When to start a PPC agency's ninety day plan and what data to gather

What to take away

Why the ninety day window changed

A ninety day plan used to mean a spreadsheet of keywords and a monthly report. That is no longer enough. Budget holders want to see decisions tied to evidence, and privacy rules mean some data you once took for granted is thinner. In the UK those rules are UK GDPR and PECR, enforced by the ICO, and they decide how much audience and conversion data reaches your reports.

Google's guidance on how to be successful with Google Ads sets out the basics: define measurable goals, organise campaigns around themes, and improve one element at a time. A ninety day window suits that rhythm.

The window also has to accommodate procurement. If the contract, tracking access or data-sharing agreement is unsigned on day one, the first fortnight disappears before any useful evidence arrives. Chase those items before the plan is written.

If the strategy behind the window is still unclear, our PPC agencies strategy and planning guide explains how objectives, budgets and channels fit together before the clock starts.

When to start the ninety day plan

The trigger is not a date in the calendar. Start when three things are true: the contract or letter of engagement is signed, you have admin access to the Google Ads account and to analytics, and the business has agreed in writing what counts as a lead. If any one of those is missing, the first fortnight goes on chasing access instead of evidence.

In practice that means the fortnight before the quarter you want to influence, so the first budget decision the plan touches is the one it was built from. If your market has a seasonal peak, work back from it so the controlled test in weeks three to six lands before demand lifts rather than during it.

The data to gather in the first fortnight

Start with what you already own: account history, conversion definitions, margin data and any first-party lists. Then pull the reports those live in — the Google Ads conversions report, the search terms report and auction insights, plus the matching GA4 acquisition and conversion-path views. Google Tag Manager holds the tracking setup and Consent Mode decides how much of it fires, a CRM export supplies leads and closed-won deals, and LinkedIn Campaign Manager covers B2B audience and lead-form data. For demand you cannot see in the account, Google Ads' Keyword Planner and Google Trends give direction on volume, and the margin file supplies the revenue, gross margin and refund fields that sit behind a lead.

Write down the definition of a lead or conversion before you touch bid strategy. Two teams can report the same account and disagree on the numbers simply because they count enquiries differently. Record it once, in the Google Ads conversion action and the matching GA4 event, and name the person allowed to change it.

For B2B clients, LinkedIn's comprehensive guide to LinkedIn ads is a sensible reference for audience targeting and format choice, because those decisions shape the lead quality you can measure later.

Ad blocking still removes part of your audience before an impression is served, so the reach figures in the plan should come from viewable impressions rather than served ones. The IAB UK guidance on ad blocking for publishers explains why, and it is worth reading before you promise reach the account cannot deliver.

A before and after view

The table below is an illustrative example for a team paying £4,000 a month in media. It shows what changes when data replaces assumption.

Sequencing the ninety days

Weeks one and two are for data collection and access. Weeks three to six run a controlled test on one theme. Weeks seven to ten scale what worked and cut what did not. The final fortnight documents decisions for the next cycle, and week one begins only once the contract, the account access and the definition of a lead are all in place.

Build a simple one-page timeline. Anyone in the business should be able to see which theme is under test this fortnight and what result would change the plan.

Avoid the common trap of changing three variables at once. Our guide to PPC agencies strategy mistakes in England covers the errors that most often derail a plan mid-quarter, including unclear ownership and untested assumptions.

Owners and checkpoints

Every data source needs a named owner: the PPC manager owns the account and bid data, the analytics or measurement lead owns conversion tracking and Consent Mode, the CRM or sales operations lead owns lead quality and closed-won data, and finance owns the margin file. If nobody owns the margin file, the plan drifts toward cheap leads that never convert. Set a fortnightly checkpoint and keep it short.

Keep the checkpoint to thirty minutes. Two numbers per theme, then a decision: continue, adjust or stop.

Write down what would make you stop a test. A plan that cannot fail quietly will fail expensively. Review spend against leads, then leads against revenue, in that order.

Common questions

How long should a ninety day plan take to build?

Two to three working days for the plan itself, then a fortnight of data collection. The plan should be usable by week three, not perfect by week twelve.

Which data sources matter most?

Account history, conversion definitions and margin data come first. External sources such as platform guides and publisher standards help you interpret them, not replace them.

What if the client has no historical data?

Run a small, tightly themed test for four weeks and treat the results as your baseline. Label every projection as an estimate until you have it.

When should the plan be reviewed?

At fortnightly checkpoints, with one formal review at day 45. Change direction only when the data supports it, not because a month has passed.