Prove the ads made money — and stop paying for clicks that never could

Track the business events the engagement is meant to support: enquiries, qualification and recorded sales, alongside spend and traffic measures. Define those events with the client before comparing results. A click can be useful without becoming a lead, and a lead can be legitimate without buying. The reporting model should make those distinctions visible.

Confirm the account identifier, business name and relationship to the client before requesting or accepting access. Ask about former agencies, multiple branches and duplicate account names. Verify that the granted role supports the work you intend to perform. A completed invitation is only the beginning; the agency should check the actual asset and usable permissions.

The agency usually cannot see every offline conversation, booked job or refund without the business supplying that information. Outcome updates connect marketing enquiries with commercial results. Name the person responsible and show them the expected routine. If updates are incomplete, label the reporting limitation rather than treating the absence of recorded revenue as a confirmed absence of sales.

No. Return on ad spend compares attributed or recorded revenue with advertising spend under the chosen definitions. It does not automatically subtract labour, fulfilment, overhead or other business costs. Explain what the report includes and the time period it covers. Use the measure to inform a discussion about performance, while the client evaluates profitability using its own cost information.

They describe different points in the recorded journey. First touch identifies the earlier recorded source; last touch identifies the later recorded source associated with the conversion path. Neither is a complete explanation of everything that influenced a customer. Use consistent definitions in reports and avoid comparing figures built from different attribution rules as if they were interchangeable.

The product describes supported Google Ads offline conversion and Meta Purchase-event workflows. Check the required setup, identifiers and delivery state for the account you are using. Recording a sale alone does not guarantee that a platform can match or use it. Validate a controlled example and review errors or missing data before relying on the feedback loop.

Submit a labelled test enquiry through the real form and verify the expected record and source information. Check the confirmation experience and who receives the enquiry. Review the page on the devices relevant to the campaign. Keep test activity distinguishable from customer results so validation does not inflate lead or revenue reporting.

No. Signals indicate activity worth reviewing, not a proven identity or intention. Look for corroborating patterns and consider legitimate reasons for unusual network characteristics. Document the evidence before blocking. This is especially important when explaining results to clients: a suspicious visit count should not be presented as a confirmed amount stolen by competitors or bots.

No single exclusion method covers every unwanted visit, and platform support has practical boundaries. Use the supported workflow as part of a wider review of traffic, enquiries and campaign settings. Verify the intended account and inspect the effect after a change. Avoid promising a fixed saving or complete protection from a tool setting alone.

Choose a cadence that fits enquiry volume and the client’s ability to supply outcomes. A new campaign may need closer attention while definitions and tracking are being validated. Review meaningful samples and recurring patterns rather than reacting to every isolated rejection. End each review with a named action so the conversation leads to an operational improvement.

Compare date ranges, currencies, attribution windows, event definitions and missing outcomes before assuming either report is wrong. Platforms may credit the same journey differently, and a sale may be recorded after the original enquiry period. Document the comparison you are making. The aim is an explainable reconciliation, not forcing unlike measures to display the same number.

Distinguish a new-customer outcome from additional work by someone already known to the business. Repeat revenue can be valuable, but it answers a different acquisition question. Confirm how the customer is matched and how refunds or corrections are recorded. Explain the distinction to the client before using a lifetime or repeat-revenue figure in campaign decisions.

Yes, the platform has workflows for both, but keep channel questions clear. A map-ranking improvement and a paid lead are different observations. Use a shared client record to provide context while retaining the appropriate measurement for each service. If several channels influenced the journey, explain the attribution model rather than awarding every channel full credit for the same sale.

Provide the account identifiers, access owners, tracking setup, conversion definitions, current priorities and unresolved data issues. Explain any exclusions or unusual configuration that another manager needs to understand. Preserve the reasoning behind recent changes. A successor should be able to interpret the report and continue the review routine without relying on undocumented knowledge from the previous account manager.

Choose one account with a cooperative client contact and a manageable enquiry volume. Verify access, test tracking, review a real lead and record an agreed outcome. Then inspect the report and client experience. This complete loop tests the parts that matter to delivery more effectively than connecting many accounts before anyone has confirmed how the outcomes will be maintained.