and publication list are identical to the previous period's.

Review this monthly, on a rolling 90-day basis rather than a single-month snapshot, because monthly coverage volume in crypto media is naturally uneven and a single month's number is a poor basis for a trend judgement.

Why Does Domain Authority Weighting Matter More Than Piece Count?

A single placement on a publication with a domain authority above 90 carries more search visibility and more weight in what AI systems treat as a credible source than ten placements on outlets scoring below 30. Piece count alone tells you nothing about which of those two scenarios you are looking at.

To calculate this, pull the domain authority score for every outlet that ran a piece in the period, using a tool such as Moz or Ahrefs, and report the average alongside the raw count every time. GeniusPR's own published case data illustrates the difference this makes in practice: a Nillion campaign produced 37 pieces averaging a domain authority of 94, a LimeWire campaign produced 70 pieces averaging DA 95, and a Chintai campaign produced 43 pieces averaging DA 70. Reporting the DA average alongside the piece count, as in these examples, tells a client immediately whether the placements are landing on outlets that carry real authority or on lower-tier sites that inflate the count without inflating the quality.

A strong result is a rising or stable average DA held across a growing piece count. A weak result is a piece count that rises while the average DA falls, which usually means the agency is filling a quota with lower-quality outlets rather than maintaining a consistent placement bar.

The most common way this metric is gamed is simply not reporting it at all, and presenting the raw piece count as though it were the complete picture. Ask for the average DA figure specifically if a report does not already include it; its absence is itself informative.

Review this monthly, alongside the piece count, never as a standalone figure disconnected from volume.

What Is Message Pull-Through and How Do You Track It?

Message pull-through measures whether a journalist used your actual framing, your specific claims and your intended positioning, or wrote a piece that used your announcement as a prompt but told a different story in its own words.

To calculate it, read every placement in the period against your original briefing document or press release, and score each one as a pass or fail against a specific question: does this piece accurately represent the core thesis you briefed. Divide the number of passes by the total placement count for the period.

A strong result is a pull-through rate that holds steady or improves as your narrative becomes more established in the market. A weak result is a declining pull-through rate even as placement count rises, which usually means the story is spreading through restatement rather than through journalists engaging with your actual material.

There is no verifiable industry-standard benchmark percentage for this metric, and any report that states one without naming its source should be treated with scepticism. What matters is the direction of your own pull-through rate over time and the specific gap between what was briefed and what was published in any individual piece that failed the check.

The most common way this metric is gamed is not tracking it at all, since it requires someone to actually read every placement against the original brief rather than just counting URLs. Ask specifically whether anyone reviews published pieces against the brief, or whether the report only counts that a piece exists.

Review this monthly, reading every placement individually rather than sampling.

How Do You Measure Post-Publication Distribution Reach?

Post-publication distribution reach measures the additional audience a placement reaches after it publishes, through channels the agency or project controls directly, separate from the outlet's own native readership.

To calculate it, track the reach generated through your owned social channels sharing the piece, newsletter distribution that references it, any paid amplification spend behind it, and syndication onto other properties, all recorded as a figure distinct from the outlet's own reported readership. Most outlets can supply an estimated readership or traffic figure for context, but that figure is not the same as the reach your own distribution activity generated afterwards.

A strong result is a distribution reach figure that represents a meaningful multiple of the outlet's native readership, showing the placement is being actively worked rather than left to generate whatever traffic the outlet itself produces. A weak result is a distribution reach at or near zero, meaning the placement was treated as complete the moment it published.

The most common way this metric is gamed is conflating it with the outlet's own readership figure, presenting a single combined number that makes it impossible to tell how much of the total came from the agency's own distribution effort versus the outlet's baseline traffic.

Review this per campaign or major launch initially, then fold it into the monthly reporting cadence for ongoing retainer work.

How Do You Attribute Inbound and Investor Interest to PR?

This is the metric most PR reports either skip entirely or overstate with false precision. Attribution across simultaneous marketing, sales and community activity is genuinely difficult, and any report that claims a clean, single-cause link between one placement and one investor conversation should be treated carefully.

The honest method is to log every inbound enquiry, investor mention or partnership request at the point of contact, and ask directly how the person found you or what prompted the outreach. Where a specific piece of coverage is named unprompted by the person making contact, record it as a directly attributed inbound signal. Where no source is named, log it as an unattributed inbound signal rather than assuming it came from PR by default.

A strong result is a rising volume of directly attributed inbound signals, specifically naming coverage, over consecutive reporting periods. A weak result is a flat or declining rate of directly attributed signals, even if unattributed inbound volume happens to be rising for other reasons entirely unrelated to PR.

The most common way this metric is gamed is attributing all inbound growth to PR regardless of source, particularly when several marketing activities are running at once. Ask specifically what proportion of logged inbound signals named a specific piece of coverage unprompted, rather than accepting a single aggregate inbound number as PR's achievement.

Review this monthly, logging every contact at the point it happens rather than reconstructing it later from memory.

How Do You Measure Branded Search Lift?

Branded search lift measures whether coverage is prompting people to search for your project by name, which is one of the cleanest available proxies for whether coverage is actually reaching and registering with a real audience.

To calculate it, pull branded search volume from Google Search Console or a keyword research tool for the two weeks immediately following a major placement, and compare it against the trailing baseline recorded before the campaign started. A genuine lift shows as a measurable increase in searches for your project name or close variants, specifically in the window following the placement rather than as part of a longer-term unrelated trend.

A strong result is a clear, dateable spike in branded search volume that aligns with a specific placement date. A weak result is no discernible movement at all, which suggests the coverage reached readers who did not act on it, whatever the outlet's stated readership figure claims.

This metric is rarely gamed so much as quietly skipped, because it is the one number an agency cannot produce for you and cannot massage: a founder pulls branded search data straight from Search Console without any agency involvement. Ask whether anyone has looked at branded search volume around specific placement dates, or whether reach figures are standing in for evidence that anyone went looking for the project afterwards.

Review this per major placement, then as a monthly trend line once enough data points exist.

How Do You Track Citation Frequency in AI Answer Engines?

AI citation frequency measures whether your project is actually named when a prospective customer, investor or journalist asks ChatGPT, Perplexity, Gemini, Claude or Copilot a buying-intent question in your category.

To calculate it, run a fixed set of category-relevant prompts across each platform on a consistent schedule, and record for each one: whether your brand is named at all, where in the response it appears, and which specific URLs the response cites as its source. The full method for running this check yourself is set out in the next section.

A strong result is your project appearing consistently across multiple platforms, cited from your own owned content or from independent tier-one coverage rather than only from paid or promotional sources. A weak result is your project appearing rarely or not at all, or appearing but cited only from low-authority or self-published sources that carry little independent credibility.

The most common way this metric is gamed, or simply missed, is checking it once and reporting that single result as though it were a stable state, when AI answer engines change source selection over time as new coverage is published and indexed. Ask how frequently the check is repeated, not just what the last result showed.

Review this monthly, on the same fixed prompt list every time, so movement is genuinely comparable period to period.

How Do You Measure Exchange and Listing-Committee Awareness?

This metric captures whether institutional and exchange-side audiences recognise a project independently of your own direct outreach to them, which is a genuine signal that coverage has reached beyond a retail or community audience.

Because exchanges and listing committees rarely share formal data on what coverage influenced their awareness of a project, this metric is necessarily qualitative rather than a clean calculation. Track it through direct feedback logged during listing conversations and business development calls: does the counterparty reference specific coverage unprompted, do they ask fewer basic questions about the project because press coverage has already answered them, does a listing conversation move faster than a comparable one for a project with no independent media footprint.

A strong result is a listing or business development contact referencing specific named coverage without being prompted to. A weak result is every listing conversation starting from zero awareness regardless of how much coverage exists, which suggests the coverage generated is not reaching the audiences that actually matter for that specific outcome.

There is no clean formula for this metric, and any report presenting one as a precise calculation is overstating its own precision. Log it qualitatively, per listing cycle, and treat it as directional evidence rather than a number to average.

How Do You Run an AI Citation Check Yourself?

This is a check a founder can run without buying a monitoring tool, in under an hour, and it should be repeated on a fixed schedule so results are genuinely comparable over time.

Build a fixed list of ten to fifteen buying-intent prompts. These should be the actual questions a buyer would type when trying to find a provider in your category, not brand-name searches. For a crypto PR agency, examples include "best crypto PR agency" or "who should I hire to launch my token." The list should stay fixed once built, so each month's check is genuinely comparable to the last.

Run every prompt across ChatGPT, Perplexity, Gemini, Claude and Copilot. Use a fresh session for each platform where possible, since conversation history can influence which sources a model draws on. Run the full list on each platform on the same day each month.

Record three things for every response: whether your brand is named, where it appears in the answer, and which URLs are cited as sources. A brand mentioned in passing at the end of a long list is a different result to one named first with a specific reason given. Record both, not just a binary yes or no.

Pay closer attention to the citation source list than to the mention itself. The specific URLs an AI answer cites tell you which pages are actually influencing what these models say about your category, which is directly actionable: it tells you which domains and pages to prioritise reaching, whether that means external outlets or your own site's content. A mention with no clear citation trail gives you a data point; a citation trail gives you a plan.

Repeating this on a fixed monthly schedule, with the same prompt list, turns a one-off curiosity check into a genuine trend line that shows whether your AI visibility is improving, stagnant or declining as new coverage publishes.

What Benchmarks Should You Set at 30, 90 and 180 Days?

Do not expect outcome metrics to move meaningfully inside the first 30 days. Expecting immediate inbound or search movement from early-stage placements sets an unrealistic bar that makes good early work look like a failure.

What Does a Good Monthly PR Report Look Like?

A genuine monthly report should be structured as follows.

A baseline reminder. Restate the original week-zero baseline figures at the top of every report, so the current numbers are always shown against their actual starting point, not in isolation.

All eight metrics from the framework above, reported individually. Not folded into a single combined score. Each metric gets its own line with its own number.

The domain authority average alongside every piece count. Never a piece count presented alone.

The current AI citation check result, including which specific prompts were run and which URLs were cited.

A qualitative note on inbound and listing-side signals, logged at the point of contact, not reconstructed from memory at report time.

A short explanation of anything that moved significantly, up or down, in plain language rather than as an unexplained number.

Three things should not be in it.

A single combined "impact score" that blends multiple metrics into one number with no visible working. This hides exactly which of the eight underlying metrics is driving the headline figure.

An aggregate reach total that mixes an outlet's own readership claim with your own distribution activity into a single figure with no breakdown between the two.

A piece count presented without its accompanying domain authority average anywhere in the same report.

Which PR Metrics Are Most Commonly Gamed?

1. Raw piece count without domain authority context. The exposing question: what is the average domain authority across this month's placements, and how does it compare to last month's?

2. A shrinking or shifting competitor set in share-of-voice reporting. The exposing question: is the competitor set and publication list in this report identical to the one used last period?

3. An aggregate reach figure that blends outlet readership with owned distribution. The exposing question: how much of this reach figure came from the outlet's own audience, and how much came from our own distribution activity afterwards?

4. Inbound growth attributed entirely to PR regardless of source. The exposing question: what proportion of this month's inbound contacts specifically named a piece of coverage, unprompted, as the reason they reached out?

5. A single AI citation check reported as a stable, ongoing state. The exposing question: when was this specific result last re-checked, on which fixed prompt list, and has it been re-run since?

6. A combined impact score with no visible breakdown. The exposing question: can you show me the individual numbers this score was built from, and how each one was weighted?

The One Principle This Framework Rests On

Every metric in this framework exists to answer one question: did this activity change something a founder can point to outside the PR report itself. Piece counts and reach figures are useful only as far as they connect to a baseline, a comparison and an outcome a business actually cares about. Measure what happened before the campaign, measure the same things afterwards, on the same fixed terms every time, and report the gap honestly, including when the honest answer is that a metric did not move.

GeniusPR runs this exact measurement discipline as part of its SEO and GEO Optimisation practice, tracking AI citation frequency alongside traditional PR metrics for its clients. Readers still deciding which agency to brief can use the companion comparison: Best Crypto PR Agencies in 2026: A Buyer's Evaluation Guide.

Frequently Asked Questions

How do you measure crypto PR results?

Measure across three levels: input metrics like pitches and releases sent, output metrics like placements, domain authority and share of voice, and outcome metrics like inbound interest, branded search lift and AI citation frequency. Most reports stop at inputs and basic outputs. A genuine measurement framework tracks all three levels against a recorded baseline set before the campaign starts.

What are the most important crypto PR KPIs?

The eight metrics that matter most are share of voice, domain authority weighting, message pull-through, post-publication distribution reach, inbound and investor signals, branded search lift, AI citation frequency and exchange or listing-committee awareness. Each should be reported individually with its own calculation shown, not combined into a single score.

How do you calculate share of voice?

Define a fixed competitor set and a fixed publication list before counting begins. Count every mention of your project and every mention of each competitor within that publication list over a set period. Share of voice is your mention count divided by the total mention count across the whole set, reviewed on a rolling 90-day basis.

Is domain authority a good PR metric?

Domain authority is a useful metric when reported alongside piece count, not as a replacement for it. A placement on a publication scoring above 90 carries more search and citation weight than several placements on low-authority sites. Reporting piece count without a domain authority average hides whether a campaign is landing on genuinely credible outlets.

How do you prove PR ROI?

Full financial ROI attribution is difficult in PR because multiple marketing activities usually run simultaneously. The honest approach is to track directly attributed signals, inbound contacts, investor mentions and search lift that specifically reference a piece of coverage, logged at the point of contact, rather than claiming a single clean cause-and-effect line for every business outcome.

What should be in a monthly PR report?

A genuine report restates the original baseline, reports all eight framework metrics individually with their calculations shown, pairs every piece count with a domain authority average, includes the current AI citation check result and its source list, and logs qualitative inbound and listing-side signals. It should not include a single combined impact score with no visible breakdown.

How do I know if my PR agency is working?

Check whether your reports show a recorded baseline, individual metrics rather than a combined score, a domain authority average alongside every piece count, and a repeated AI citation check on a fixed prompt list. If a report cannot answer what the numbers looked like before the campaign started, it is describing activity, not proving a result.

How long before crypto PR shows measurable results?

Output metrics like placements and domain authority can show movement within the first 30 to 60 days. Outcome metrics like inbound signals, branded search lift and AI citation frequency typically need 90 to 180 days of consistent activity before a genuine trend becomes visible. Expecting outcome-level movement inside the first month sets an unrealistic bar.

How do I check if my project appears in ChatGPT answers?

Build a fixed list of ten to fifteen buying-intent questions a real buyer would type in your category. Run each one in a fresh ChatGPT session, and record whether your brand is named, where in the answer it appears, and which URLs are cited as sources. Repeat this on the same fixed list every month to build a genuine trend rather than a one-off snapshot.

What PR metrics should I ignore?

Treat a raw piece count with no domain authority context, a single combined impact score with no visible breakdown, and an aggregate reach figure that blends outlet readership with your own distribution activity as unreliable on their own. Each can be presented to look impressive while concealing whether the underlying work is actually producing quality coverage or a measurable business result.