PR first or community first is usually argued as a philosophy question and it is not one. Media coverage and a community do different jobs at different points, and the order that works is decided by which asset a specific project is missing.
The failure both directions produce is symmetrical. Coverage that lands with nowhere to send the reader converts nothing. A community with nothing to discuss goes quiet in a month. Neither is a criticism of the channel.
PR first or community first: what each one actually produces
| PR first | Community first | |
|---|---|---|
| Speed | Weeks, and a release can go out sooner | Months to build anything real |
| What it creates | Findability and third-party credibility | Retention and a feedback loop |
| Cost shape | Discrete campaigns | Continuous staffing |
| Who it reaches | People who have not heard of you | People who already arrived |
| Durability | Coverage stays indexed and quotable | Decays the moment attention stops |
| Failure mode | Traffic that lands and leaves | An empty room that reads as no traction |
| Effect on the other | Fills a community with strangers | Gives journalists something real to write about |
1. Do you have news, or only a project?
Journalists cover events, not existence. Funding with named investors, an incident and its handling, original data, a regulatory development affecting you, or a genuinely new capability are stories. A rebrand, a roadmap update or a partnership with no product consequence are not.
With no event, PR reduces to paid distribution: it puts a release in front of an audience but does not produce a reporter deciding you matter. Whether that is worth buying now is a different question from whether PR works. The distinction is set out in how to pitch a crypto journalist.
2. Is there somewhere for the reader to go?
Coverage sends people to your site, your channel and your product within a few hours and then stops. If the channel is empty, the documentation is thin or the product is not usable, that traffic converts at close to nothing and does not come back.
This is the most expensive mistake in the sequence, because the coverage is spent. You cannot re-run the same announcement once the destination is ready.
3. Is there anything to talk about yet?
Communities form around a shared activity. Users of a live product, testers, holders with a stake, people waiting for a specific thing on a known date all have a reason to be in a room together. A project with none of those is asking people to gather around an intention.
Twenty people talking is a community. Two thousand silent members is worse than no room at all, because the ratio is visible to everyone who looks. Getting the first fifty right is covered in how to build a Telegram group from zero.
4. Can you staff it?
Community is a continuous cost, not a campaign. It needs someone answering across time zones, moderation during the hours your team is asleep, and a publishing rhythm that does not break when the team is busy shipping.
PR is discrete. You can run a campaign, stop, and run another in three months without the first one decaying. A team that cannot commit continuous attention should not open the room yet, and should say so rather than open it and let it die.
5. Which asset is your bottleneck?
If people cannot find you, PR is the instrument. If people find you and do not stay, community is. Most teams can answer this from what they already see: traffic with no returning users points one way, a small engaged group with no new arrivals points the other.
Answering it honestly matters more than the choice itself, because both channels are usually funded on the assumption that the other is working.
6. What does the next conversation require?
Exchanges, launchpads and investors read both, and they read them differently. Coverage is a credibility signal that can be verified in a minute. Community is a demand signal, and the number they check is not membership but the ratio of visible activity to it.
If a listing application is imminent, coverage is the faster asset to produce. If the application will be judged on whether anyone actually wants the token, the room has to be real first.
PR first or community first in practice
The decision is made once and implemented weekly, and the implementation is where it fails. Put the answer in writing with the date and the reason, because six weeks later somebody will argue for the other order and nobody will remember what the original evidence was. One line is enough: which asset was missing, what was chosen, and what would have to change for the order to be revisited.
Then give the chosen side the scarce resource. A team that decides on coverage and spends its only marketing hire on daily chat moderation has not decided anything. The test is whether the calendar of the person doing the work matches the answer, not whether the answer exists in a document that everyone agreed to and nobody reads again.
What to agree before signing an agency
For media work the deliverable is placements, not effort, so the contract should name what counts: outlet tier, whether syndicated copies count separately, whether a mention inside a roundup counts at all, and who writes the draft. Agree the approval loop and its deadlines in the same document, because a release that misses its news window is worth a fraction of what it cost to produce.
Ask what happens when nothing lands. Media relations carries a real risk of a quiet month, and the workable arrangements either carry the commitment forward or convert it into paid distribution with the difference stated openly. Payment monthly against a defined scope is normal. Payment per placement moves the incentive toward volume on outlets nobody reads, which is the opposite of what you are buying.
Budgeting a quarter you can defend
Both channels get cut first when a quarter tightens, and the cut lands on whichever has the less legible number. Protect against that by funding community as a staffing line rather than a marketing line, because it behaves like one: the cost is a person’s time, it does not pause cleanly, and restarting it costs more than continuing would have.
Keep coverage money in discrete campaigns tied to events already visible on the roadmap. That shape makes a deferral cheap rather than damaging, and it stops a retainer from consuming the budget during months when there is nothing to announce. A retainer with no event in the quarter is buying readiness, which is a real thing to buy, but it should be bought deliberately and named as such.
Tracking which channel produced what
Tag every link you control and accept that most of the effect will be untagged. Coverage produces searches for your name rather than clicks on the article, so watch branded search and direct arrivals in the days around a placement, not only referral traffic. Community effects appear as returning users and as questions in support that stop being asked at all.
The comparison worth running is before and after, not channel against channel. Pick one number both are meant to move, such as new arrivals who come back within a week, and record it weekly from the start. Deciding what to measure after the campaign has run is how teams end up arguing about attribution instead of about what to do next quarter.
Keep the record somewhere that survives whoever made it. A sheet holding the date, the channel, what was spent and what the weekly number did afterwards is enough, and it is what makes the argument for the next budget in a single screen. Teams that cannot reconstruct what they did six months ago end up rerunning the same debate with the same absence of evidence.
Hiring the person who holds the room
The role sits closer to support than to marketing. The test in an interview is not humour, it is whether the candidate can answer a hostile question about a delay without either promising something or sounding defensive. Ask how they handled a specific incident in a previous room and listen for whether they escalated it or absorbed it, and what the difference cost.
Decide early whether the person is in-house or contracted, because the room learns their voice either way and replacing them costs the same. Write down the escalation path, the things they may never confirm, and who they are allowed to wake at three in the morning. Most community failures are decision-authority failures wearing a different name.
When coverage lands before you are ready
It happens: a journalist publishes on their own schedule, an incident forces a statement, a partner announces first. Triage within the hour rather than the day. Pin a post in the channel that says what is true, put a short note at the top of the site explaining what the arriving reader came for, and make sure a named person is answering for the rest of the day.
Do not open a new channel to catch the traffic. A room created the same day the coverage lands will be empty when the coverage stops, visibly so. Send arrivals where activity already exists, even if that is a smaller space than you would like, and treat the episode as the deadline for the destination work that was already scheduled.
Restarting a room that went quiet
Silence compounds, because each new visitor reads the date of the last message before deciding whether to speak. Restarting is a scheduling problem: fix a weekly time when a named person posts something that requires an answer, and hold it for a month before judging the result. Questions restart rooms. Announcements into a quiet room confirm what the visitor already suspected.
Clear out whatever makes the room look abandoned first: stale pinned messages, dead links, a welcome note describing a milestone that has passed. Then reduce the surface. Two channels people actually use beat eight that are empty, and merging them is the fastest visible change an operator can make in an afternoon.
Running both with one person
Small teams run both, badly, by interleaving them. A workable split is blocks rather than a mix: part of the week on outreach and drafting, the rest on the room, with the two never competing on the same day. Pitching is slow, quiet work, and it is the first thing dropped when a channel needs attention at an inconvenient moment.
Protect the outreach block by scheduling the community cover around it explicitly, even where cover means a pinned message saying when somebody will be back. A room that knows the answer arrives at a fixed hour behaves better than one that expects an answer immediately and receives one about half the time.
Signals that the order was wrong
Revisit the decision when something specific happens rather than on a schedule. A placement in a real outlet that produced no measurable arrival means the destination was not ready. A room that grows only when coverage lands, and shrinks in between, means there is still nothing to discuss. Either finding changes what the next quarter should buy, and neither requires a debate about philosophy.
Give the current order a defined run before judging it. Coverage takes weeks to appear and longer to compound in search results. A room takes about a month to establish a rhythm anyone can feel. Switching before either has had that run produces two half-built assets, which is the outcome both orders exist to avoid.
Preparing the founder for interviews
The person being interviewed is the deliverable, and most preparation goes into the wrong half of the problem. Facts about the product are easy. What needs rehearsing is the answer to the question nobody wants: the delay, the incident, the departure, the number that has not moved. An unprepared answer to that question becomes the quote, whatever else was discussed for the other forty minutes.
Agree what is on the record before the call and put it in the same message that confirms the time. Say what cannot be discussed and why, rather than declining to answer live. Journalists work with that. What they do not forgive is a correction demanded after publication for something the founder said without noticing it was new information.
Using coverage after it lands
A placement keeps working only if you make it reachable. Link it where a visitor deciding whether you are real will look, quote it where credibility is being assessed, and include it in the materials that go to exchanges and investors. Coverage that lives only in the week it was published has been bought and then left on the shelf.
Tell the community about it once, in the room, without the language of an announcement. The audience that already follows you needs no persuading, but it circulates what it sees first. Then record the outlet and the journalist in a file, so the next pitch goes to somebody who has already written about you and knows what happened last time.
PR first or community first: the order that usually works
Open a small room early and keep it small, before there is anything to announce. Build coverage when there is an actual event and a destination worth arriving at. Then let each feed the other: coverage brings strangers the community converts, and community activity gives journalists something concrete to write about next time.
The version that fails is a large announcement into an empty room. The version that quietly works is a small room that was already alive when the coverage landed.
Media relations, outlet selection and distribution are on our crypto PR page, and moderation, staffing and retention on our crypto community management page.