Direct Media Buying on CoinGecko, CMC and CoinDesk

Buy direct, count on your own server, and expect a discrepancy. Across the endemic crypto properties the gap between the publisher dashboard and my ad server usually lands between eight and twenty-five percent on standard display, and closer to the top of that range on anything sold through a reseller. The number you plan against should be the one your own tags produced, and that single decision changes what these placements are worth more than any negotiation does. Where I buy: CoinGecko coin pages and the trending placements, CoinMarketCap coin pages and watchlist units, CoinDesk site and newsletter direct, and the shorter list that carries readers who move money: The Block, Blockworks, DL News, Decrypt, Bankless, and regional publications in Hebrew, Korean and Turkish where your listing warrants them. Everything runs as an insertion order with the publisher's own ad operations team. If a seller will not name the sites, we are not buying. Rates and negotiation: Published rate cards on these properties are opening positions. The levers that move them are flight length, category exclusivity on the coin page, and the timing of your campaign against their quarter. I negotiate against what comparable buyers paid in the same quarter rather than against the card, and I will tell you when a rate is fair and the placement is still wrong for you. Serving and counting: Your tags, in an ad server you control, GAM or Kevel, whichever you already run. Daily discrepancy reconciliation from day one, invalid traffic filtering applied on your side rather than trusted from the publisher's, and frequency capping at the ad server rather than the site. Price-checking scripts and headless sessions are a real share of coin-page traffic, and a placement that looks cheap on CPM often is not once that share is removed. What the campaign is judged on: Wallet connects, funded accounts and first trades, sent server-side into one event schema, with click-through treated as a diagnostic rather than a result. On a listing or a product launch I run a geographic holdout so that any lift claim has something to stand against. Terms I write into the insertion order: A viewability floor with a make-good rather than a viewability report. Exclusion of sanctioned jurisdictions and, where your counsel requires it, of US persons, enforced at the ad server and verified in delivery logs. Creative approval turnaround in hours, because listing dates move. Cancellation with pro-rata refund on unserved impressions. What I refuse: Trending placements on tokens whose volume shows wash patterns, because that is your brand sitting next to it. Reseller inventory with an unnamed site list. Any package that bundles editorial coverage with the media buy, which is what you are paying for when a rate card quotes an article as included; it damages the publication you are buying and it is disclosed badly or not at all. And I will not renew a plan simply because it spent cleanly last quarter. Who this is not for: Projects before listing with no live product, and buyers who want one banner for one week. Below roughly fifty thousand dollars in working media a quarter you cannot read the results across these properties well enough to act on them, and you are better served putting the money into one publisher and one honest test. How it runs: Two weeks to plan, rates and insertion orders, then flighting with weekly reads. I hold the vendor relationships, you hold the contracts and the ad server. Nothing about the way I work requires you to keep me in order to keep the buy.