Frequently asked questions

The analysis combines four sources: campaign configuration (account structure, bidding strategies, search terms, exclusions), the measurement layer (conversion actions in Google Ads, GA4 events, tags and triggers in GTM), organic queries from Search Console, and how well the landing page matches the intent of the query. The audit checks whether an ad click can be traced to an enquiry and whether the same contact is counted more than once. Meta Ads and LinkedIn Ads are separate scope modules. The scope and deadline are agreed before work starts.

The workload grows with the number of campaigns, markets, landing pages and data sources. Auditing a single Google Ads account requires different work from a multi-channel audit covering GA4, GTM, Search Console and the website. A description of the accounts, channels and the problem is enough for an initial quote; if the scope includes the website, its address is needed. The price does not include implementing changes.

Each finding states where it occurs (account, campaign, tag, event, URL), the data supporting the conclusion, the recommended change and how to verify it after implementation. Example: when the conversion is a button click, the report shows which event is sent, when it fires and which event should confirm that the form was accepted. Hypotheses and missing data are labelled separately. The audit ends with delivery of the PDF report; implementation, campaign management and subsequent technical assistance are not part of the service.

For an audit, the following are needed: the advertising channels, the purpose of the analysis and, if the website is in scope, its address. For a translation: the type of material, language pair, volume and deadline. First contact is made within 24 hours on business days. Once the terms are accepted, the next step is to grant the agreed access or hand over the files.

Permissions granted to the designated account are used to read data and configuration; the audit does not include changes to the accounts. For the Google scope, the list usually includes Google Ads, GA4, GTM and Search Console, and for the Meta Ads module the advertising account, Pixel and Conversions API. Examples of leads and information on which of them the sales team considers valuable also help. Sending an invitation alone does not confirm access to the data.

The report describes what to change, why and how to assess the effect. It is not a platform user guide or a step-by-step configuration. The description of the problem, the evidence and the acceptance criterion allow the provider to plan the implementation and check whether the change worked.

An example of a check: comparing conversions in Google Ads, events in GA4 and tag firings in GTM to establish what is actually counted as an enquiry. Clicking the form button and an accepted enquiry are different events, and the system gathers the data needed for this check. The analysis also covers how well the page matches the ad message. In the report, confirmed problems, hypotheses and missing data are labelled separately.

A B2B audit examines two things: whether enquiries come from companies matching the target profile, and whether the sales outcome (qualified lead, opportunity, contract) returns to the advertising systems. Google Ads accepts such data through offline conversion import based on GCLID and, in enhanced conversions for leads, also through hashed contact data. Without this, bidding strategies optimise for the conversion they can see, that is, the form itself, rather than for a lead that is valuable to sales.

The chain of events is checked from the end: whether enquiries reach the inbox or CRM, whether the form sends the data, whether the confirming event (e.g. generate_lead in GA4) fires after the enquiry is accepted rather than on the button click, and whether it reaches Google Ads as a conversion. Only then are the match between ad and page and the intent of the search terms assessed. Measurement errors also include a refusal of consent to measurement, after which some conversions may not be recorded, and the delay between the click and the conversion. The number of enquiries in the CRM is compared with conversions in GA4 and Google Ads, and the discrepancy indicates the layer where the error occurs.

The starting point is a definition of a good enquiry: the right company, a need that matches the offer and a realistic possibility of purchase. Sales data (CRM, export, an organised list) makes it possible to compare sources by accepted leads and rejection reasons. The delay has to be taken into account: fresh enquiries are not compared with completed conversations, and the conversion window in Google Ads (30 days by default for Search and Display, up to 90 days for clicks) limits which conversions are recorded at all. Without sales data, the quality assessment is incomplete.

In Google Ads, conversion actions are divided into primary and secondary: primary ones go into the “Conversions” column and are used in bidding, while secondary ones are for observation only and are reported in “All conversions”. Clicking a phone number does not confirm a call, and opening an email client does not mean a message was sent, so such events should be secondary. In GA4, an enquiry is recorded with the generate_lead event; when a value is assigned, the event requires the currency and value parameters.

Search works well when there is measurable demand for queries describing the solution. LinkedIn allows targeting by company, industry, job function, job title and seniority, which matters for a small, well-defined group of buyers, but criteria that are too narrow limit reach and the speed of data collection. Google can also build interest through YouTube and Demand Gen. The choice takes into account the budget needed to collect data, enquiry quality, sales opportunities and the time it takes for contacts to mature.

Correct measurement is a prerequisite for a meaningful assessment. Events from the browser (Pixel) and from the server (Conversions API) are deduplicated when event_name and event_id match and the second event arrives within 48 hours of the first; without this, the same enquiry is counted twice. Campaign results are assessed together with data on engagement, brand searches and contact quality, because attributing the result solely to the last click overlooks earlier interactions with the ad.

This is limited by the number of events needed for an assessment and by the delay between the click and the result. The conversion window in Google Ads is 30 days by default for clicks on Search and Display (up to 90 days); conversions after it expires are not recorded, and the delay analysis is limited to the selected window. For this reason early signals (search term match, impression share, event accuracy) are assessed separately from business results, and when the sales cycle is longer than the conversion window, offline conversion import is needed.

The audit checks whether the ad message matches the page, the content and evidence of expertise, the form (number of fields, validation, confirmation message), how the page works on mobile, and performance measured with Core Web Vitals metrics: LCP up to 2.5 s, INP up to 200 ms and CLS up to 0.1 at the 75th percentile of visits. A visible problem can be described directly, but its effect on campaign results requires data or a test.

A customer may see an ad, read an article and return weeks later by searching for the company name, and the last click does not describe that whole journey. Google Ads currently offers two models: last click and data-driven attribution (the default for most conversion actions); rule-based models (first click, linear, time decay, position-based) have been discontinued. GA4 provides data-driven attribution and two last click variants (paid and organic channels, and Google paid channels only), while the others were removed in November 2023. No single report will show all the channels and people involved in the decision, so the findings are compared with sales data.

Events reach a container on the site’s own domain (hosted on, among others, Cloud Run, App Engine or a managed provider such as Stape), where server tags can filter and enrich them and pass them on, e.g. to GA4, Google Ads or Meta Conversions API. Part of the measurement still takes place in the browser. Implementation increases control over data leaving the website, but it does not replace user consent, it incurs hosting and maintenance costs, and it does not guarantee complete data or immunity to browser blocking. Whether implementation is justified depends on whether control over data sent to vendors or better measurement quality is needed, and whether these justify the cost. The audit assesses whether sGTM is justified but does not include implementation.

In an interface, what matters is consistent terminology, preserving variables and tags, and the length of the text, which must fit within the UI element. Context, e.g. a screenshot or a feature description, reduces errors arising from short strings taken out of context. A quote requires the file format, volume, language pair and information about the product.

The specific pair and scope are confirmed on the basis of a file or a text sample. It also helps to know who the material is for and where it will be used, e.g. on a website, in documentation or in an interface, because register and terminology depend on this.