Estimated reading time: 11 minutes
The key takeaway in a sentence: According to Taboola’s study “The State of Performance Marketing in Spain,” 59% of Spanish marketing professionals believe that the decline in traffic caused by artificial intelligence is the main risk factor for traditional performance channels, ahead of attention fragmentation (50%) and rising CPC and CPM costs (36%). And 79% are already considering incorporating new channels into their media mix frequently or very frequently.
In other words: the Spanish industry has stopped debating whether AI will change customer acquisition. It now accepts that it is already doing so and is looking at where to reallocate its budget.
What the study says exactly
The report was based on a sample of 100 marketing professionals in Spain—from both brands and agencies—with an average age of 37 and responsibilities that include planning performance campaigns, managing advertising budgets, and optimizing CPC and CPM.
When asked about the emerging factors that pose the greatest threat to traditional performance channels, the ranking is as follows:
| Risk factor | % reporting it |
|---|---|
| Traffic reduction caused by AI | 59% |
| Fragmentation of Consumer Attention | 50% |
| Rising CPC and CPM Costs | 36% |
| Algorithmic unpredictability | 34% |
| Privacy Restrictions | 32% |
| Decrease in organic reach | 27% |
And the direct impact on budget management: 79% of respondents say they consider adding new channels to their media mix frequently (54%) or very frequently (25%).
Source: Taboola study reported by PR Noticias (September 2, 2026) — https://prnoticias.com/2026/09/02/mas-alla-de-search-y-social-los-nuevos-canales-del-marketing-en-espana
Why These Six Factors Are Actually One
It’s tempting to read the table as a list of separate issues. It isn’t. The six describe the same phenomenon from different angles: the intermediary between your brand and your customer has changed, and it no longer lets you see inside.
- AI reduces traffic because it answers the query without requiring a click.
- Attention is fragmented because discovery channels are multiplying (assistants, feeds, newsletters, communities, retail media).
- CPC is rising because all advertisers are focusing on the inventory that can be measured.
- The algorithm is becoming unpredictable because platforms are shifting optimization into black boxes.
- Privacy is becoming a constraint because the conversion signal that fueled those bids is getting weaker and weaker.
Here’s a statistic that confirms this from the platforms’ perspective: In August 2026, the IAB published the document *Measuring Visibility in the AI Era* as part of its Project Eidos initiative and found that only 16% of brands systematically measure their visibility in AI, while more than twenty providers are already selling this service using query sets and scoring systems that are incompatible with one another. In other words: nearly the entire market perceives the threat, yet almost no one is measuring it.
The strategic aspect: what to do, step by step, in this new era
The natural reaction to the 79% seeking new channels is to rush out and open new channels. That’s the wrong reaction—or at least an incomplete one. Opening a channel you don’t know how to measure only multiplies the original problem.
The correct order is: first, regain the ability to explain what’s happening; then, diversify. Here are the specific actions, grouped into four blocks.
Block 1 — Restore Measurement Before Adjusting the Mix (Weeks 1–2)
If you can’t answer “why performance changed,” any decision about the media mix is a gamble.
- Set up a measurement layer outside of the advertising platforms. Use your own data warehouse (BigQuery, Supabase, or even a consolidated spreadsheet if you’re a small business) that receives daily data from GA4, Search Console, Google Ads, Meta, and your CRM. The rule is simple: if the only place your data exists is in the dashboard of the platform that sells you inventory, you don’t have measurement—you have a statement of account.
- Reintroduce the direct attribution question. Add “How did you hear about us?” as an open-ended or optional field on all forms and in the checkout process. With digital signals becoming less reliable, the customer’s own stated response is once again one of the most trustworthy sources available. It takes an afternoon’s work and provides a time series that doesn’t depend on any cookies.
- Track traffic from AI assistants. Segment visits with referrers from ChatGPT, Perplexity, Gemini, and Copilot in GA4 and your CRM. Save them as an audience and as a report segment starting today, because historical data cannot be created retroactively.
- Set a baseline before the platforms migrate your campaigns. Starting September 1, 2026, Google began automatically migrating campaigns with Broad Match and automatically generated assets to AI Max, and Microsoft has enabled its own AI Max by default for new Search campaigns. Freeze a report today showing CPA, ROAS, and volume for the last 8 weeks per campaign. Without that baseline, in November you won’t be able to tell the difference between “AI Max is working” and “September was a good month.”
Block 2 — Turning SEO into Visibility in AI (Weeks 2–8)
The 59% who fear a loss of traffic are describing a real problem, but the goal is no longer just the session—it’s the mention.
- Define your set of brand-specific prompts and measure them at regular intervals. Include 30 to 50 questions that a real customer would ask before buying from you (“best automation agency for SMBs in Valencia,” “WooCommerce vs. Shopify for a B2B catalog”). Run these prompts on the four main chatbots once a month—always the same ones—and track three things: whether you’re mentioned, at what point in the conversation, and which sources the model cites.
- Use the IAB framework to hold tools to a high standard. The document organizes metrics into four Ps—presence, prominence, portrayal, persuasion—and distinguishes between “directional” and “decision-grade” data. When dealing with any AI visibility provider, ask which of the four Ps it actually measures, whether its output is directional or decision-grade, and what its query set is. Anyone who cannot answer those three questions is selling a number, not a measurement.
- Rewrite the key pages in a format that’s easy to extract. For a model to cite you, it needs to be able to extract the following: an explicit definition in the first 100–150 characters (“X is…”), a comparison table on each page, specific figures with dates, and a FAQ section using the schema
FAQPage. A nice-looking but empty introductory paragraph is invisible to an assistant. - Focus on the sources that the models cite, not just your own domain. If, in your prompts, the assistants repeatedly cite three directories, two industry publications, and one comparison site, those six URLs are your new link-building opportunities. Whether it’s a profile, a review, original content, or a collaboration—the format doesn’t matter; what matters is being featured in the source that the model reads.
- Check the facts before jumping to conclusions. The generative AI performance report in Search Console has had an error in its impression logging since August 13, 2026. It’s a logging error, not an actual drop in visibility. Check the date before drawing conclusions based on a graph.
Block 3 — Diversify the mix judiciously (months 2–4)
Now it’s time for new channels. But with one rule: no new channel is added without a written hypothesis and defined selection criteria before spending the first euro.
- Try ChatGPT Ads with a learning budget, not a scaling budget. The platform has been available in Spain since late August 2026, with self-service Ads Manager in beta, and in the European Economic Area, ad personalization is not initially available, so the focus is on conversational context and ad relevance, not on the audience. A 4- to 6-week test budget, creatives geared toward the comparison phase—not the conversion phase—and KPIs focused on reach, not last click. The mistake to avoid is treating it like Google Ads within a chat window.
- Strengthen the channels you already have. Email and WhatsApp are the only two platforms where you can reach your audience without paying an “algorithm toll.” Concrete action: Audit how many contacts with valid consent you have, when the last useful message was sent, and what percentage of your revenue comes from your own list. If it’s less than 15%, that’s where your most profitable investment for the next six months lies—and it doesn’t require opening any new channels.
- Consider retail media and marketplaces if you sell physical products. This is the inventory where purchase intent remains explicit and measurable. Watch out for the fine print: On August 10, 2026, Amazon automatically enrolled existing Sponsored Products campaigns in creator placements, maintaining bids and budgets, so check what’s actually running in your accounts before assuming you know where your ads are appearing.
- Replace generic content with content that includes your own data. In an environment where AI generates an endless stream of average content, the one thing it cannot replicate is your data. An annual study with aggregated data from your customers, an industry benchmark, a useful calculator—these are cited, linked to, and help maintain visibility in the AI landscape all at once.
Block 4 — Protecting What You Already Have (cross-curricular, this week)
- Check what your pixels and customer lists are transmitting. In July 2026, the FTC filed a lawsuit against Hims & Hers, along with the states of Utah and California, for sharing sensitive health data with Meta, Snap, and others through customer lists and pixels. In Spain, the same practice falls under the GDPR and is regulated by the AEPD. Almost no team has reviewed what those lists contain or on which pages the pixel is placed, usually because the person who set it up is no longer with the company. It’s an audit that takes just an afternoon and can prevent a six-figure problem.
- Check your bids for brands that aren’t yours. In August 2026, the FTC reached a $2.1 million settlement with Doxo for using third-party names, logos, and URLs in search ads in a way that implied a nonexistent affiliation. If your search strategy includes bidding on competitors, make sure the ad copy and landing page do not suggest any connection.
How to Prioritize When You Can Only Do Three Things
For a small team or an SME without a dedicated marketing department, the order is as follows:
- Baseline and proprietary measurement layer (Actions 1, 2, and 4). Without this, everything else is just opinion.
- A set of brand prompts and key pages in an exportable format (actions 5 and 7). This is what safeguards your future visibility and doesn’t depend on your advertising budget.
- Pixel and list audit (Action 14). High risk, low cost—it can be resolved in an afternoon.
New channels will come later. The 79% of respondents in the study are right to want to diversify, but those who diversify without tracking their own results will be having the exact same conversation in 2027 that they’re having today.
Frequently Asked Questions
It is the discipline of measurable customer acquisition applied to an environment where the intermediary between the brand and the customer is increasingly a language model rather than a search results page. Three things are changing: the metric for success is shifting from clicks to mentions plus assisted conversions; the attribution signal is deteriorating and must be reconstructed using self-reported and first-party data; and optimization is taking place within automated systems that advertisers cannot inspect.
No. It means they are no longer sufficient on their own and must be measured from outside the platform. Search continues to capture explicit purchase intent; what has changed is that it is no longer the only measurable touchpoint, and much of the optimization is now automated by default.
There’s no one-size-fits-all figure, but the approach is replicable: a limited learning budget (an amount you can afford to lose entirely without impacting the business), a 4- to 6-week testing window, and hypotheses and decision criteria established in writing before you begin. If a new channel consumes a budget that was already generating proven returns, that’s not diversification—it’s a cutback.
Using a fixed set of prompts run at regular intervals to track presence, position, and cited sources. There are commercial tools available, but before purchasing one, it’s advisable to apply the IAB framework’s criteria: what exactly it measures, whether the data is directional or suitable for decision-making, and what set of queries it uses.
At Inprofit
We are a MarTech agency with offices in Alicante and Valencia. We work precisely at the intersection described in this study: SEO and GEO strategy, development and e-commerce (WooCommerce, Shopify, Odoo), and automation with n8n and Make so that the measurement layer and acquisition flows do not depend on a third-party dashboard.
If you’d like to start with the diagnostic—a baseline analysis of campaigns prior to automatic migrations, a set of brand-specific prompts, and an audit of pixels and lists—please contact us, and we’ll analyze your actual data.
Especialista en SEO y Paid Media | Google Ads, Meta Ads, LinkedIn Ads & Search Console en vena Optimizo visibilidad orgánica + escalo adquisición pagada con ROAS obsesivo y estrategias data-driven.
Especialista en Performance Digital, Core Web Vitals, E-E-A-T, algoritmos de subasta, attribution multi-touch y maximizar LTV/CAC.
