A marketing audit can be a practical freelance service for someone who understands the basics of campaign measurement and can explain findings clearly. Instead of managing a client’s advertising indefinitely, you review a defined set of materials and deliver a short list of useful next steps.
AI can help organize information and suggest questions, but it does not remove the need for judgment. A good beginner offer is small enough to check carefully, priced for the time involved and clear about what it cannot promise.
Define a small marketing audit you can actually deliver
Start with one business question, such as: “What should we investigate before spending more on this campaign?” Agree to review one advertising account or campaign, a specific date range and a limited selection of ads. Your deliverable might be a short report with three observations, three proposed tests and a discussion of measurement gaps.
A workspace such as OhlasAi can help with parts of that process: it accepts campaign exports and offers feedback on creative materials in relation to a stated goal and audience. Treat its output as a starting point, not a finished audit. You remain responsible for checking the underlying figures, identifying missing context and deciding whether a recommendation makes sense.
Write down what is outside the project too. A fixed-scope audit does not have to include rebuilding an account, creating new ads, installing tracking or managing future spend. If the client wants implementation, quote it separately after you understand the work.
Choose a type of campaign you know how to interpret. If you are unfamiliar with a platform’s reporting or the client’s business model, ask for time to learn, bring in qualified help or decline the job. The goal is a useful review, not a report filled with confident-sounding guesses.
Request the right client data and permission before uploading
Send a short request list before the project begins. Ask for the campaign goal, reporting period, currency, budget and an export with the metrics needed to assess that goal. Request a few representative ads or creative samples, plus relevant landing-page links. A brief explanation of the target audience and offer will make the review more useful than a spreadsheet alone.
Confirm that the person commissioning the work has authority to share the materials and agrees to any external software you plan to use. Explain what you intend to upload, why you need it and who will review the results. Do not ask for account passwords when an export or appropriately limited access will do.
Reduce the data before it leaves the client’s system. Remove customer names, email addresses, phone numbers, individual transaction details and internal notes unless they are genuinely necessary and their use has been agreed. Aggregate results where possible, and check creative files and screenshots for hidden or incidental personal information. Also consider whether confidential pricing or unreleased product details can be omitted.
Review the provider’s terms, security arrangements and data-handling practices before uploading; client permission alone does not answer every privacy question. The UK Information Commissioner’s guidance, How should we assess security and data minimisation in AI?, covers data minimisation and security considerations, including risks associated with outside providers. It is UK-specific guidance, so check the rules and contractual obligations relevant to your own work.
Agree on how files will be transferred, stored and deleted, and record what was actually shared. If the client cannot authorize an upload, you can still perform a manual review or work with an approved, appropriately anonymised sample.
Turn campaign observations into testable recommendations
Begin by checking what the numbers mean. A campaign with a low cost per reported conversion is not necessarily profitable if the conversion includes actions that rarely lead to sales. Ask which actions are counted, whether the conversion value is reliable and whether reporting periods and attribution settings are consistent. Google’s documentation, Understand your conversion tracking data, explains settings and distinctions that matter when reading Google Ads reports, including “Conversions” versus “All conversions.”
Next, compare performance against the client’s stated objective rather than a universal benchmark. For a lead-generation campaign, that could mean looking at spend, leads and the client’s information about lead quality. For a sales campaign, it may mean comparing reported revenue with advertising cost while acknowledging other costs and any tracking gaps. Do not assume that a reported conversion is a completed sale.
Separate an observation from its possible explanation. “Ad A received more clicks than Ad B during this period” may be supported by an export. “Ad A’s headline caused the difference” is a hypothesis unless the comparison was designed to isolate that factor. Check sample sizes, dates, targeting, budgets and other differences before drawing conclusions.
Turn a promising hypothesis into a bounded test: name the change, the audience or campaign it applies to, the metric to watch and the decision the client might make afterward. For example, test a clearer offer in one headline against the current version while keeping other elements as comparable as practical. Agree on the test budget and review period with the client; small datasets may not support a decisive answer.
AI-generated suggestions can help you brainstorm alternatives or spot questions you missed. Verify every quoted number against the source export, inspect the actual ad and landing page, and remove recommendations that depend on information you do not have. If tracking is unreliable, fixing or investigating measurement may be the most valuable first recommendation.
Price the audit around your time, software costs and scope
A fixed fee should cover more than the minutes spent reading a report. Estimate time for the initial conversation, data preparation, analysis, writing, a handover and the agreed revision. Add an allowance for software and other project costs. Then compare the result with what the client is buying and adjust the scope if the fee is not workable.
For a simple illustration, suppose you expect four hours of work and want to charge €30 per working hour. That gives a labour estimate of €120. If you also budget €2 for software usage and €8 for other project costs, the illustrative project total is €130. These are assumptions, not standard market rates; your own calculation should account for unpaid administration, taxes and the possibility that work takes longer than expected.
Prepaid software requires a separate cash-flow check. At https://www.ohlas.uk/pricing, OhlasAi lists a €10 minimum purchase and a rate of €1 per 10,000 credits. Purchases are one-time payments without automatic renewal. A €10 top-up therefore buys 100,000 credits, but the cash paid for that top-up is not necessarily the same as the value of credits consumed by one project.
For example, if usage tracking showed that a particular audit consumed 20,000 credits, those credits would represent €2 at the published rate. The initial cash outlay would still be €10 if you had to buy the minimum top-up for that job. The unused balance is not profit from the project; it is prepaid capacity you may be able to use later. This is only an arithmetic example, not a claim about how many credits an audit requires.
Measure actual usage on a pilot before estimating future software costs, and allow for variation in file size, questions and revisions. Put the fee, payment schedule, deliverables and revision limit in writing. Also state who pays for any advertising tests: audit software costs and the client’s campaign budget are separate expenses.
Deliver a concise report with clear limits and next steps
Clients should be able to understand the report without reading every calculation. Open with the business question, the materials reviewed and the reporting period. Then give a short summary of the most important finding and the action you suggest taking first.
For each recommendation, show the evidence, explain why it matters and identify an owner or decision. A useful format is “observation, implication, proposed test.” Include enough detail for the client to trace a figure back to its source, but avoid presenting a long list of minor changes as if they all have equal priority.
Be explicit about limitations. The export may exclude offline sales, the period may contain a promotion, or the client may not have supplied lead-quality data. Label assumptions and distinguish a data problem from a campaign problem. Do not promise a specific return, conversion rate or ranking: an audit identifies opportunities and uncertainties, not guaranteed results.
Finish with a manageable sequence, such as checking tracking first, testing one creative change next and reviewing the outcome after an agreed period. A short handover call can help the client challenge assumptions and decide what to implement. Send the final report through the agreed channel and handle source files according to your data-retention agreement.
Run one paid pilot before expanding the service
Offer the first project to a suitable client with a clearly defined campaign and accessible data. Explain that the offer is limited in scope, quote the fee in advance and use a written agreement. A pilot can be priced to reflect your experience, but avoid offering unlimited revisions or promising to keep working until the client gets a particular result.
Track your time by task and record the software usage attributable to the project. After delivery, compare the estimate with the actual effort. Did cleaning the export take longer than expected? Were you able to verify the findings? Did the client understand the recommendations and know which decisions were theirs to make?
Use the answers to improve your checklist, intake questions and fee before taking on more work. If clients repeatedly need tracking repairs or implementation, consider offering those as separate services only when you have the skills and capacity. A repeatable, honest audit is a stronger foundation for a freelance business than a broad promise to improve every campaign.
