SEO after the ranking hype: Google makes ranking tracking more difficult – Why visibility alone doesn't generate business

SEO without measurement certainty: How companies can still make better decisions

AI search is changing the rules of the game: Why companies need to rethink their SEO metrics

Search engine optimization is becoming more challenging – not because rankings and keywords have become meaningless, but because they are increasingly unreliable indicators of what truly matters to businesses. Google is making it more difficult to automatically analyze its search results, while AI responses are shifting attention and the attribution of marketing effectiveness remains incomplete. This makes it harder to draw reliable conclusions about reach, demand, and business success from individual ranking positions.

The answer isn't to abolish key performance indicators (KPIs) or abandon data altogether. Instead, companies need to more precisely distinguish what they are actually measuring, how reliable these metrics are, and what business decisions they can justify. Visibility remains important – but it only gains its value when it reaches the right people, builds trust, and contributes to relevant actions. The following text shows how SEO can be realigned in an era of less certain measurement: away from the mere obsession with rankings and towards a comprehensible connection between discoverability and business.

Those who only count positions lose sight of the business

Search engine optimization doesn't become easier simply because less reliable data is available. Rather, it needs to be managed more clearly: with economic objectives, traceable cause-and-effect relationships, and a conscious distinction between actual market changes and measurement errors. Therefore, Google's decision to make the automated collection of its search results more difficult doesn't just affect individual SEO tools, but the very foundation of many previous success stories.

The crucial change, however, is not that keywords, rankings, or technical optimization suddenly become meaningless. Rather, the assumption that a highly detailed observation of individual positions can reliably predict reach, demand, and business success is becoming less valid. A more sensible perspective views discoverability as part of a larger value chain: people need to discover an offer, recognize its relevance, develop trust, and ultimately make a financially relevant decision.

A sound analysis must distinguish between three developments: the increasing difficulty of accessing search result data, the altered distribution of attention due to AI responses, and the limited attribution of marketing impact. These problems overlap, but are neither technically identical nor solvable with the same metric. Their interplay creates a new management challenge: companies must make better decisions with less certainty in measurement.

The key figure is not the business

A traditional visibility index condenses rankings into a manageable metric. Sistrix, on the other hand, weights positions based on criteria such as search volume and expected click-through rate, and then aggregates the results. This provides an indicator of a website's organic presence, but it doesn't directly measure actual visits or calculate its economic return.

This distinction is economically crucial. A better position can generate additional attention, but this only translates into a financial advantage if the people reached are a good fit for the offering and their needs are effectively met. A highly specialized service with a few qualified contacts can be more valuable than a widely used advice column whose audience primarily seeks free information. The key is not maximum visibility, but visibility within a relevant decision-making context.

Even without technical measurement problems, it would therefore be risky to declare a proxy metric the actual company goal. As soon as teams are primarily evaluated based on rankings achieved, an incentive arises to produce easily measurable improvements. More difficult-to-measure contributions, such as better purchasing decisions, less uncertainty, or increased trust, are more likely to be neglected. The appropriate response is not to abolish key performance indicators (KPIs), but to define their function more narrowly.

Google's defenses are changing the rules of the game

Since September 2025, the previous method of querying up to 100 organic results using the parameter num=100 is no longer possible. This necessitates more queries to achieve a comparable capture of deeper positions. In 2026, additional difficulties arose due to redirects, blocked access, and conspicuously different results pages. Such changes increase the technical effort required for data collection and can affect its costs, updates, and completeness.

An important distinction must be made here: An SEO crawler that examines a company's website for broken links performs a different task than a service that automatically retrieves Google results. Similarly, the Googlebot, which crawls external websites for the search engine, must be distinguished from the programs that read Google's search results interface. Therefore, the current restrictions on rank tracking do not automatically mean that every technical website analysis is affected.

A defense against extensive AI-related data collection is a plausible explanation and is supported by the providers involved; however, the specific motivation behind the conspicuous variations in results is not independently proven. Google explicitly prohibits unauthorized automated search queries, including ranking checks, and points to the strain on its resources. This documents a conflict of interest, but not the precise intention behind each individual technical change.

Redirects obscure the actual hit

A clear example of a problem is Google redirects via addresses containing the `goto` attribute. The result address initially points to Google; the actual destination only becomes visible after the redirect is resolved. A data interface can therefore recognize an organic ranking but still return a Google address instead of the actual destination address. If this data set is processed without verification, an existing ranking can be incorrectly attributed or overlooked.

On August 27, 2026, DataForSEO reported that 99.99 percent of the results pages delivered via its interface no longer contained unresolved goto addresses for organic search results. The reference point was results pages, not individual links. However, such technical success was no guarantee against subsequent changes: In an external sample taken on October 30, 508 results pages with 4,344 organic search results were analyzed; in 2,552 cases, or 58.7 percent, a goto address remained instead of the complete target address.

This measurement describes a specific state for a particular selection of queries and a specific point in time. It is not a representative statement about all Google searches or all ranking tools. This very limitation makes the finding analytically valuable: a general provider evaluation cannot be derived from it, but rather the necessity of examining raw data, resolution rates, and measurement times.

Credible results can be false

Even more problematic are results pages that appear complete and plausible but deviate significantly from the usual results for the same search query. Several providers have documented such anomalies in automated queries. The fundamental risk is that no obvious error is apparent, and a ranking system may therefore register the response as a regular observation.

An analysis considered 56.1 million assessable search results pages from 53 countries between August 20 and October 21, 2026. For the first search results page, the proportion of observations deemed conspicuous by the researchers' own definition increased from 0.17 percent in the initial period to 8.7 percent since September 13. In a separate desktop subset, where both the first and second pages could be assessed, the proportion of conspicuous first or second pages was 22.6 percent. These values ​​should not be considered equivalent due to the different underlying populations.

Even the detection has its limits. In simplified terms, a page is considered suspicious if less than half of the previously regularly represented websites are found in the collected search results. Genuine algorithmic changes can also produce unusual combinations. Therefore, these findings represent a serious warning signal from practical experience, but are neither a universally applicable error rate nor independent proof of deliberate manipulation by Google.

Reliability comes at a price

Sistrix continues to describe its top 10 data and visibility index as reliable, but also points to additional checks and, in some cases, slower updates. If new results fail quality control, the data is collected again; until then, the last checked version remains visible. Furthermore, on-demand updates have been disabled since September 2026 to ensure these additional checks are not circumvented.

This creates a conflict between timeliness and reliability. A provider that checks data more carefully can deliver more robust, but older, figures. A service that updates more quickly can show short-term changes earlier, but requires correspondingly powerful error detection. Under such conditions, a smoother curve does not necessarily mean a calmer market. It could also indicate that fewer new observations have been published.

Therefore, to evaluate a ranking dataset, the actual data collection date should be considered in addition to the position. Equally relevant are the proportion of updated keywords, unusual measurements, and documented methodological changes. Companies should not judge their service provider solely on whether the interface continues to display daily figures. More important is whether it is clear which values ​​are new, verified, unchanged, or currently unreliable.

Many measurement points do not replace quality

The recommendation to assess visibility across multiple metrics rather than individual rankings is generally sound. This reduces the impact of short-term fluctuations while making longer-term trends more apparent. However, the observations must be sufficiently comparable. Above all, their errors must not systematically lead in the same direction.

Random fluctuations can partially balance out with appropriate aggregation. Systematic biases, however, do not disappear with larger datasets. For example, if certain search query types or later results pages are particularly frequently affected, a topic area may appear permanently weaker, even though its actual presence remains stable. Similarly, two tools cannot automatically provide independent confirmation if they both use the same upstream data.

A robust analysis therefore requires a comparison group that remains as consistent as possible and documented data coverage. New keywords should not be inadvertently introduced into existing time series; missing measurements should not automatically be treated as lost rankings. Simplification here means interpreting fewer unreliable details and, at the same time, thinking more carefully about the reliability of the remaining information. This is not a retreat from data-driven work, but rather an improvement of its discipline.

AI changes the value of a ranking

Even accurately measured rankings no longer guarantee the same level of attention as before. Search results can now fulfill an information need directly on the results page, without requiring a subsequent visit to an external website. Studies on user behavior show a correlation between displayed AI summaries and fewer clicks on traditional search results.

A study analyzing browser data from 900 adults in the US and 68,879 different Google searches from March 2025 found significant differences: On search pages with AI summaries, 8 percent of visits resulted in a click on a traditional search result, compared to 15 percent without such summaries. A link within the summary was clicked on 1 percent of visits to pages with AI summaries. However, the results pages were collected retrospectively, and search queries with and without AI summaries differed in their composition; this does not constitute controlled causation.

Another analysis compared aggregated desktop data from Search Console for December 2023 and December 2025, based on 300,000 keywords. Compared to a trend derived from the control group, the top-ranking page for keywords with AI-generated results showed a click-through rate approximately 58 percent lower. This value is dependent on the selection, time period, and model assumptions. It cannot be directly applied to Germany or to all industries.

Fewer clicks can be more valuable

Declining click-through rates do not automatically imply a proportional decrease in business value. If fundamental questions are answered before a website visit, subsequent visitors may have a more specific interest. Therefore, for economic evaluation, the quality of inquiries, sales conversions, revenue per visit, and the actual remaining contribution margin are more important than the sheer number of visits.

For US online retailers, visits demonstrably driven by AI-generated offers in September 2026 showed an 83 percent higher revenue per visit and a 90 percent higher conversion rate compared to non-AI visits. The underlying analysis was based on a retail dataset containing more than one trillion visits; this total size should not be equated with the number of AI-generated visits included. Therefore, no general guarantee of profitability for German industrial companies or specialist publishers can be derived from this.

The crucial factor is selection. AI users may already be more ready to buy, searching for different products or shopping in different price ranges. Therefore, a higher conversion rate doesn't prove that the AI ​​offering generated this additional purchase intent. Similarly, higher revenue per visit doesn't necessarily translate into a correspondingly high share of total revenue. A small channel can achieve excellent individual results and still contribute only a limited amount to the overall business.

The added value shifts to the answer

Economically, a portion of the value creation shifts to the surface level, where information is selected and compiled. This can be advantageous for providers if their solutions are presented effectively and qualified prospects are connected. Problems arise, however, when expensively compiled information is used without generating sufficient attention, contacts, or revenue.

A mention in an AI response therefore has no fixed economic value. Presenting a company as a suitable solution provider can support a selection decision. However, simply adopting an easily replaceable definition may result in little benefit for the original provider. The same applies to source references: a link creates a contact option, but neither automatically leads to a visit nor a business relationship.

This raises a strategic question for content production. Companies should distinguish between information that should be widely available and content that provides a distinct reason for direct contact. Original data, verifiable practical experience, technical calculations, and in-depth tools serve a different function than interchangeable overview texts. Additional content only creates economic value if it solves a relevant problem better than existing alternatives.

Your own data becomes the foundation

Effective management requires information that is as closely aligned with the company's core business as possible. Google Search Console provides insights into search performance as tracked by Google; web analytics describes measurable visits and actions. Customer management can supplement this view with qualified leads, sales opportunities, and conversions. Together, these elements create a more robust basis for decision-making than relying solely on external ranking curves.

The Search Console is not a perfect reflection of reality. Google itself documents changes to measurement methods and occasional errors in data recording. One example is an error that affected impressions from May 13, 2025, to April 27, 2026, while clicks, according to Google, were not affected. Therefore, switching to your own data or data provided directly by the platform does not replace the need to verify its definitions and quality.

For longer-term analyses, building up your own data history early on is beneficial. Automated exports enable additional links and evaluations; however, anonymized search queries don't suddenly become available in plain text. Companies must consider data costs, access rights, and failure monitoring. Owning your own data means more control over the analysis, but not complete independence from the platform's provisioning rules.

Xpert.Digital is a data-driven B2B industry hub led by Konrad Wolfenstein . The company acts as an external, quasi-in-house solution for industrial partners, closing operational gaps in marketing, content, and sales – without requiring additional resources on the client side.

More information here:

• The quasi-in-house solution: How Xpert.Digital closes operational gaps in B2B marketing and sales – Smart Content-Driven Business

New reports don't fill every gap

The notion that AI visibility is not observable at all in Google's own tools is now too simplistic. Since August 31, 2026, the dedicated reports on generative AI performance have been rolled out worldwide. They show impressions for supported features such as AI Overviews and AI Mode, as well as analyses by page, country, device, and time period.

This additional perspective improves monitoring but doesn't eliminate the entire impact gap. The specialized report focuses on impressions and doesn't provide complete, separate evidence of the resulting clicks and business outcomes. Furthermore, the relevant viewability data is already included in the general performance report and shouldn't simply be added on.

In practical terms, this means that the performance of a URL in generative search functions should be considered in conjunction with its overall search performance and subsequent website results. An increasing number of AI impressions can be a relevant early indicator, but it is not proof of increased demand. A new name for visibility does not replace a new economic rationale.

Server logs show no purchase intent

Server and CDN protocols offer another technical perspective. Their strength lies in examining requests, response states, and access problems, not in directly explaining human decisions. A corresponding name in the user agent is insufficient to identify a Google crawler, as this identifier can be spoofed. Google recommends verifying the origin via its IP ranges or appropriate DNS methods.

An automated retrieval of data does not provide complete proof of subsequent AI use. The retrieval alone does not indicate whether content appeared in a response, how it was presented, or whether it resulted in customer contact. Similarly, an increasing number of retrievals does not automatically guarantee success. For the company, what matters is not the amount of time a program spends on its pages, but the resulting benefit.

The methodological boundary is crucial: Technical observations should answer technical questions. Additional data is required to make statements about demand and willingness to buy. Even the desire for better measurability does not justify unlimited data collection. A sound architecture defines what information is needed for what purpose, who has access to it, and how long it is stored.

Channel assignment remains an approximation

A potential customer might first encounter a provider through an AI response, later search for their name, and finally contact them directly. Web analytics might then show the most recent search as the source of the visit. The earlier influence remains unmeasured, even though it may have been relevant to the decision. The same fundamental problem exists with events, trade publications, and personal recommendations.

The technical distinction is nevertheless improving. On May 13, 2026, Google Analytics introduced a dedicated channel for AI assistants, which separately attributes visits based on a recognizable AI origin. According to the current channel definition, Google's AI Overviews and AI Mode, however, still belong to organic search and not to this separate assistant channel. Therefore, anyone who derives all AI influence solely from its volume is underestimating a portion of the relevant usage, even based on this definition alone.

A more robust practice combines clear campaign labeling, documented channel rules, and supplementary customer surveys. Even a voluntary question about the initial touchpoint can be helpful, but its effectiveness depends on recall and willingness to respond. Distinguishing between a lack of attribution and a lack of impact is particularly important. Conversely, not every unclear visit should be retrospectively attributed to a supposedly successful SEO or AI strategy.

Themed spaces instead of rigid keyword lists

The URL remains a useful operational unit because its function, content, and technical quality can be specifically addressed. However, instead of aligning each page solely with a single search term, it should serve a clearly defined problem and its related information needs. A basic keyword can still serve as a linguistic anchor. The crucial factor is whether the page explores the relevant topic clearly and in sufficient depth.

For an industrial solution, this might include technical requirements, economic limitations, integration effort, and risks. Such questions are interconnected, even though they are formulated very differently in search queries. Google also describes the ability for its AI functions to break down an initial question into several related sub-queries. Therefore, a single user question does not necessarily correspond to a single classic keyword query.

However, a thematic approach shouldn't lead to an endless collection of information. An introduction for beginners, technical documentation, and a concrete product decision may require different page types. A clear main page with logically linked sections is advisable, provided the needs justify it. Individual rankings remain diagnostically useful but lose their role as the sole objective.

A URL becomes part of the decision path

This can be illustrated using the example of a provider of automated warehouse solutions. A central page could be based on the fundamental concept of automated order picking. However, its task would not be to use this term as often as possible. It would have to explain which order structures the technology is suitable for, what throughput requirements exist, and what interfaces to existing software are necessary.

The website becomes economically valuable when it reduces the uncertainty of a real investment decision. Transparent calculation methods, clearly stated assumptions, and documented limits of use can contribute to this. A serious analysis of amortization must differentiate between investment costs, operating costs, utilization, and conversion risks. A blanket promise of return would be less helpful than a transparent explanation of when a solution is not worthwhile.

Success monitoring would combine various metrics. Search data shows the development of relevant contacts, web analytics the measurable use of advanced search options, and customer management the quality of follow-up inquiries. An industry article or a product demonstration can also contribute to the same decision-making process. The goal is to optimize the connection between information needs, trust building, and qualified contact. A position in the search results remains one possible entry point, not the ultimate outcome of the effort.

User satisfaction requires clear criteria

A stronger user focus is not a call to replace every metric with a vague sense of quality. Satisfaction should ideally be examined in relation to the specific task at hand. Can a potential customer find the information they need? Do they understand the requirements of an offer? Can they complete their next step without unnecessary obstacles?

Longer dwell times are not automatically positive. They can indicate thorough interest, but also confusion. A short visit might signify a failed contact or a quickly answered question. Accordingly, the evaluation should differentiate between an overview text, a technical guide, and a sales page. A single usage metric is insufficient to address these different needs.

Google describes the use of aggregated and anonymized interaction data to assess relevance. However, this does not mean that any metric from your own web analytics can be directly controlled as a ranking factor. A more sensible approach is to improve task fulfillment rather than artificially generating supposed user signals. Technical on-page work on accessibility, internal linking, and understandable page titles remains a necessary component of this.

Distribution creates connection, not reach

A results-oriented distribution chain begins with a specific target group and doesn't end with the first visual contact. A specialist article can generate attention, an in-depth website can build trust, a webinar can answer open questions, and a consultation can assess suitability. This gives the channels different tasks, instead of each having to deliver the highest possible individual value independently.

For planning purposes, it's crucial that every transition is clearly defined. A section for technical professionals shouldn't end abruptly on a generic company homepage without explanation. A page for business decision-makers requires different information than a maintenance manual. The real optimization often lies in this alignment: Is the anticipated expectation met, and is the next step meaningful for the user?

Not every company needs to be active on every platform simultaneously. Additional channels incur production, maintenance, and coordination costs. Focusing on a few suitable touchpoints can be more economical than a broad presence without any connection. The benefit of such a distribution strategy lies in combining search engine optimization with expert communication, sales, and customer support.

The brand has an impact, but not a magical one

A well-known brand can make it easier to find your way around. Therefore, for effective marketing purposes, it's advisable to separate brand-related search queries from general topic-based queries. Otherwise, an increase in already known demand could be mistakenly interpreted as a success in reaching new target groups. This would be particularly problematic if budgets were then allocated to measures that merely capitalize on existing interest.

The opposite error is also possible. If a specialist article only leads to a search for the company name later on, its earlier contribution is easily underestimated when considering only the most recent touchpoint. The development of brand search queries is therefore a helpful indicator, but not the sole proof of effectiveness. Product launches, trade fairs, and existing customer relationships can all influence this same metric.

For smaller companies, this doesn't mean that only large advertising budgets guarantee visibility. A specialized brand can be relevant in a narrow problem area without reaching a broad public. The crucial factor is that expertise is demonstrable: through consistent messaging, clearly defined contacts, and verifiable results. A brand is an economic offer of trust, not a universal technical key for every search engine.

Attribution is not proof of effect

Attributing a conversion to a channel answers a different question than measuring additional impact. Attribution distributes observed success across documented touchpoints. Causal analysis, on the other hand, investigates which part of the success would not have occurred without a specific action. Even a sophisticated attribution model does not automatically solve this second problem.

That large datasets alone are insufficient is demonstrated by a scientific study of 15 advertising experiments with approximately 500 million user observations. The methods based on observational data often failed to reliably reproduce the results of randomized trials. This is not direct proof of SEO effectiveness, but it is a methodological indicator: detailed behavioral data does not necessarily eliminate the differences between individuals who are already ready to buy and those who are influenced by marketing.

For SEO, controlled comparisons can be implemented under suitable conditions. Comparable groups of pages can be improved incrementally, while others initially remain unchanged. Seasonal trends, different starting points, and potential spillover effects must be taken into account. With small B2B sample sizes, an honest range is often more reliable than a seemingly precise return on investment.

Budgets need a new return-on-investment logic

For budget management, the key question should be what additional economic impact a further investment is likely to generate. The potential contribution margin of additional sales must be compared to the costs for content, technology, data, distribution, and support. Revenue alone is insufficient, as different offers have varying margins and follow-up expenses. The time lag between investment and expected return must also be considered.

This changes the priorities. A further, more detailed text isn't worthwhile simply because a search term has high search volume. What matters is the likelihood of qualified demand, one's own competitiveness, and the ability to meet that demand. A technical fix can be more economically important than additional content if it restores a central access point or a relevant closing process.

In addition, companies should assess their dependencies. Those who rely heavily on a single platform for new customer acquisition are particularly vulnerable to its changes. Therefore, additional suitable channels can be beneficial, even if their short-term return initially appears lower. However, diversification also requires investment and attention. It should mitigate specific risks, rather than simply adding as many channels as possible without prioritizing them.

AI analyses require human oversight

AI can play a supporting role in the evaluation process: organizing search queries thematically, grouping similar information needs, and suggesting anomalies for further investigation. However, this does not shift the responsibility for data quality to the model. A convincingly formulated explanation remains incorrect if the underlying information is corrupted or improperly combined.

Binding calculations should therefore be based on verifiable rules. A language model can suggest hypotheses for a decline; the determination of the actually affected URLs, time periods, and countries must remain traceable. Thematic classifications also require sample checks. Productive use consists of controlled division of labor, not the automatic adoption of every plausible interpretation.

A new form of pseudo-precision would be particularly dangerous. If a tool uses incomplete data to generate a comprehensive narrative about brand impact, user satisfaction, and market share, the report can appear more authoritative than its underlying data. It is more sensible to differentiate key findings according to their status as evidence: observed, plausibly explained, or still requiring verification. Less scrutiny of detailed rankings must not be replaced by uncontrolled interpretation at a higher level.

The SEO market is reorganizing itself

More complex data collection and additional quality checks are changing the economic landscape for tool providers and agencies. Increased data requests, the resolution of redirects, and repeated checks increase the workload per reliable observation. This may have market consequences, but it is not yet a certainty.

Larger providers can potentially spread infrastructure costs across more customers. Smaller specialists can remain competitive through narrower topic coverage, proprietary data, and individual review. Higher costs could lead to increased prices, less frequent updates, or a focus on economically relevant search areas. Which combination prevails depends on technology, competition, and willingness to pay.

Agency services also need to be more clearly explained. The number of optimized keywords or automatically generated reports is insufficient proof of performance if the connection to the business remains unclear. Reliable diagnoses, sensibly prioritized measures, and integration with product communication and sales are becoming more valuable. Holistic thinking doesn't mean that a single person has to master all disciplines. It means anchoring one's work within a shared economic logic.

Ability to act instead of perfect measurement

A practical restructuring begins with an inventory of the existing success logic. Which key performance indicators (KPIs) serve only as a guide, which trigger actions, and which influence budget decisions? Automatic reactions to individual daily values ​​are particularly critical. A perceived drop in rankings should only trigger major interventions after a thorough review of measurement quality, data age, and additional indicators.

Subsequently, a manageable chain of effects should be defined for each important topic area. This links relevant search contacts with appropriate website content, meaningful follow-up actions, and, ideally, well-documented business results. Details remain available for operational review. At the management level, however, a smaller number of clearly interpretable metrics, supplemented by information on data quality, suffices.

The new way of working demands two forms of simplicity. In terms of content, offers and next steps must become more understandable for users. In terms of management, companies must stop confusing available detailed data with established knowledge. The more demanding benchmark is making better decisions with incomplete information: based on verifiable data, independent value, and a consistent link between visibility and business.

In industrial B2B, sustainable business relationships rarely emerge overnight. They develop step by step – through visibility, professional relevance, recurring touchpoints, and growing trust. Xpert.Digital's 4-stage model addresses precisely this: It offers a structured path that begins with a manageable entry point and can evolve into deeper collaboration in business development if needed.

Instead of relying on loud marketing promises, this model puts the relationship at the forefront. Companies start with clearly defined, easily calculable measures and then decide, based on their own experience, how far they want to expand the collaboration. A key factor for this undisturbed trust-building process: The platform completely avoids annoying advertising ads, so the editorial focus remains solely on the companies' expertise.

More information here:

• From visibility to trust: Your scalable path with Xpert.Digital

☑️ Our business language is English or German

☑️ NEW: Correspondence in your native language!

 

I and my team are happy to be available to you as your personal advisor.

You can contact me by filling out the contact form here wolfenstein@xpert.digital:or simply call me at +49 7348 4088 965. My email address is

I'm looking forward to our joint project.

 

 

☑️ SME support in strategy, consulting, planning and implementation

☑️ Creation or realignment of the digital strategy and digitization

☑️ Expansion and optimization of international sales processes

☑️ Global & Digital B2B trading platforms

☑️ Pioneer Business Development / Marketing / PR / Trade Fairs