Buy the good inventory. Skip the rest.

Four things, in order: decide which inventory is worth buying, build the audience and suppression logic, produce creative that fits each screen, and verify what was delivered against what you were charged for. The buying platform is the easy part — the value sits in supply curation and measurement, because that is where budgets quietly disappear. A good engagement leaves you with a named inclusion list, curated deals, a verification setup and reporting that a finance director can read without a translator.

Yes, once you buy it deliberately. The waste is real and measured: the ANA logged $26.8bn wasted in one quarter with only 36.5% of spend reaching publishers. But the same programme showed the other side, with 47.1% of spend reaching publishers and MFA exposure down to 0.39% among advertisers enforcing quality rules. The difference is not luck or budget size; it is an inclusion list, curated deals, capped long-tail volume and someone reading the placement report every week. That is the job.

By deciding what we will buy rather than listing what we will not. Blocklists chase an infinite problem; an inclusion list of named sites, apps and curated deals turns the question around. On top of that we run suitability and invalid-traffic verification on every line, cap open-exchange volume, exclude the domains and app bundles that exist only to serve ads, and review the full placement report monthly with you. Anything unexplained comes off the list until it is explained.

It helps a great deal, and its absence is not a blocker. Plenty of strong programmatic work starts with contextual and geographic buying, curated publisher deals and well-made video, all of which perform without an identifier. In parallel we get your own signal into shape: customer lists, site and app behaviour, CRM stages and offline outcomes. Within a couple of months that usually becomes the most valuable targeting asset you have, and it is one nobody can take away from you.

As the reach layer that makes the rest cheaper to convert. Connected TV buys attention that search cannot create and social increasingly has to fight for, and it now trades programmatically with the same targeting and frequency controls as display. We plan it as one budget with your paid search and paid social activity, manage frequency across screens rather than per platform, and judge it on modelled and tested lift rather than the click it never earns.

With several lenses and no illusions about any single one. Platform reporting drives in-flight optimisation, your sales or billing system says what actually closed, blended efficiency describes the business, and geo holdouts answer the causal questions attribution cannot. On the delivery side we verify viewability, invalid traffic and suitability, and share the whole placement report rather than a curated slice. Where the answer is a range, we give you the range and the method behind it.

Happily, and we prefer it. Working in your seat keeps the deal IDs, audiences and log-level history with you, which is exactly how it should be. If you have no seat we can buy through ours and hand over the setup later, documented so it survives the transition. We are also comfortable sitting alongside another agency that owns a different channel — we will share the methodology and the data, because a supply path only one party can inspect is the problem we exist to remove.

Enough to buy a curated list with meaningful frequency, which for most brands means mid five figures a month across a small number of formats. Below that, thin spend spread over many screens produces impressions and little else, and we will say so rather than take the retainer. Fees are a transparent management charge on a scope we agree up front, not an opaque margin inside the media cost — you should always be able to see what the media cost and what we cost.

Delivery quality improves within the first fortnight, because cutting poor supply is immediate. Efficiency on outcomes usually takes six to twelve weeks: the audience signal needs to be onboarded, creative needs a fair read per format, and a holdout takes time to be worth trusting. We report leading indicators weekly so nothing is a black box in the meantime, and we would rather tell you early that a format is not working than defend it for a quarter.

We can, and it makes the media better. Our marketing analytics and server-side tracking teams build the collection layer these campaigns are judged on, so conversions come back from your systems rather than a platform's estimate. It is equally fine to keep measurement elsewhere: we will report against whatever source of truth you nominate and share our method with whoever maintains it.