How to Set Up Google Ads Performance Max for Agencies

To set up Google Ads Performance Max for an agency client, translate the client’s commercial objective into a written brief, verify that conversions represent useful outcomes, prepare approved assets and landing pages, assign owners for data and decisions, and define how the first complete review will work. A reliable agency setup is a documented handoff between the client’s business, the account, the website, and the team that will respond to demand.

This process applies whether the client sells online, generates sales leads, serves local customers, or combines several models. Use the current Google Ads Performance Max overview and campaign creation guidance for the live interface, while using the agency’s operating notes to prevent generic settings from replacing client-specific judgment.

Write the client brief and ownership map

Start the setup with a short brief that answers what the client wants the campaign to change. Define the product, service, market, customer, sales cycle, primary business outcome, acceptable cost or return, and the period in which the result will be judged. “More conversions” is not enough if the client cares about qualified pipeline, contribution margin, retained customers, or appointments that are actually kept.

Record constraints that change campaign decisions: territory, store hours, inventory, minimum order, service capacity, lead response time, sales coverage, promotions, seasonality, legal approvals, and brand rules. For a lead program, describe the difference between a contact and an accepted opportunity. For commerce, record returns, cancellations, stock, shipping, discount, and margin conditions that make revenue meaningful or misleading.

Create an ownership map before launch. Assign a client owner for offer, stock, claims, sales classification, and final approval. Assign agency owners for account changes, creative QA, tagging or feed checks, reporting, and escalation. Name the website, CRM, booking, commerce, or analytics owner who can repair a broken customer path. If a decision has no owner, treat it as a setup risk.

Save the brief with the campaign name, intended scope, launch date, review window, budget authority, and stop conditions. This gives the next strategist enough context to understand why the account was structured as it was.

Audit access, tracking, feeds, and landing pages

Make a checklist of every system the campaign depends on. Confirm that the agency and client have the right access to the advertising account, analytics, conversion configuration, CRM or commerce platform, product or service feed, website, call system, and reporting source. Document ownership without copying secrets into the article, account notes, or shared documents.

Inspect existing conversion actions for duplicates, stale goals, test events, unqualified micro-actions, or values that no longer match the client model. Check whether offline stages or imported sales arrive on a useful schedule. For ecommerce, compare feed prices, stock, shipping, tax or offer language, product URLs, and variant behavior with the site. For lead generation, inspect form validation, phone routing, calendar availability, and CRM fields.

Run a real customer-path test. Submit a realistic lead, make a controlled call, book an appointment, or complete a test order. Verify the event fires once, the right data is captured, the handoff reaches the client, and the record can be classified. Test a duplicate or invalid case where relevant, so the team knows which events must not become primary optimization signals.

Create a baseline that joins platform and client information. Capture recent spend, conversion volume, qualified rate, response time, opportunity or order rate, average value, revenue, margin, cancellations, and returns when they are available. Note the reporting time zone, attribution limits, sales-cycle delay, and data that will remain directional.

Choose campaign scope, goals, and budget deliberately

Decide whether one campaign can support the client’s distinct jobs. Separate products, locations, offers, or business units when they require different budgets, conversion owners, margins, inventory, or landing pages. Keep them together when splitting would remove useful data and no meaningful decision is lost. Write the reason in the setup notes.

Choose the objective and primary conversion actions only after the audit. A high-volume action is not automatically the best signal. If the client wants qualified sales, do not let an easy but unqualified form or page interaction stand in for a later stage without documenting the tradeoff.

Set a sustainable starting budget and record how it relates to demand, capacity, average value, margin, and the expected review window. See agency Google Ads campaign budget planning for related budget considerations, then use the client’s economics rather than a generic benchmark. Record who can approve a change and what evidence is required.

Decide the bid or value approach in line with measurement maturity. Do not make strong efficiency promises while events are duplicated, values are modeled without validation, or the sales cycle is longer than the observation period. Write down any target, value rule, or learning constraint that will affect how the first weeks are interpreted.

Build approved asset groups and destination paths

Create asset groups around a distinct client proposition: a product family, service, audience, location, offer, or landing-page job. Prepare headlines, descriptions, images, video, logos, feeds, prices, terms, and calls to action that the client has approved and can support. Record who approved each claim and when it must be reviewed.

Read each asset by itself and in likely combinations. Remove language that implies universal availability, guaranteed results, unlimited capacity, a price without conditions, or a promotion after its end date. An agency’s QA checklist should catch unsupported claims, but the client remains the source of truth for stock, service eligibility, qualifications, warranties, and terms.

Map the group to a focused final URL or product path. A visitor should be able to understand the offer, eligibility, price or value context, delivery or service boundary, proof, and next action. Test mobile layout, forms, checkout, phone links, chat or booking handoffs, and error states. Log the final URL, asset group, offer, market, owner, and expiry or review date.

Add audience or customer inputs, search themes, brand details, product filters, or exclusions for a clear purpose. Treat them as evidence and controls, not a substitute for measurement or a guarantee of customer quality. If a client has strict brand, legal, or product rules, include the review evidence in the launch record.

Rehearse the launch and record the baseline

Before enabling the campaign, perform a two-person review when possible. The strategist checks the campaign objective, scope, budget, conversion actions, locations, audience inputs, assets, destinations, and exclusions. The client owner checks the offer, price, stock, service boundaries, terms, capacity, brand, and operational handoff. Capture both approvals and any open limitation.

Follow the customer path on a phone and desktop. Use a realistic in-scope case, an out-of-scope case, a failed payment or form case where applicable, and an ordinary customer path. Confirm that the final URL is correct, the business responds, the event records once, the CRM or commerce system has the right details, and the customer receives a useful confirmation.

Set a dated launch baseline. Record the starting settings, previous results, expected lag, seasonality, inventory or staffing conditions, and the first review date. Do not interpret a launch-day signal as a stable performance trend. If the client needs an early health check, separate delivery and tracking checks from the later business-outcome assessment.

Establish reporting and change control before optimization

Build the report around decisions rather than every available metric. Include spend and delivery, primary and secondary conversions, qualified outcomes, revenue or pipeline, margin when available, and operational measures such as response time, cancellations, returns, or stock. Explain the relationship between platform attribution and client records, including any data delay.

Segment by the boundaries that matter: product, service, location, market, asset group, new versus returning customer, lead status, or offer. Avoid presenting a blended average when one profitable line hides a failing one. Show what the agency knows, what the client has verified, and what remains an assumption.

Use a change log with one entry per meaningful decision. State the observation, hypothesis, setting or page affected, owner, date, expected result, review window, and rollback condition. Keep client approval close to implementation evidence. When a change is urgent because a page, price, feed, or conversion path is broken, record it as a risk response rather than implying it was a controlled test.

Use Google’s Performance Max fluctuation guidance when a client requests a reaction to a short period. Review complete periods appropriate to volume and sales cycle, and annotate promotions, holidays, website releases, stock, staffing, and CRM delays. Change one decision family at a time where practical so the next report can explain what happened.

Agency launch checklist

Before pressing launch, confirm that:

1. the client brief names the business outcome, audience, market, offer, capacity, and review window;
2. the account, analytics, CRM, feed, website, call, and reporting access has a clear owner;
3. primary conversions, values, imports, attribution limits, and qualification rules have been reviewed and tested;
4. campaign scope, budget, bid or value approach, and change authority are documented;
5. asset groups, claims, feeds, destinations, offers, and brand or legal approvals are current;
6. the customer path works for representative in-scope and exception cases;
7. reporting distinguishes platform delivery from qualified or verified business outcomes; and
8. the change log, escalation path, pause rule, and first review date are ready.

If the setup is for a local client, compare the local Performance Max setup guide and best local Performance Max framework for additional boundary and capacity checks. adz.so does not configure a client’s live account from this educational article. Use the sequence to make client and agency responsibilities explicit before implementation.