Summary: A marketing audit reconciles advertising, the website and measurement with confirmed sales. Start with 15 business questions, then review 15 technical and content areas. The output is a list of evidenced problems, owners and repeat checks, without a promised CPA reduction.
The ad account reports purchases, while the shop records fewer paid orders. Or enquiries arrive but the team replies too late. Before increasing spend, establish where those two views diverge.
A six-month review can be useful. Recheck affected processes sooner after changes to the website, CRM, consent, market or campaigns. A calendar does not replace ongoing monitoring of critical events.
What should you record before the audit?
Write down five things: the business location and actual selling countries; team and buyer languages; product, order value, margin and sales cycle; the source of confirmed sales; advertising, privacy and sector restrictions. Otherwise identical metrics may describe different businesses.
15 business questions before checking the tools
1. Which outcome pays for the business?
Define a paid order, qualified enquiry, appointment or contract. A lead is not revenue. Shops also need returns, shipping and margin; service businesses need qualification and completed sales.
2. Are markets reported separately?
Separate at least country, language and product line. A strong market can hide a weak one. Compare suitable periods and the actual terms of each offer.
3. Does the buyer understand the offer?
The first screen should explain who you help, the task, service area and next action. Calling, booking, buying and enquiring are different paths. Copy and arguments should fit the buyer.
4. What role does each channel play?
Search can capture existing demand; Meta can introduce an offer and reach visitors again. SEO, email and CRM have further roles. Judge channels against their actual purpose rather than identical attribution expectations.
5. Does the conversion represent a real operation?
A page view, button click and received enquiry are different events. A purchase needs a unique shop record; a lead needs a traceable receipt. Showing a success screen alone is insufficient.
6. Are duplicate events identified?
Pixel and CAPI should not count the same operation twice. Check event names, aligned identifiers and the relationship to orders. Use an authorised test environment; do not trigger unplanned real purchases or leads.
7. Does consent control the data flows?
A banner is the interface. Check initial choice, refusal, reopening and storage across browser and server tags. Google’s consent documentation covers the relevant EEA use cases. Technical configuration and legal basis require separate assessment.
8. Are platform reports reconciled with the shop and CRM?
Windows, models, time zones and event definitions can differ. Compare payments, orders and qualified leads. A discrepancy deserves investigation but does not automatically prove broken tracking.
9. Can the business handle more demand?
Check response time, stock, appointment capacity and qualification. More advertising offers little help when enquiries remain unanswered or the advertised size is unavailable.
10. Does the campaign structure explain results?
Keep markets, products and demand types distinguishable. Too many groups split a small data volume; one total can hide material differences. Use a structure that supports decisions.
11. Do the creatives fit the market?
Language, currency, shipping, real evidence and presentation must fit the offer. Translating a headline alone does not prepare an ad for another market.
12. Does the landing page continue the ad’s promise?
Product, price, message and CTA should agree. Check readability, speed, telephone and form on mobile. Record obstructive elements and the actual step where the journey breaks.
13. Does the enquiry contain the necessary information?
Market, language, entry page and lawfully collected source information can help routing. Collect only necessary fields. Every additional required step needs a clear purpose.
14. Does content support the buying decision?
Answer questions about recognising the problem, comparing methods and choosing a provider. A service page cannot replace an explanation; an educational article should not contain unsupported service promises.
15. Is there a decision log and an owner?
Record the change, reason, date, review point and business metric. Assign responsibility for forms, tracking, CRM and ad accounts. Otherwise the latest intervention is easily mistaken for the cause of the result.
15 areas of the practical tool review
1. Speed and Core Web Vitals
Test the homepage, landing page, product, form and checkout. Good field thresholds are LCP ≤2.5 seconds, INP ≤200 milliseconds and CLS ≤0.1 at the 75th percentile. A PageSpeed lab test does not replace real-user data. web.dev explains the thresholds; good values do not guarantee rankings.
2. Mobile journey
Use a real smartphone: readable text, reachable buttons, appropriate keyboards, useful errors and unobstructed fields. Review your own device mix rather than an unsupported Austrian traffic percentage.
3. Google Business Profile
For an eligible local business, reconcile verification, category, address, hours, phone, URL, images and reviews. A three-month gap between posts does not prove lost rankings. Google describes relevance, distance and prominence. See GBP for Vienna SMEs.
4. Pixel, CAPI and analytics
Check event source, definition, timing, value, currency and deduplication against the shop. Missing CAPI does not imply a fixed percentage of lost sales. Meta Blueprint covers freshness, matching and deduplication. See CAPI and Pixel.
5. Event Match Quality
Use EMQ to diagnose matching parameters rather than profit. Check each field’s origin, format and permissible use. A score below seven does not prove wasted spend. Further guidance: CAPI setup.
6. Conversion and baseline
Compare the same countries, devices, sources and events over a suitable period. A low overall rate cannot distinguish poor traffic, an unsuitable offer and a technical obstacle on its own.
7. SEO and Search Console
Review indexability, canonical, sitemap, hreflang, impressions, clicks and search intent. Find pages where demand lacks an appropriate answer. Total traffic alone does not explain revenue contribution.
8. Content accuracy
Review prices, dates, screenshots, sources, links and limitations. Older content is not automatically weak. Google cautions against date changes without substantive updates.
9. Competitors
Compare three relevant queries and providers: offer, evidence, entry price, response path and buying obstacles. A short review generates hypotheses; it does not establish their revenue or validate their promises.
10. Social media
Review reach, relevant visitors, enquiries and retention alongside reactions. Engagement below one percent does not prove failed delivery. Define the numerator, denominator and purpose first.
11. Email and CRM
Check delivery, permission, segments, triggers, response time, clicks and orders. Open rates have limitations, including Apple Mail Privacy Protection. Old automation can keep sending the wrong offer.
12. CPA, ROAS and margin
Use confirmed orders, discounts, returns, shipping and fees. For leads, add qualified opportunities and sales. CPA €20–35 or ROAS 3–5 without product economics is not a universal DACH standard.
13. Cookie banner and privacy
Review choice, withdrawal and actual data flows. CAPI and Consent Mode do not replace a valid basis. EDPB consent guidelines are a starting point, not legal approval for your shop. See cookie banner review.
14. Directories and NAP
Name, address and phone should agree across the website, GBP and relevant WKO or Herold listings. Correct real discrepancies; assess directory quality before adding entries. See Austrian business directories.
15. AI visibility
Repeat a recorded set of commercial questions. Capture date, language, named brands, sources and potential visits. A missing mention does not prove defective content. Google AI uses ordinary SEO foundations, without a special AI file or schema requirement. See the GEO guide.
Scope and cost of a professional audit
| Format | What to agree before starting |
|---|---|
| Self-review | This checklist and existing reports; no external service |
| Focused review | One system or evidenced problem; agree access, scenarios and output |
| Full audit | Advertising, website, consent, CRM and orders; define depth, price and deliverables |
| Initial discussion | Establish the task; does not replace technical testing or promise a full free audit |
An open-site review can identify visible problems. Internal processes require agreed access. Implementation and repeat checks need a separate agreement unless included in the engagement. Context: agency pricing.
What can you prepare this week?
- Record one confirmed business conversion and its source.
- Review the mobile journey from ad to form or checkout without an unplanned real transaction.
- Reconcile tracking with shop/CRM in an authorised test environment.
- Check GBP, telephone, URL and NAP in the relevant listings.
- Choose three evidenced problems with an owner and review date.
A short Events Manager check, NAP comparison and PageSpeed test can provide an initial view. The earlier 5/10/15-minute slots are planning aids, not a complete diagnosis or repair commitment.
How do findings become decisions?
Critical: unusable data, missing enquiries, duplicate purchases or consent choices not honoured. Commercial: offer, market, landing page or lead handling limits sales. Optimisation: improve structure, creatives and content once the earlier obstacles are understood.
For each finding, record evidence, possible impact, owner and verification method. A before/after comparison can contain changes in season, price or demand; it does not automatically isolate the audit’s effect.
Frequently Asked Questions
How often should I run a marketing audit?
A six-month rhythm can be a useful planning choice. Recheck affected processes after material changes and monitor critical events continuously. There is no established mandatory interval for every SME.
Can I perform the audit myself?
The checklist supports an initial self-review. Internal data flows, deduplication and difficult discrepancies may need specialist help. Agree scope, access and price before starting.
Which problems should I look for first?
Start with confirmed sales and the enquiry journey: missing receipts, duplicate events, incorrect definitions and outdated business information. Without a documented sample, no most-common failure rate can be stated.
How do I measure the economic effect?
Record the baseline, specific obstacle, change and evaluation period. Compare confirmed sales, margin and processing after a suitable cycle. An audit alone carries no guaranteed ROI.
Why include AI visibility?
It can be another discovery path. Review repeatable mentions, accurate descriptions, sources and visits separately. A single answer or technical eligibility does not establish lasting visibility.
Are general benchmarks useful?
Only with a traceable sample and comparable category, margin, country and event definition. Your own confirmed data is the first comparison; another business’s average does not explain your bottleneck.
Is the ad account enough?
It can reveal settings and reporting. Business assessment also needs the website, consent, orders or CRM and the lead-handling process.
When is an independent review useful?
For unexplained discrepancies, a new market, a substantial spend increase or evaluating a provider. Limit access and depth to the actual task.
Describe the selling market, buyers, confirmed sales and observed problem. We can establish the data and review scope needed. Discuss the task.
