Direct-Sold Trafficking and Creative QA Inside Ad Manager

Underdelivery is usually reported to you by the advertiser, and by the time it is, the only remaining fix is a makegood at a rate nobody agreed to. The pattern is consistent: the campaign was set up correctly, went live, and drifted. So the work is arranged around three checkpoints, a pre-launch gate, a delivery read on day one, and another on the third day, which is the window in which a pacing failure still costs a schedule change rather than free inventory. The pre-launch gate: Every line is checked against the signed IO for dates, geo, formats, frequency and rate, and then against the network for the things the IO cannot see. Priority band, first, because a sponsorship sold as a roadblock and trafficked as standard is the single most common cause of a missed guarantee. Competitive exclusion labels next, since two exclusions applied to overlapping inventory starve both campaigns and neither one reports an error. Frequency caps multiply when they are set at order and line level together, and the result looks like weak demand rather than a configuration mistake. Creatives get their own pass: HTML5 payload against the weight you promised, TLS on every asset, click macro present and in the right position, cache buster, and VAST wrapper depth counted, because a third-party wrapper nested three deep behind a verification vendor is a video ad that times out on a Vietnamese 4G connection and reports as an error in somebody else's dashboard. Viewability guarantees: We check the arithmetic before the deal is signed where you let us. If Active View measurable rate on the placement sits at 62 percent, a 70 percent viewable guarantee cannot be won on that inventory no matter how it is trafficked, and it is cheaper to say so in the proposal than to discover it in week three. Where a guarantee is already signed, we say what has to change on the page for it to be winnable, and if nothing can, we say that too. In flight: Pacing is set even or front-loaded deliberately per campaign, not left on the default. Delivery is read at the checkpoints and once weekly after that, with an escalation note written for the account manager in language a buyer can read. Discrepancies are worked against the buyer's server with a stated tolerance and an agreed source of truth, so the conversation is arithmetic rather than a standoff. Makegoods are proposed by us with inventory attached, before the advertiser asks. What your commercial team gets: A rate card annotated monthly with what is actually available by geo and format, taken from forecast rather than optimism. A one-page runbook for the question ad ops answers most often, which is why a line is not serving. And a monthly delivery report that shows underdelivery risk while there is still time to act on it. Not included: We do not build or design creatives; we QA what your advertiser sends and reject what fails. We do not invoice, chase payment, or handle billing reconciliation with your finance system. We do not sell, and we do not negotiate programmatic guaranteed deals, though we will traffic and monitor them. We do not take responsibility for delivery on inventory whose traffic we have not reviewed. Who this is not for: Publishers whose direct-sold book runs below roughly forty thousand US dollars a month, where the retainer is a larger number than the revenue at risk. Teams looking for overnight coverage in a European timezone; we work Singapore hours with a defined escalation path outside them. And anyone who wants a headcount rather than a standard, since the point of this engagement is that campaigns are trafficked the same way whoever is on shift.