Campaign launch automation

A three-platform launch is the same intent expressed three times in three interfaces: campaign, ad set or ad group or ad squad, budgets, geos, placements, naming — an afternoon of click-work where one slipped digit becomes real spend on Monday.

The manual launch today

1. Build the campaign in Meta Ads Manager: objective, ad sets, budgets, placements, naming
2. Rebuild it in TikTok: campaign, ad groups, optimization event, the same targeting re-expressed
3. Rebuild it again in Google: campaign, ad groups, networks, bidding
4. Upload the creatives into each, attach copy, set states
5. Cross-check budgets and naming against the brief — the step that gets skipped at 6pm
6. Schedule the go-live and hope nothing was left 'active' by muscle memory

What breaks

Naming drift between platforms (which later breaks prefix-scoped reporting), a daily budget entered as a lifetime budget, the wrong geo on one ad set, a campaign accidentally live overnight, and the subtle one: objectives that don't map cleanly across platforms getting set by habit instead of intent.

One instruction, native everywhere

Opera builds each platform's real structure — Meta campaign/ad set/ad under the right act_, Google campaign/ad group/ad under the right customer ID, TikTok ad groups, Snapchat ad squads — from one brief, with your naming convention applied identically:

"Create paused prospecting campaigns across Meta, TikTok and Google using this creative pack — $150/day each, JO naming, broad targeting."

The preview lists every object about to be created, per platform, with budgets and names. You approve the list; Opera executes it; the audit log keeps the receipt.

The first 72 hours: pacing the ramp

Go-live isn't the end of the launch workflow — it's the start of its verification window. The pattern worth codifying: a pacing check the morning after (spend per campaign vs plan, flagged outside ±20%), a delivery check at 48 hours (anything with zero spend is a setup problem — disapproved ad, audience too narrow, budget trapped at the wrong level), and a first efficiency read at 72 hours that is explicitly labeled directional — early CAC on a learning-phase campaign is a rumor, not a result. Opera runs these as scheduled checks against the launch's own plan: the brief's budgets become the pacing baselines automatically, so "is the launch doing what we approved?" is a flag, not a Friday-afternoon spreadsheet. The discipline protects against both failure modes of new campaigns: the silent non-starter nobody notices for a week, and the panic edit made on day-two data that the learning phase would have fixed by Thursday.

Exclusions and retargeting, not forgotten

The launch errors that cost the most are omissions: the prospecting ad set missing its existing-customer exclusion, the retargeting campaign without its frequency cap, the geo overlap between two markets' broad audiences. Because Opera builds from the convention page, exclusion lists and audience hygiene are part of the template — applied by default, visible in the preview, impossible to forget at 6pm. The preview's audience column exists precisely so the reviewer can confirm the negative space (who is excluded) as easily as the positive.

Cloning a winning structure to a new market

The second-best launch instruction after a brief is a reference: "build Germany like the US, MA naming, $120/day, German and English ad sets." Opera reads the US structure — campaign shape, ad-set splits, placements, exclusions — and re-expresses it under the new market's prefix, budgets and languages, paused, previewed. Market expansion becomes a parameterized copy of what already works instead of a from-scratch rebuild that drifts from the original in four small ways nobody intended.

The preview, excerpted

What you actually review before anything exists:

#	Platform	Object	Name	Budget	Geo	State
1	Meta	Campaign	US_Prospecting_Meta_Jun26	CBO $150/d	JO	Paused
2	Meta	Ad set	US_Prosp_Broad_18-44	—	JO	Paused
7	TikTok	Ad group	US_Prosp_Broad	$150/d	JO	Paused
11	Google	Campaign	US_Prospecting_Google_Jun26	$150/d	JO	Paused

…and so on through every ad. Senior review happens here — grep the names for the prefix, diff the budgets against the brief, check the geos per ad set — instead of clicking through three ad managers after the fact. Errors get fixed in a table, not in production.

Objective mapping, decided once

"Conversions" is a different machine on each platform — Meta optimizes to a pixel/SDK event under its attribution setting, Google to a conversion action, TikTok to an optimization event, Snap to swipe or view goals. The convention page maps your intent ("optimize to first transfer") to each platform's mechanism once; every launch inherits the mapping. That ends the quiet failure where two markets "launched the same campaign" and optimized to different things for a month.

Paused until a human says go

Everything is created paused — no exceptions, no configuration that changes it. Going live is an explicit human approval, which means the launch can be built Thursday, reviewed Friday, and flipped on Monday at 9am with one click.

Rolling it out

Safe enough for production

Opera is built to touch production reports and live ad accounts without breaking anything:

• No destructive writes. Updates are append-only by default — your existing data and formulas are never overwritten.
• Preview before execution. You see exactly what Opera will change before a single cell is written.
• Campaigns paused by default. New campaigns are created paused, with approvals required before any spend.
• Full audit logs and client-level isolation. Every action is logged, and each client's data and rules stay separate.