Assessing content and AI contributions

Why reports miss some origins

Traffic reports show recorded visits, not necessarily first discovery. The contact that created awareness may differ from the source of the final visit.

Consider someone who sees a company and phone number in an AI answer and calls without visiting its website. Web analytics has no session to attribute. This is a specific scenario, not a universal claim about AI behaviour or all tracking systems.

After discovering the brand elsewhere, a user may visit through another recorded source:

• Search. The user searches the company name.
• Direct visit. The user types your website address.

A visit source does not necessarily explain where someone first learned your name. First discovery differs from last click. Do not substitute one measure for another.

UTM tracking records clicks, not someone remembering an article and contacting later. Tagged links still leave journey gaps.

Last-click reporting can miss earlier content or AI exposure, but does not prove those contacts caused the enquiry. It records only the observed part of the journey.

Add customer and sales evidence

Reports record actions; customer conversations clarify earlier events. Ask, save the answer and compare with available data without replacing one with the other.

1. Ask how customers first discovered you. Ask 'Where did you first hear about us?' A phone call describes contact, not discovery. Record partial memories without inventing the rest.
2. Store responses in the CRM. Store source categories such as ads, search, AI, articles and recommendations, with customer wording and 'does not remember' available. Do not replace the customer's answer with staff assumptions.
3. Review initial results after a month. Define the denominator for source shares and report missing or unknown responses separately so the distribution is interpretable.

Compare customer responses with web, ad and call records for the same period. Differences may reflect first discovery versus a later visit, not necessarily an error.

Combined data still leaves gaps. It cannot alone explain a call made using a number in an AI answer, and customer memory may be incomplete. Combine evidence to test hypotheses, not claim perfect journey reconstruction.

Review search demand and direct visits

Brand searches and direct visits to internal pages add clues, but do not independently attribute AI or publication impact.

Trends in branded searches

1. Collect name ideas. Separate unambiguous brand searches from ambiguous terms and variants.
2. Review a year's history. Demand tools show interest in terms; website search reports show actual recorded arrivals. Use both without confusing them.
3. Plot publication dates over time. Compare changes using equal intervals, such as weeks or months.

Search growth after publication is a hypothesis to test, not proof. Advertising and recommendations may also influence awareness. A trend alone cannot identify first discovery or attribute it to AI.

Entry pages for direct visits

Review three months of direct visits grouped by landing page.

• Review homepage and internal pages separately.
• Review blog, services and pricing pages.
• Compare visits with publication dates.

A direct landing on an article narrows investigation but does not reveal origin. The visitor may have saved the URL or returned to a familiar page. Record page, period and observed change separately from assumptions about AI referrals.

Connect visibility observations with revenue

Compare discovery sources with financial outcomes. Acquisition performance includes total spending and revenue from new clients, not only the last click. Customer responses add context but do not replace financial records.

Bring data into a shared report:

• Monthly acquisition spending. Keep cost definitions consistent across reporting periods.
• Revenue from new customers. Separate new-client revenue from existing accounts and allow for different spending and payment dates.
• How customers first heard of you. Record publication or AI mentions alongside the share of customers with a known source.
• Indirect observations. Compare brand search, direct visits, customer answers and revenue.

Revenue-to-cost ratios are not profit. This does not establish profitability after all costs. Include production, selling and delivery expenses before drawing financial conclusions.

Partial journey data cannot precisely assign acquisition cost to a channel. Revenue rising with publications is not proof of causation; aligned trends still require testing.

Analytical takeaway: combine financial records, customer answers and indirect observations. Agreement suggests hypotheses to test; differences reveal missing evidence.