Foundation built the habits. This is where they combine into real workflows — full campaigns, multiple stakeholders, and the messier judgment calls that come with running things end to end.
This level assumes Foundation's five habits are already automatic. If any of them still feel effortful, worth revisiting Foundation first — Practitioner builds directly on top, not from scratch.
Pick one active campaign to use as your working example throughout this module — every lesson builds on the same one.
Pull up your actual business goal or brief for that campaign, or write a one-line version if none exists formally. Lesson 1 starts here.
Have your real stakeholder list handy. Lesson 4 works best with actual names and roles, not a hypothetical.
This level moves faster than Foundation. Lessons assume you can already do the individual tasks — the focus here is combining them into something that holds up across a full campaign.
Going from "leadership wants more qualified leads next quarter" to an actual campaign structure — budget split, targeting, channel mix — usually means staring at a blank brief.
Use AI to translate a business objective into a first-draft campaign structure across Google Ads and Meta. It won't be final — but a real starting point to react to and edit beats a blank page every time.
Our leadership goal for next quarter is: [goal — e.g. "30% more qualified leads at the same CAC"]. Draft a first-pass campaign structure across Google Ads and Meta: suggested budget split, targeting approach, and funnel stage for each channel. Flag any assumption you're making that I should confirm.
Treating the first draft as final rather than a starting point to react to. The value is having something concrete to edit against — not a finished plan you ship unchanged.
Foundation's creative-testing lesson worked one platform at a time. A real campaign needs a coherent creative approach that flexes by channel and funnel stage — without becoming five disconnected efforts.
Map one creative concept across awareness, consideration, and conversion stages, and adapt it correctly for Meta versus Google's actual formats — same idea, different job at each stage.
Here's our core creative concept: [one-line description]. Adapt it into an awareness-stage Meta ad, a consideration-stage Meta ad, and a Google RSA for the conversion stage. Keep the underlying idea consistent, but let the tone and length shift to match the goal of each stage.
Writing one version and re-purposing it unchanged across every stage. A conversion-stage message and an awareness-stage message are trying to do different jobs — they shouldn't read the same.
Foundation's pacing check was a single prompt. In practice, you need this running every week without it becoming a chore you quietly skip.
Turn the Foundation pacing prompt into a saved, reusable routine — and know what to change about it as the campaign matures, since a check that's useful in week 1 isn't necessarily useful in week 8.
Here's the pacing check prompt I use weekly: [paste your Foundation prompt]. This campaign is now in week [X] of its flight. Suggest one adjustment to this check that would make it more useful at this stage than it was in week 1.
Running the exact same check indefinitely without revisiting whether it's still asking the right question at this stage of the campaign.
A director wants headline numbers, a client wants reassurance, a teammate wants the raw detail — writing three versions of the same update by hand is where a lot of real time goes.
One source update, reshaped by AI into stakeholder-specific versions — including how to handle the harder case, like a client asking a pointed question about underperformance, without producing something evasive or overly defensive.
Here's this week's raw performance summary: [paste]. Write three versions: (1) a 2-sentence version for a director who only wants the headline, (2) a reassuring but honest client-facing version, (3) the full detailed version for a teammate picking up this account. Don't soften real problems in any version — adjust only length and tone.
Letting the client-facing version become evasive about a real problem. Reassuring tone and honest content aren't the same thing — losing the second to protect the first erodes trust faster than the bad news would.
Foundation's diagnostics were single-channel. Real underperformance often spans ad → landing page → conversion, and the actual cause is easy to misattribute to the wrong stage.
Combine GA4 funnel data with ad platform data in one diagnostic pass. A campaign can look completely healthy at the ad level while the real problem sits two steps downstream.
Here's ad-level data from Google Ads/Meta and funnel data from GA4 for the same period, both pasted below. Walk through the funnel stage by stage — ad → click → landing page → conversion — and identify exactly which stage is underperforming relative to the others, not just which channel.
Diagnosing a performance drop using ad-platform data alone. A fine click-through rate with a collapsing conversion rate is a landing page problem, not an ad problem — optimizing the ad won't fix it.
AI can draft the reassuring version of an update. It can't decide whether that reassurance is actually honest. If a stakeholder update sounds better than the underlying numbers justify, that's a decision to fix before you send it — not something to prompt your way around.
Run these against a live campaign, not a hypothetical — Expert builds on judgment you can only get from doing it for real.
