Before Assessing Reach: How to Review a Publication for Possible Signs of Automated Traffic
A practical guide to interpreting audience data shared by a publication, checking indicators of automation, and deciding what information to request before commissioning a placement.
Declared volume alone does not prove audience quality
A traffic figure can be a useful starting point, but it does not explain who visited, where they came from, which channel they used, or what they did next. Nor can it show on its own whether the traffic is relevant to the proposed collaboration. Before comparing publications, separate three questions: how much traffic is reported, how it was measured, and how it relates to the campaign’s objective. Figures may refer to different periods, tools, and concepts. Users, sessions, page views, and impressions are not interchangeable. Also, a metric from an analytics tool cannot be directly compared with an estimate from an external tool: coverage, methodology, and data processing may differ. A high number is therefore no guarantee of a useful audience, and a lower figure does not necessarily mean the publication has no value.
- Clarify which indicator is being shared and the period it covers.
- Check whether the figure covers the entire site, a section, or a specific page.
- Relate the audience to the topic, geographic market, and purpose of the collaboration.
- Treat SEO metrics and audience metrics as different signals, not as substitutes for one another.
What evidence to request—and its limitations
Ask for information that helps you understand the data, not unrestricted account access or readers’ personal information. An aggregated analytics summary may be enough for an initial review if it includes the period, the property or section analyzed, the main traffic sources, and an explanation of the indicators used. If the proposal depends on a specific article or section, data for that page or comparable content is more informative than the site-wide total. Evidence also has limitations. A screenshot may be outdated, cropped, or refer to a different scope; an aggregated report may conceal internal variations. A third-party estimate may help with comparisons, but it does not confirm traffic observed by the publication. Interpret each item according to its origin and scope, and avoid requesting identifiable data or login credentials.
- Request the exact period and time zone, if relevant to interpreting spikes.
- Ask whether the figures cover the entire domain, a category, or a URL.
- Request an aggregated breakdown of channels and main countries, where relevant.
- Clarify whether the figures are data observed by the publication or an external estimate.
- For a previous campaign, ask whether they can share aggregated, anonymized results without personal data.
Cross-check patterns, sources, and consistency across data
No single signal diagnoses automated traffic. The review involves looking for context and consistency across multiple pieces of evidence. For example, it may be useful to examine changes over time, the distribution by channel, geographic origin, and the relationship between visits and landing pages. A sudden change warrants a question, but it may have a legitimate explanation: a campaign, a mention, seasonality, a news story, or a technical change. It also matters what each source means in the report. Direct traffic does not always identify a visit with no known source; attribution can lose information because of privacy settings, apps, redirects, or tagging. Similarly, engagement indicators depend on how they are configured and should not become an automatic verdict. Compare equivalent periods and look for explanations before interpreting a discrepancy as an anomaly.
- Check for concentrated spikes and ask what editorial or promotional activity coincided with them.
- Check whether the reported sources are consistent with the channels the publication uses to distribute its content.
- Compare the geography with the target audience, bearing in mind that a difference may have legitimate causes.
- Look at several metrics and periods; do not treat an engagement rate or average duration as conclusive proof.
- If the data comes from different tools, note that their definitions and methods may not match.
A signal is not a verified conclusion
Some patterns may justify follow-up questions: abrupt increases with no known context, activity highly concentrated at repeated times, unexpected geographic distribution, or a weak relationship between visits and content consumption. A traffic source that does not fit the stated activity may also draw attention. These are indicators to investigate, not proof that automation is taking place. Automated traffic can have legitimate uses, such as crawlers and monitoring tools, and not all non-human activity represents fraud. Conversely, artificial activity can be mixed with real visits. Without access to suitable data and technical context, a third party generally cannot conclusively verify the origin of every visit. Describe what you observe precisely: “the report shows an unexplained spike” is more rigorous than “the traffic is fake.”
- Frame hypotheses and questions, not accusations based on a single anomaly.
- Distinguish traffic from known crawlers from possible metric manipulation.
- Ask for context about campaigns, technical changes, or events that could explain the pattern.
- If the risk is material, request a methodological clarification or an independent review suited to the case.
Practical questions that do not require confidential data
You can have a direct and respectful conversation with the publication. Instead of requesting access to an analytics account, explain what you need to validate in order to decide and offer to accept aggregated data. This limits exposure of sensitive information and helps both parties discuss the same indicators. For a sponsored collaboration, also ask how the content will be distributed and how the commercial relationship will be identified. Transparency about the collaboration is independent of traffic quality. For paid links, Google states that advertising or paid links should be qualified with rel="sponsored"; rel="nofollow" may also be valid. The appropriate attribute depends on the case and the editorial implementation.
- What period and part of the site do the shared figures cover?
- Which tool or method was used, and how is each metric defined?
- Can you share an aggregated month-by-month trend or a sample of comparable pages?
- Which channels and countries account for most visits? Were there campaigns or relevant changes during that period?
- How will the content be promoted, and how will it be identified as a sponsored collaboration?
- Which attributes will be applied to commercial links, as appropriate?
Document uncertainty and decide whether you need more information
A sound decision does not require absolute certainty: it requires knowing what has been checked, what remains unknown, and how important that uncertainty is to the campaign. Record the review date, the source of each data point, its scope, explanations received, and outstanding questions. This will help you compare opportunities using consistent criteria and later revisit why a decision was made. If the potential issue significantly affects the objective or budget, ask for clarification before committing. If the collaboration still makes sense with the information available, you can document the limitations and reduce exposure by aligning the scope or investment with your level of confidence. If there is not enough information and you cannot accept the risk, postpone or decline the opportunity without presenting uncertainty as an accusation.
- Record the data point, period, scope, source, and limitation of each piece of evidence.
- Separate observed facts, explanations provided, and your own hypotheses.
- Define in advance what additional information would change your decision.
- Compare publications using the same method, and do not mix figures of different kinds.
- Use analytics and comparison tools to support the process, not to certify traffic quality.
Conclusions to avoid
The aim of the review is not to label a publication based on an estimate, but to better understand the available evidence and make a proportionate decision. Third-party tools may differ; a screenshot or aggregate figure may not answer the specific question; and a striking variation may have a legitimate cause. None of these elements, on its own, proves or rules out automated traffic. It is also unwise to infer SEO results from traffic volume. Visibility, audience, and the value of a collaboration depend on different factors. Evaluate each opportunity with context, transparency, and documentation, without turning a metric into a promise of rankings, authority, or future traffic.
- Do not equate an external estimate with a direct measurement by the publication.
- Do not treat a spike, unexpected geography, or low engagement as sufficient proof.
- Do not interpret the absence of visible signals as a guarantee of authentic traffic.
- Do not attribute SEO results to a collaboration that cannot guarantee them.
Frequently asked questions
What is the first data point I should request to evaluate a publication’s traffic?+
Start with an aggregated figure that has a clearly stated period, scope, and definition. Then, if relevant to your decision, request a basic breakdown by channel or geography. Without that context, an isolated number is difficult to interpret.
Does a low engagement rate prove that traffic is automated?+
No. It may be a reason to ask for context, but it depends on measurement settings, content type, device, source, and other factors. On its own, it cannot confirm automation.
Can I verify traffic without access to the publication’s analytics?+
You can make a reasonable assessment using aggregated reports, comparable periods, explanations of sources, and external estimates as a reference. Without access to internal data and technical context, you cannot conclusively verify every visit.
What should I do if I find an unexplained spike?+
Ask the publication to confirm the period and data source, and whether there were campaigns, mentions, technical changes, or seasonal factors. Record the response and decide whether the uncertainty affects the collaboration enough to request more information or postpone it.
Are visits estimated by an SEO tool equivalent to actual visits?+
Not necessarily. They are estimates based on methodologies that may differ from the publication’s analytics data. Use them to inform comparisons, not as exact proof of observed traffic.
Sources and references
- Google Search Essentials — Google Search Central
- Spam policies for Google web search — Google Search Central
- Qualify outbound links — Google Search Central