Where We Buy

What we use it for. Connected TV first. It has the deepest CTV supply through direct publisher relationships, it is the largest independent DSP, and it lets us bring our own measurement partners. For a CTV campaign with a conversion goal, or an omnichannel campaign where a brand study is the north star and you don't have first-party audience data, it is usually where we start. Asset managers are the typical case.

What we watch for. Identity. With a server-to-server UID2 integration, built on emails your customers consented to give you, targeting and attribution on CTV hold up. With pixels and browser-captured IDs, they don't, and since iOS 26 they hold up less. For display and online video with a cost-per-acquisition goal, we often look elsewhere.

A detail that matters in finance. On the finance networks, guaranteed deals open far more inventory than private marketplaces, at a higher price. We buy both, and we say which one carried the result.

What we use it for. Display and online video with a conversion goal, where signed-in Google data makes targeting and attribution work after the cookie workarounds stopped. YouTube, which is only bought here. And the brand lift study comes with the buy, which matters when a budget can't carry a paid study.

What we watch for. The numbers are Google's. A conversion counted in DV360 is checked against your analytics before it reaches your reporting, and the same ad server counts across every platform so no two of them claim the same sale.

What we use it for. Consumer brands, where Amazon's shopper signals are the best signed-in data outside Google, and streaming TV inventory that other DSPs don't reach. For display with a conversion goal, it is the second option we consider after DV360.

What we watch for. It reports on its own terms, and it is built to keep you in its own inventory. We take the reach and read the result in your analytics.

What we use it for. Paid search optimized to your objective rather than to the click: a lead that becomes a customer, a quote that becomes a policy, an order. Bid strategies run against the conversion you define, and each business line gets its own. A lead-generation account run this way cut its cost per lead by more than a third while leads nearly doubled.

What we watch for. Search finds the demand that already exists. When it is the only channel, the cheapest leads are often the wrong ones, so site engagement by channel and the lead-to-customer rate from your records stand beside the cost per lead.

What we use it for. Consumer reach and retargeting with a conversion goal, in your account so the history stays yours. It is in most of our consumer plans and in few of our finance plans.

What we watch for. Attribution that counts generously, and conversion events that fire on the wrong page. Before we spend, the Conversions API is checked against your records, and a "purchase" has to mean a purchase. In health, we check what the pixel sends before it sends anything.

What we use it for. Business audiences where the objective is to be thought of first: a media company selling to agencies and brands, an asset manager reaching advisors, a firm whose buyers have a job title. The brand study is the point. We ask the same question every quarter and pool the results, so each study adds to the last.

What we watch for. The CPMs are the highest in digital, so it is a brand channel with a study attached, not a lead channel. Reach wide at a modest frequency rather than hammering a small list; the lift came at a frequency of about three, not ten.

What we use it for. Context when identity is gone. Reddit targets the conversation rather than the person, which is where display is heading, and it reaches audiences that research before they buy: investors, patients, people comparing products. It is newer in our plans than the platforms above, and it is in our reporting as its own channel.

What we watch for. Tone. A finance or health ad that reads as an ad gets the response Reddit gives ads. Creative built for the place, or none.

What we use it for. Moments where the audience is in one place: an advisor conference, a trade show, a city during a launch. Vistar reaches the screens programmatically. For the screens that matter most, the airport concourses, the streets around the venue, the kiosks inside, we go direct to Clear Channel, Outfront and the venue, because the programmatic marketplace doesn't carry them. Sometimes that means calling every vendor in a city. One conference, end to end.

What we watch for. Impressions in out-of-home are estimates of people passing a screen. We report them as that, and we look for the result where it would show: site traffic from the city during the dates, a brand study where the budget supports one.

What we use it for. The placements no auction sells: a homepage takeover on a Federal Reserve day, a custom video unit, a publisher's own advisor audience, a direct deal on a finance network's streaming inventory. In finance these are a large share of the plan, and most of them underperform the same inventory bought by CPM. We audit every one against its alternative and keep the ones that earn it, at a price we negotiate ourselves.

Who we buy from directly. NBCUniversal, Netflix, Bloomberg, CNBC, Fox, Dow Jones, Yahoo, Investopedia, Reuters, Spotify, Clear Channel and Outfront, to name some. Relationships that go back years, with the people who sell the inventory, not a form.

Both desks. Every large publisher has a direct sales team and a programmatic team, and they do not always talk to each other. We work with both: the programmatic team to open a guaranteed or private deal we can run in our seat, the direct team when a placement or a price only exists on their side. Which desk a placement came through is in your reporting.

What we watch for. Fixed prices and the publisher's own numbers. Every direct placement is ad-served and verified so it is counted the way everything else is, and a sponsorship that can't be measured doesn't get renewed.

What we use it for. One count. Every platform reports its own conversions and claims the same sale. The ad server sits across all of them, so there is one record of what ran, where, and what followed, and your reporting is built from it rather than from eight platforms' opinions of themselves.