Viral Market

Proof

Three campaigns, with the numbers that don’t flatter us included.

Most agencies show you a logo wall. A logo tells you somebody paid once. These are three real campaigns with what was done, what came back, and where the number is softer than it looks.

AI hardware · $11M seed led by a16z

Ghost

One exclusive, 9 million views, seven days — and batch one sold out

The problem

A company nobody had heard of, coming out of stealth in four days, launching a $3,499 computer for running personal AI agents at home — into a category most people had never encountered. Ghost did not even have an Instagram account.

What we did

  • Messaging, press materials, product imagery and a written Q&A prepared before anything was pitched
  • The founder's story made the centre of it: a 19-year-old former quant who left trading to build personal AI
  • One embargoed exclusive placed with TechCrunch, answers and imagery supplied the day before so the reporter had nothing left to chase
  • The founder's launch film published on X at the exact minute the TechCrunch story went live
  • a16z partners, angels and the team quote-posting within minutes, carrying it into far larger networks
  • Instagram pages and creators turning the article into posts and reels for audiences outside tech
  • Second-day coverage, a New York Sun review-unit request and a TBPN guest pitch all handled in launch week
The honest limit. This worked because the news was genuinely strong: an $11m a16z seed, a 19-year-old founder and a product nobody had seen. The method makes a good story travel; it does not manufacture one. Note also that the Instagram pages used have 10.72m combined followers — the honest number is the 1.24m reel plays actually measured, not the follower total.
Blockchain infrastructure, enterprise

Validation Cloud

17 placements, and a name that now sits beside Fortune in search

The problem

An infrastructure company with real enterprise customers and almost no third-party coverage. The product was credible; nothing online said so.

What we did

  • Positioning built before anything was pitched, around enterprise adoption rather than token mechanics
  • Tier-1 business press targeted first, crypto-native press second
  • Executive commentary placed against scheduled market events
  • Every placement amplified to the people who had already asked about it on calls
The honest limit. Placement counts and outlet names are verifiable — the articles are public. View estimates are ours and are not measured. We did not control the headlines.
Decentralised AI infrastructure

0G Labs

50 placements over five months, with no launch to announce

The problem

The hardest brief in PR: sustained coverage for a company with no funding news, no product launch and no event in the window.

What we did

  • Coverage built on the category rather than the company — the AI infrastructure argument, repeatedly
  • Explainers pre-placed ahead of scheduled industry moments so there was something to cite when they arrived
  • Consistency of territory over volume of outlets: the same subject, from the same spokespeople, month after month
  • Measured on referring domains rather than impressions, because that is the number that compounds
The honest limit. Referring domains are independently checkable; view estimates are not. Coverage without news is slower and costs more per placement than coverage with it — that is the honest trade.
Bitcoin / Web3

OPNET

CoinDesk, Cointelegraph and Decrypt inside one week

The problem

A launch window of days, and three tier-1 crypto outlets that all wanted the story first.

What we did

  • Two explainers pre-placed before the announcement, so the context already existed when the news broke
  • An embargo run properly: a named reporter, a fixed time, and no second offer
  • The announcement timed against the pre-placed pieces rather than against the internal launch date
The honest limit. An embargo depends on a reporter agreeing to it. This worked; it does not always, and when it fails you lose the exclusive and keep the news.

Why we print the soft numbers

Every agency in this market quotes a combined audience figure. Validation Cloud’s is 169 million, and we have written on the card above that nobody reached 169 million people — it is the total audience of the outlets, which is a different thing entirely. If you have been in twelve newspapers, you have probably been read in one.

The numbers worth buying on are the dull permanent ones: 0G Labs’ 2,700 referring websites, which you can check yourself in any SEO tool, and a company name that now appears beside Fortune’s in search results and in AI answers.

Talk about your campaign →See what we turned down