How long should PPC client onboarding take?

Thirty days is a workable default for an established account with spending history, because the baseline stage needs a settled window of past data. A brand new account with no history has nothing to baseline, so onboarding compresses to roughly two weeks: access, tracking build and verification, then scope.

Should the agency or the client own the ad account?

The client should own the Meta business portfolio and the Google Ads account, with the agency added as a partner or through a linked manager account. Ownership decides who keeps the data, the learning history and the ad account itself when the relationship ends, so agree it during onboarding rather than at the exit.

What if the client's conversion tracking turns out to be broken?

Stop and fix it before you optimise anything, and say plainly that a baseline drawn from broken data is not comparable to what comes after. Treat the repair as its own small scoped project with a deadline, then restart the baseline window from the date the tracking was verified so every later comparison uses one consistent definition.

Can I use industry benchmarks instead of building a baseline?

No. Public benchmark figures vary widely by account size, category, market and how each source defines its metrics, so they work as orientation but not as a target for a specific client. The account's own recent performance, segmented so brand demand is separated from acquisition, is the only baseline you can be held to fairly.

What documents should exist at the end of onboarding?

Four: an access register listing every account, property and named contact; a one page conversion definitions record; the segmented baseline with its window stated; and the scope document covering decision rights, exclusions, escalation and the provisional target. If a document is missing, the corresponding gate is not actually closed.

PPC client onboarding checklist

A ppc client onboarding checklist is the written sequence an agency or freelancer works through between signing a Meta or Google Ads client and taking real control of the account: collecting access, verifying that conversion tracking reports what the client believes it reports, recording a baseline of past performance, and agreeing in writing who decides what. Most onboarding failures are not skill failures. They are ordering failures.

Why a ppc client onboarding checklist has to be a sequence

Each stage produces the input the next one needs. You cannot verify tracking on an account you cannot open. You cannot record a baseline from numbers you have not verified. You cannot agree a target without a baseline that describes what the account already does. Run the four stages in parallel and you commit to a number in week one that the data contradicts in week four.

Say the quiet part at kickoff: this is not thirty days of doing nothing, it is thirty days of not making irreversible changes while the picture is still forming. An unexplained quiet month reads as neglect.

Days 1 to 5, the access inventory

Request access the day the contract is signed, not the day you plan to start. Platform invitations, two factor handoffs and a finance team reissuing a card can absorb a week between them.

Meta business portfolio, ad account, Page, Instagram account and the pixel or dataset, requested as partner access rather than someone's personal user access, plus the current domain verification status and web event priority list
Google Ads account linked to your manager account, plus Merchant Center, Google Analytics 4, Google Tag Manager and Search Console
Billing: whose card is on file, who receives invoices, and what the account spending limit is set to
Edit or publish rights on the site, CMS or theme, plus the name of the developer who owns them
Read access to the CRM or order system the client uses to count revenue
Named humans: who approves creative, who approves budget, and who answers the phone when the account is disabled at 22:00

Settle ownership now. The business portfolio and the Google Ads account should stay in the client's name, with your agency attached as a partner or linked manager account rather than as owner. That decision determines how painful the eventual separation is.

Days 5 to 12, verify tracking before you trust a number

The most expensive assumption in this trade is that inherited tracking works. Verification is not reading a dashboard. It is causing an event on purpose and watching where it lands.

Run one real transaction or form submission and confirm it appears in the ad platform, in GA4 and in the client's own order system
List every conversion action in the Google Ads account and check which are primary; a secondary action excluded from bidding explains a lot of otherwise baffling behaviour
Hunt for two conversion actions counting the same event, such as a tag manager purchase and a thank you page purchase both firing
Confirm Meta purchase events carry a value and a currency rather than just firing
Write down the attribution setting and lookback window in force on each platform on your start date, because it silently defines every later comparison

If server side events run alongside browser events, confirm deduplication is configured rather than assuming it, since a duplicated purchase inflates the baseline you are about to record. Expect the ad platform and the analytics tool to disagree even when everything is correct; that gap is usually a question of attribution model and window rather than a broken tag.

The output of this stage is a one page conversion definitions record: what counts as a conversion, where it fires, what value it carries, which window it is measured in. Nearly every later disagreement resolves against that page.

Days 10 to 20, record the baseline before you change anything

Baseline means this client's own history, not an industry average. Public benchmark reports vary widely by account, category and market, so treat them as orientation rather than a target. What matters is what this account did before you touched it.

Pull twelve months for shape only: seasonality, quiet periods, spend jumps. Then fix one window for every figure you quote. Ninety days is usually the honest compromise, long enough to hold volume and recent enough to describe the account you inherited.

Then segment it, because an account level number hides what you are being hired to fix. Suppose the last 90 days show ₺180,000 in ad spend against ₺720,000 in platform reported revenue. That is 720,000 ÷ 180,000 = 4.0 ROAS, and it is the figure that tends to end up in the proposal. Now separate brand search. If brand took ₺30,000 and returned ₺300,000, brand is running at 300,000 ÷ 30,000 = 10.0. Everything else is the remainder: 180,000 − 30,000 = ₺150,000 of spend and 720,000 − 300,000 = ₺420,000 of revenue, so 420,000 ÷ 150,000 = 2.8.

The blended 4.0 is roughly 1.4 times the 2.8 you are genuinely being asked to move, and the difference is demand the client already owned. Set the target against the 2.8, hold that same 90 day window and the same revenue definition for the rest of the engagement, and the month three review stops being an argument about which number anyone meant. A ROAS calculator handles the arithmetic; the segmentation is the part that carries the insight. Apply the same split to remarketing against prospecting before you quote anything.

Days 20 to 30, put the scope in writing

By day twenty you have facts instead of impressions. Write the scope while everyone is still cordial.

Decision rights: which changes you make alone, which need a message, which need a meeting. Budget movement inside a campaign is usually yours; a new campaign type or a spend rise past an agreed threshold is not
The target, stated with its window and its definition, built on the segmented baseline rather than the blended one
What is explicitly not yours: landing page changes, product feed quality, photography, discounting, stock availability. Naming these stops the slow drift where every unsolved commercial problem becomes the ads problem
Turnaround and escalation: how fast creative gets approved, and who is reachable during a suspension
The tool inventory, listing every automation layer, script or platform touching the account and which decision each owns. If an AI ad management platform is part of the arrangement, and ZenoxAds works across Meta and Google Ads, record here what it is permitted to change
An exit clause covering what gets handed back and in what form, which belongs in its own handover document

If you push several accounts through this process at once, the load sits in the repetition rather than the thinking, which is where managing ad accounts at agency scale becomes its own discipline.

What not to do in the first thirty days

Do not restructure the account in week one. You destroy the comparison you will need in month three and you inherit the blame for every dip that follows
Do not pause the worst campaign before the baseline is written down. Pause it in week three, with the number recorded
Do not accept "the tracking is fine" from anybody. It is fine surprisingly rarely, and you own it the moment you start optimising against it
Do not commit to a target the baseline has not yet supported. Mark the proposal figure provisional and confirm it at day thirty
Do not start with creative, however tempting. New creative on unverified measurement produces a result you cannot interpret

The four gates, on one page

Each gate is a yes or no, and you do not step past one that is still no.

Gate one, day five: you can open every account and property yourself, with the client still the owner
Gate two, day twelve: you have watched a real conversion land in all three systems and the conversion definitions page exists
Gate three, day twenty: the 90 day baseline is written, segmented, with one window and one revenue definition
Gate four, day thirty: scope, decision rights and the provisional target are agreed in writing

A stalled gate is information rather than failure. An account that cannot clear gate two within twelve days is telling you the first month's real work is measurement, not media buying. Clients would much rather hear that in week two than discover it in month four.