Someone is going to ask what the press release returned. You need an answer that isn't "impressions" and isn't a shrug. Here's a model that holds up under questioning, including the parts where it admits it can't measure something.
Start with total cost, not the invoice
The distribution fee is the smallest part of the cost for most companies.
- Distribution: $150 to $300 mid-tier, $800 to $8,000+ on legacy wires. (Real numbers here.)
- Writing: agency rates run $300 to $1,000. If you wrote it yourself, count your hours honestly, because three hours of a founder's time is not free.
- Review cycles: legal, executive, partner approvals. Two hours across three people is common.
- Outreach: if you pitched journalists alongside distribution, that's the expensive part. Building a list and sending twenty tailored emails is four to six hours.
- Follow-through: newsroom page, social, sales enablement, screenshots archived.
A "$249 press release" is often a $1,200 to $2,000 activity once you count labor. That's not an argument against doing it. It's an argument for doing it only when there's news, and for not doing it eleven times a year out of habit.
The returns you can actually count
Four categories, roughly in descending order of confidence.
1. Direct attributable revenue (rare, but count it)
Sessions from the release link or its placements that convert, tracked with a campaign parameter. This is usually a small number. Sometimes it's zero. Record it anyway, because honesty about the small number is what makes the rest of your model credible.
2. Assisted conversions (the real one)
Conversions where the announcement page or a release referral appears anywhere in the path: first touch, middle touch, any touch. Press rarely closes. It frequently opens.
Pull this from your analytics attribution report for a 90-day window. Apply a fractional credit: if the release is one of four touches, count a quarter of the deal value. This is arbitrary, and you should say so out loud, but it's defensible and consistent.
3. Branded search lift, converted to value
Compare branded search clicks in the 30 days after distribution to the 30 days before. If branded organic clicks rose by 180, and your historical branded-search conversion rate is 4% at an average order value of $900, that's roughly 7 conversions and $6,300 of influenced revenue.
Two caveats you must state: branded search moves for many reasons, and you cannot prove the release caused the lift. Present it as correlated value, not attributed value. Repeat the measurement across several releases and the pattern becomes more trustworthy than any single reading.
4. Substitution value
What would you have paid for the same asset another way? Three legitimate line items:
- Credibility surface: third-party URLs you can cite. If you'd otherwise buy sponsored placements, that has a market price.
- The record itself: a dated, indexed, third-party statement of a fact. For fundraising, procurement, partnership diligence, and due diligence, this has real value even when nobody clicks it.
- Reusable content: the release becomes a newsroom post, a customer email, a sales one-pager, a LinkedIn post, and a section of the next investor update.
Count substitution value conservatively, at what you'd genuinely pay, not at "ad value equivalency," which inflates everything and convinces nobody.
A worked model
Northwind Dental, a four-location practice, announces a new pediatric clinic and a partnership with a local school district.
Costs - Distribution, mid-tier: $249 - Writing, in-house, 3 hours at $75 loaded: $225 - Owner review and quote approval, 1 hour: $90 - Local outreach to 8 regional reporters, 3 hours: $225 - Total: $789
Returns, 90 days - Direct attributable: 2 new-patient form fills from a regional site referral. At $1,100 lifetime value: $2,200. - Assisted: 3 conversions where the announcement page was a middle touch. Quarter credit at $1,100: $825. - Branded search lift: 60 additional branded clicks, 5% conversion, $1,100 value: $3,300 correlated, reported separately, not added to the total. - Substitution: a local coverage clip reused on the homepage and in the school district newsletter. Conservatively $300.
Countable return: $3,325 against $789. Roughly 4x, with a separate correlated figure that may or may not be real.
Note what this model does: it refuses to stack the soft numbers into the hard total. That's the whole discipline.
What attribution cannot tell you
Be explicit about the limits, because whoever you present to will find them otherwise.
- View-through effects are invisible. Someone reads your name in a syndicated article, doesn't click, and searches you six weeks later. Attribution sees organic-branded, not press.
- Offline effects are invisible. The procurement lead who forwarded the release internally leaves no trace.
- Windows truncate. B2B cycles run longer than any default attribution window.
- Multi-touch models are assumptions. Every credit split is a choice, not a measurement.
- Placements decay. A link that converted in month one may be gone by month four. (Syndication explained.)
The right response isn't to give up on measurement. It's to report a hard number, a soft number, and a clearly labeled unmeasured category, and to never quietly merge them.
When the honest answer is "brand"
Some releases have no plausible revenue path, and pretending otherwise is how PR budgets get killed later.
A leadership appointment, a certification, an office opening, an anniversary, a rebrand: these exist to maintain the record and the perception of momentum. The correct ROI statement is:
This release cost $340 and produced a dated, third-party record of a fact we'll reference in sales, hiring, and diligence for years. We do not expect measurable revenue from it.
That sentence is more persuasive than a fabricated 12x. It also sets the precedent that when you do claim revenue impact, you mean it.
The three-question test before you spend
- What decision would this release change for someone outside the company? If none, it's a brand item, so budget it accordingly.
- Would we still do it if the placement report were blank? If yes, you're buying the record, and you should stop evaluating it on traffic.
- What's the cheapest version that accomplishes the actual job? Often a newsroom page plus five targeted emails, not a $1,400 wire.
Measured this way, press releases come out as a modest, reliable, unglamorous tactic, which is what they are. Anyone selling them as a growth channel is selling you the wrong thing. See also how to measure a press release and do press releases still work.
Know the cost before you model the return. [Our pricing is public](/pricing).