Full-funnel programmatic: how to add a DSP without breaking what works

Most advertisers arrive at programmatic the same way: search and social are working, results have plateaued, and growth now means reaching people who aren't searching yet. That's exactly the job of a DSP — but bolting one onto a healthy operation goes wrong in predictable ways. This is the playbook.

The division of labour

The DV360 vs Google Ads comparison covers the platform differences. This is about what each one does for the funnel — and why measuring the first row with the fourth row's metrics is how good channels get killed.

Rule 1: fund it with incremental budget

Moving budget out of a converting search campaign into a new awareness channel produces the worst possible read: search conversions drop (you defunded it), programmatic shows a high CPA (it's upper funnel), and the conclusion is "programmatic doesn't work". If incremental budget genuinely isn't available, pull from the least efficient existing line — never from the channel you'll be measured against.

Rule 2: don't cannibalize yourself

Three settings prevent most of the damage:

• Exclude existing converters from prospecting — stop paying to reach people who already bought.
• Cross-channel frequency cap — the single biggest efficiency win, as covered in brand safety controls.
• Separate retargeting pools between programmatic and social so the same user isn't chased twice with the same message.

Rule 3: agree how you'll judge it before launching

Write down, before spending a peso: what each campaign is for, which metric represents success, and what result would justify scaling. Awareness gets measured by reach, frequency and lift; performance by conversions and incrementality. The measurement guide covers the models — but the discipline is agreeing on the ruler before you see the numbers.

A 90-day sequence that works

Days 1–30 — Foundation. Tag the site and start building first-party audiences. Launch retargeting first: it's the fastest to prove, since the audience already knows you. Set up brand safety and frequency controls.

Days 31–60 — Prospecting. Add lookalikes from your converters and one or two contextual segments. Introduce a second channel (video or CTV) with the same audience logic. Start a geo holdout if you can.

Days 61–90 — Read and scale. Compare against baseline: did total conversions rise, did search volume for your brand grow, what does the holdout say? Scale what proved out, cut what didn't, and only then consider deals for premium inventory.

Who runs it

The plan above assumes someone operates it weekly — structure, bids, creative rotation, supply curation. That's the real gate, more than budget. If your team isn't there yet, a managed seat gets the operation running now and hands it over when your people are ready. Either way, the sequence stays the same.