X Ads Campaign Build for Publishers and Considered Buys

X is worth buying for two jobs: reaching people by what they are posting about right now, and re-engaging an audience that already knows you, at a cost per click that in our accounts sits well below the LinkedIn equivalent. It is not worth buying to generate cold B2B leads. We say that before quoting rather than in the third-month review. The build, in order: Measurement before spend. The pixel deployed and verified, the Conversions API sending server-side events with shared event IDs so nothing is counted twice, and a short list of conversion events that mean something rather than eleven micro-events nobody will read. If the measurement layer cannot be in place before launch, we move the launch date rather than start blind. Objectives: Traffic or website conversions for anything with a real next step. Video views only where the video is the product rather than the wrapper. We build against the objective the buyer will be judged on, which occasionally means declining the objective the platform recommends. An engagements objective buys engagement, and engagement is not something you can put into a forecast or defend in a budget meeting. Targeting we use: Keyword targeting against terms people are genuinely posting, which is the capability this platform has and the others do not. This is the one part of a build where we will spend real hours. Conversation and event targeting, where the event has an actual audience rather than being an inventory unit invented in order to be sold. Follower look-alikes seeded from a hand-picked set of accounts, chosen by reading who follows them rather than by whose name is familiar. Custom audiences from your own lists and from retargeting, deduplicated against existing customers. Brand safety, which is not optional here: Adjacency and sensitivity controls set explicitly rather than left at platform defaults. A keyword and account blocklist written for your business rather than lifted from a template. Third-party verification through IAS, DoubleVerify or Zefr where you hold a contract with one. In publishing accounts we also block adjacency to your own newsroom's coverage, because a paid unit sitting beside your reporters' story creates a problem that lands on an editor's desk, not a marketer's. What we decline: Takeover placements. They are priced in the six figures per day, they buy reach no attribution model will resolve afterwards, and if your executives want one, the media agency that sells it will be delighted to help. Automated engagement tactics of any kind. Follower-count objectives. And amplification of organic posts chosen because they did well organically, which selects for what your existing followers liked rather than what reaches anyone new. Handover: A built and tested account, creative specifications with the supplied assets checked against them, the targeting rationale in writing, a readout after the first four weeks recommending continue, restructure or stop, and a walkthrough with whoever will operate it. You can run it yourself afterwards. Many clients do, and that is a normal outcome of this engagement rather than a failure of it. Who this is not for: Businesses selling to buyers who are not on the platform, which in our experience covers most freight and industrial procurement and most home and furniture trade channels. We know that because we tried it in those accounts at our own risk and it did not work. It is also not for anyone who needs the platform defended as a safe environment to their own executives. We can control adjacency. We cannot control the platform.