What Campaign Operations Covers

The campaign that spent for nine days against the wrong conversion action

A marketer duplicates last quarter’s campaign to save build time. The new ads and landing page are correct, but the duplicate retains the old conversion goal and audience exclusions. Spend begins. Nine days later, the monthly review celebrates a low reported cost per conversion before somebody notices the “conversion” is an earlier, lower-value action.

The damage is larger than the affected spend. Automated optimisation received the wrong objective, the intended audience was partly excluded, and managers made budget decisions using an invalid comparison. The setup looked live; it was not operating as designed.

Campaign operations owns this middle layer: the controlled work between approved strategy and interpretable performance. It deserves a named owner, a builder-checker separation and a checklist, just as finance and software releases do.

What campaign operations actually covers

Strategy decides audience, offer, objective, channel role, budget logic and success criteria. Creative production develops messages and assets. Campaign operations turns those approved inputs into a correctly configured, traceable and reportable launch.

1. Brief intake: confirm objective, audience, exclusions, markets, channel, budget, dates, conversion definition, landing page, assets, tracking, approvals and dependencies.
2. Build and trafficking: create campaigns, groups or sets, ads, audiences, budgets, schedules, links and platform fields under the approved structure.
3. Naming and tracking: apply controlled names, identifiers, URL parameters, conversion actions and reporting mappings.
4. Pre-launch QA: a person other than the builder checks the build against the approved brief and records evidence.
5. In-flight operation: monitor pacing, delivery, policy status, tracking, anomalies and approved changes.
6. Change logging: record what changed, when, why, by whom and under whose approval.
7. Reporting assembly: reconcile definitions and sources, state attribution basis, compare valid periods and surface decisions.

Campaign operations is not a substitute for strategy, media buying judgement, creative or analytics engineering. It makes their decisions executable and auditable.

Naming conventions and tracking governance

A naming convention is the join key between plan, platform, analytics, CRM, finance and report. If the platform campaign says “LeadGen-final-2,” the URL says “summer,” and the CRM says “Q3 paid,” analysts must reconstruct identity manually. Two operators inventing their own abbreviations split one effort into several rows.

Define each level before building:

• Campaign: stable ID, channel, market, objective, offer or initiative, audience stage and planned period as appropriate.
• Group or ad set: audience, targeting method, placement, bid or another level-specific distinction.
• Ad and asset: concept ID, format, variation, dimensions or language, with the source file using the same identifier.
• Landing page: canonical destination ID and approved variant, separate from query parameters.
• Tracking: controlled source, medium, campaign ID or name, content and other permitted parameters mapped to reporting fields.

Use lowercase or another declared case, one separator, no personal initials as ownership, no unsupported spaces or punctuation and a controlled vocabulary. Do not encode changing performance status into a permanent name. Maintain a data dictionary with field, allowed value, example, system destination and owner.

Google Analytics’ current custom-campaign URL guidance lists UTM parameters including utm_source, utm_medium, utm_campaign, utm_id and utm_content; it also notes that values are case-sensitive and inconsistent values fragment reporting. Use the fields your measurement design needs and test how auto-tagging, redirects, consent and the destination preserve them. Do not paste tagged links into places where the same parameters can expose sensitive or personal data.

Give one owner authority over the convention. Changes require a version, effective date, platform impact and mapping plan for history. Naming is not finished after launch; it governs every new asset, duplication and report.

The pre-launch QA checklist

QA should compare the staged or draft build with an approved brief. The checker should not be the builder. If staffing requires one operator, use a time-separated self-review plus an authorised approver and identify the residual risk rather than calling it independent QA.

Objective and conversion

• Campaign objective and optimisation event match the brief.
• Correct conversion action or goal is included for bidding and reporting; unintended account defaults are excluded where the platform permits.
• Primary versus secondary or observational actions are understood.
• Conversion value, count, window, attribution and source settings match the measurement plan.
• A permitted test reaches the expected systems without creating false production records.

Audience, exclusions and market

• Included audience, prospecting or remarketing logic and expansion settings match approval.
• Existing customers, employees, converters, unsuitable locations and other exclusions are applied at the correct level.
• Geographic option, language, age or other targeting is justified and lawful for the campaign.
• Uploaded audiences came from approved sources and follow platform, privacy and consent requirements.

Budget, bid, schedule and delivery

• Daily or lifetime budget, currency and total plan reconcile.
• Bid strategy and limits match the brief and have the intended conversion inputs.
• Start, end, daypart, account time zone and launch hold are correct.
• Placement, device, network and brand-safety controls match approval.

Creative, copy and policy

• Correct approved versions, aspect ratios, crops, captions, destinations and call-to-action are attached.
• Names, prices, dates, disclaimers, trademarks and claims match source approval.
• Platform previews are inspected across placements; key information is not cropped or hidden.
• Policy review and any regulated-market approval are complete before publishing.

Landing page and tracking

• Final URL is live over the intended protocol, resolves without an unintended redirect and works on mobile.
• Every link carries the approved parameters and lands on the intended page.
• Page, form, thank-you or other event tracking fires under the approved consent state and does not duplicate.
• Submission, error, validation, confirmation and CRM route are tested with authorised test data.
• Analytics source, campaign and content dimensions show the expected values after processing.

Approval and evidence

• Builder, checker, marketing owner and any legal or brand approver are named.
• Brief, screenshots or exports, test record and approval are stored with a stable campaign ID.
• The launch owner and first monitoring time are scheduled.

The three checks that catch the most expensive classes of error

Conversion action: confirm the campaign is optimising toward the intended business event and that reporting columns use the declared action set. Google Ads’ current conversion-goal documentation explains that goals group conversion actions and can influence bidding. A duplicated campaign can inherit defaults or goal choices that are valid elsewhere and wrong here.

Audience exclusions: inspect them at every applicable level. A missing exclusion can spend on existing customers or employees; an inherited exclusion can remove the campaign’s intended market. Confirm list freshness and inclusion/exclusion direction.

Landing-page tracking: click the actual ad preview or final tagged link under a permitted test route, complete the event and verify every required destination. A platform preview, browser event and CRM record answer different questions. One green tag indicator does not prove the full path.

In-flight operations: what a good week looks like

Set a monitoring cadence by spend rate, risk and launch stage. A new high-spend campaign may need checks shortly after launch and more often during the first days; a stable low-spend campaign may not. Define the schedule and alert thresholds before launch.

Pacing

Compare actual spend with planned spend for the elapsed period, accounting for platform delay, day-of-week pattern, learning, seasonality and approved changes. Forecast end-of-period spend under a stated method. Do not make a bid or budget change solely because one morning is slow.

Anomalies

Review cost, delivery, conversion volume and value against an appropriate baseline. Segment by campaign, audience, creative, placement, geography, device or another useful dimension. A cost-per-action spike can come from low volume, broken tracking, auction change, landing failure or genuine performance; diagnose before editing.

Policy and operational status

Check disapprovals, limited delivery, payment, audience, feed, asset and destination alerts. Assign an owner and due time. Never evade a platform policy through minor copy or domain changes; correct the issue or use the authorised appeal route.

Audience and creative condition

Monitor frequency or analogous exposure measures, reach, response and performance by creative, interpreted within platform context. Do not declare “saturation” from one metric alone. Record the evidence and decision.

Change log

For every material edit, capture timestamp and account time zone, campaign ID, field before and after, reason, requester, approver, operator and expected review point. Platform history helps, but it may not preserve the business reason or connected landing-page and CRM changes. Undocumented edits destroy month-to-month explanation.

Reporting that survives month-to-month comparison

Write a metric dictionary. Define spend, impressions, reach, click, landing-page view, lead, qualified lead, sale, conversion value and cost fields; identify the source system, window, currency, timezone, exclusions and processing delay. Keep the definition version with every report.

Separate platform-reported conversions from analytics and CRM outcomes. A platform can attribute or model events under its own rules; analytics can apply a cross-channel model; the CRM records accepted leads and sales under business rules. None is automatically the single truth for every question. Reconcile them without forcing equality.

State attribution on the report. Google Ads’ current attribution-model guidance says data-driven attribution is the default for most conversion actions and last click remains available, while several older rule-based models are no longer supported. Settings and defaults change. Inspect the account and source documentation at reporting time; do the same for Meta, LinkedIn, email and other platforms using their current official help.

Compare like periods: equal days, similar day-of-week composition, currency, market, attribution window, conversion definition and data maturity. Annotate launches, pauses, budget changes, tracking incidents, offers and landing-page releases. Show preliminary data as preliminary rather than silently revising last month.

A monthly report is not a platform export. It should answer: what happened, which data is trusted for which decision, what changed, what remains uncertain and what action is approved next.

Where the work should sit

Strategist builds their own campaigns

This works at low, infrequent volume when the strategist is technically current and can protect build and QA time. It stops working when launches collide, the same person must check their own setup, naming drifts or strategic work is repeatedly postponed. Do not invent one numeric campaign threshold; track build hours, defects and launch queue.

In-house marketing operations hire

This fits sustained multi-platform volume, deep internal systems and daily coordination. The company gains context and direct ownership but must hire, train, provide cover and maintain specialist knowledge. One hire can still become a bottleneck if the role spans build, analytics engineering, CRM administration and every report.

External campaign operations team

This fits variable volume, agencies with repeated builds, or companies that need a managed checker and execution layer without another permanent function. It brings supervision and cover but adds coordination: complete briefs, approval windows, access, change control and brand education. If the client cannot supply those, external work stalls.

Choose from evidence. A remote marketing assistant may support a narrow workflow; a digital marketing support team can cover several connected operations; strategy and media accountability may remain elsewhere.

Access, approvals and brand safety

Use individual platform users, business-manager structures, roles and partner access rather than shared logins. Give builders only required accounts and permissions. Separate build, publish, billing, user administration and high spend authority where the platform and team allow. Enable multi-factor authentication and review connected apps.

Write approval limits: who may create drafts, publish, change budget within a range, exceed monthly plan, alter conversion goals, upload audiences, change landing pages or respond to policy alerts. High-risk claims, regulated categories and new markets need named specialised review.

Route changes through a ticket or controlled request with campaign ID, reason, field, effective time, approval and rollback. Chat may notify; it should not be the sole audit record. Urgent changes still receive a log after containment.

At offboarding, remove the person or partner from every ad, analytics, tag, CRM, email, social, landing-page, reporting and asset system individually. Revoke tokens and connected tools, rotate secrets where necessary, transfer ownership, preserve approved records and verify no scheduled campaigns remain under an inaccessible account.

Build the operating rhythm

• Weekly intake and build window: briefs due by a stated time, dependency review, builds staged and checker assigned. Emergency requests disclose what control or other work they displace.
• Weekly QA and launch window: independent review, corrections, approval, scheduled launch and first-monitoring appointment.
• Weekly performance check: pacing, anomalies, policy, tracking and change decisions recorded.
• Monthly report and retrospective: reconcile sources, explain changes, review incidents and agree next actions.
• Quarterly governance review: update naming, metric dictionary, account access, conversion actions, attribution assumptions, platform changes and checklist versions.

The cadence reduces urgent work because launch needs and approval capacity become visible earlier. It does not ban genuine exceptions. It gives an exception an owner, impact and record. Broader coordination belongs in marketing admin support for campaign management; platform tactics belong in the dedicated Google PPC, Meta and LinkedIn campaign guides.