Digital Advertising 2026: First-Half Review

We’re hitting June 2026 with a digital advertising landscape that January was already hinting at, and these six months have confirmed: AI Overviews has redrawn search, Performance Max is still the ecommerce workhorse, and the move to cookieless measurement went from “project” to “daily reality with consequences.”

This review pulls from what I see in client accounts spending between €3,000 and €30,000 a month, what public sources confirm, and what’s worth preparing before Q4 arrives.

30-second summary:

• AI Overviews now appears on a meaningful share of informational searches in Spain; organic traffic to direct-answer pages has fallen, but commercial-intent queries are holding steadier
• Performance Max still captures 40-60% of spend in catalog ecommerce accounts I audit
• The shift to Consent Mode v2 and modeled conversions took months to settle; accounts that delayed it are paying for it with worse bidding decisions
• TikTok Ads investment in Spanish ecommerce is up, but CPA is wildly uneven by sector
• Most valuable lesson of H1: review attribution before debating budget

What actually changed in digital advertising in H1 2026?

Three things matter more than the rest.

The first is AI Overviews. Google launched it globally in 2024 (Google Blog), and since late 2025 it’s been live on Spanish-language informational queries. The impact on organic traffic is real for pages that answered specific questions (“how do you X”, “what does Y mean”). For commercial pages (services, products, comparisons), the effect has been more limited, because AI Overviews still hands the click to the user when there’s purchase intent.

The second is Performance Max consolidation. For accounts with product catalogs, it’s still the campaign that generates the most volume, but also the one that accumulates the most errors when set up poorly: thin creative assets, no separation between brand and non-brand, no audience exclusions.

The third, less visible but more structural, is the end of third-party cookie-based measurement. Chrome started phasing them out in 2024 (Google Privacy Sandbox), and by 2026 modeled conversions is the default mode in Google Ads. Accounts that turned on Consent Mode v2 and Enhanced Conversions on time are measuring reasonably well. The ones that didn’t are making bidding decisions on incomplete data without realising it.

How has AI Overviews affected search traffic?

The hit to informational SEO is the best-documented piece. When Google shows a complete answer at the top with citations, the share of users who click any organic result drops.

For paid search the picture’s different. Search ads still appear above AI Overviews on commercial queries, and most transactional-intent searches (“buy X”, “Y price”, “Z consultant”) either don’t trigger AI Overviews or do so very narrowly.

What I do see in accounts with meaningful organic traffic: the split between organic and paid clicks has shifted. Terms that used to pull plenty of organic traffic (“how to set up Google Ads”, “what is an RSA”) are losing organic impressions, which pushes us to cover that demand with informational Search campaigns or with Dynamic Search Ads. The budget stays the same; what changes is where the click comes from.

A question worth asking: is there any keyword in my account where organic traffic has fallen more than 20% in the last 90 days? If the answer is yes, that term is probably on a page hit by AI Overviews. Two options: rewrite the content so it gets cited by the AI answer, or fill the gap with a search ad.

Performance Max after 3 years: where it wins, where it loses

Performance Max is hitting three years in serious ecommerce accounts. The mid-2026 conclusion: it wins on coverage, loses on control.

It wins when three conditions hold: a product catalog with a clean Merchant Center feed, enough conversions to feed the algorithm (at least 30-50 a month per campaign), and a clean split between brand and non-brand campaigns to avoid cannibalization.

It loses when the feed has errors, when products with very different margins are mixed without segmentation, and when brand-search exclusions aren’t applied. The typical case I see in audits: PMax looks like it has excellent ROAS, but cross-checking with Google Analytics shows that 30-40% of conversions came from brand searches the account would have captured anyway.

In ecommerce accounts with a clean catalog structure, Performance Max usually captures 40% to 60% of total spend when set up correctly. If it’s pulling more than 70%, it’s probably cannibalizing other campaigns with attributed data that doesn’t really belong to it.

For small accounts (€500-€2,000/month), PMax is still hard to optimize for lack of conversion volume to feed the model. In that range, a standard Search + Shopping setup still gives more control and fewer surprises.

Cookieless measurement: what works and what doesn’t in 2026

The phaseout of third-party cookies in Chrome has had a quiet but cumulative effect. In 2025 it could still be postponed. In 2026 it can’t.

What works reasonably well:

• Enhanced Conversions in Google Ads. Sends hashed email/phone data from the form, lets Google match it with logged-in users and reconstruct conversions the cookie no longer captures.
• Consent Mode v2 implemented correctly. Accounts with properly modeled explicit consent are seeing 15-30% more reported conversions than those that lost data through bad configuration, based on what I observe in accounts that migrated in 2024-2025.
• Meta Conversions API (CAPI) with server-side data. The Meta cookie stopped being useful a while back; CAPI with hashed data is how you keep measurement reasonable.

What doesn’t work:

• Relying on GA4 alone, without Enhanced Conversions or CAPI. The gap between GA4 and the ad platforms has grown so wide that taking decisions from isolated GA4 is a mistake.
• Last-click attribution. With cookies limited, last-click consistently undervalues Display, YouTube, and top-of-funnel campaigns.

In audits, the first thing I check now isn’t campaign structure but measurement health: whether Consent Mode v2 is right, whether Enhanced Conversions is on, whether modeled conversions are a reasonable share of total. Without that, optimizing campaigns is optimizing on contaminated data.

Meta vs Google vs TikTok: how budget splits in H1 2026

In Spanish ecommerce accounts I see, the typical split looks like this:

Meta is still the acquisition engine in visual-product ecommerce, but CPA pressure has gone up. The accounts that perform best are the ones combining Advantage+ Shopping Campaigns with creative tested weekly.

TikTok has cemented its position in sectors where the audience is genuinely young. For accounts targeting over-35 audiences or B2B products, CPA tends to run too high and you spend more on testing than what comes back in conversions.

Google Ads is still the most stable, but the Performance Max dependency means configuration errors hit harder than before. A poorly structured PMax campaign can burn €5,000 in a month before the account owner notices, when the end-of-month report drops.

What to prepare for H2 2026

Four fronts deserve attention before September.

Q4 with less comparable historical data. A year ago at this time of year, attribution worked differently, AI Overviews wasn’t live, and Consent Mode v2 was mid-migration. Comparing Q4 2026 to Q4 2025 will be misleading without context. Document the changes in your measurement setup so you can explain the differences.

Black Friday with steeper competitive pressure. Meta CPMs and Google Ads CPCs rise every November. In 2026 the bump may be sharper because small brands that paused spend in 2024-2025 are coming back. Lock down brand-protection budget with exact keywords before October.

Merchant Center feed audit. If your PMax depends on the feed (and it does), a feed cleanup in July/August can move Q4 ROAS more than any bidding tweak. Products with poor images, missing attributes, or negative margins are the first candidates to exclude.

Plan B for AI Overviews on commercial queries. If Google decides to turn on AI Overviews for purchase-intent queries during H2, search-ad pressure goes up. Having landing pages that answer specific questions with verifiable data is the closest you can get to “ready.”

Most valuable H1 lesson: fix attribution before debating budget

If I had to call out one lesson from these six months, it’d be this: budget decisions taken on flawed attribution data cost us more than any fine bidding tweak.

In recent audits I found two especially clear cases. The first: an account about to cut YouTube spend over “low ROAS”, when the actual problem was last-click attribution undervaluing the channel. The second: an account that doubled Performance Max budget because “it was the most profitable,” when in reality PMax was absorbing brand conversions arriving through organic SEO.

Before debating whether to push Meta budget up 20% or pull Google budget down 10%, it’s worth answering: what attribution model are you using? Is GA4 reasonably aligned with the platforms? How many conversions are being modeled? If the answers are fuzzy, any budget reallocation is a bet, not a decision.

Frequently asked questions on digital advertising in 2026

Does AI Overviews affect my Google Ads campaigns?

Directly, not much yet. Search ads still appear at the top even when AI Overviews is active. The indirect effect is bigger: organic traffic to informational pages drops, which forces you to cover that demand with Search ads or with Dynamic Search Ads. In accounts that lean heavily on SEO, the overall cost per acquisition has gone up.

Is Performance Max still the best option for ecommerce in 2026?

For accounts with a product catalog, spend above €2,000/month, and at least 30-50 monthly conversions, yes. For smaller accounts or single-product setups, a standard Search + Shopping structure usually gives more control. The Google Shopping vs PMax comparison details the selection criteria by account size.

Is it worth starting to invest in TikTok Ads in 2026?

Depends on the sector. Fashion, beauty, home decor, and visual product with under-35 audiences work reasonably well. For B2B, professional services, or long-cycle products, TikTok CPA tends to run too high. Before starting, plan a testing budget of €1,500-€3,000/month for at least 8 weeks.

How much priority should measurement get in my H2 strategy?

The highest. Without Consent Mode v2, Enhanced Conversions, and CAPI working, bidding decisions get made on partial data. In audits it’s the first thing I review before looking at campaign structure. A Google Ads account audit guide covers the critical points to verify.

Is it a good time to cut Meta spend and raise it in Google?

There’s no universal answer. It depends on product type, audience, and how attribution is working in your account. What I do recommend: before moving budget between platforms, verify that measurement is comparable. Moving spend based on last-click attribution in GA4 tends to penalize Meta and Display artificially.

Conclusion: H2 2026 will reward measurement discipline

The first half of 2026 has confirmed what was already in the air: digital advertising has gotten less intuitive and more dependent on data infrastructure. The accounts performing best aren’t the ones with the best creative or the finest bidding; they’re the ones with measurement in order.

For H2, the priorities are clear: finish the Consent Mode v2 migration if it’s still pending, clean the Merchant Center feed before Q4, set up brand protection for Black Friday, and accept that last-click attribution no longer works for budget decisions.

If you want a review of your account’s measurement setup before Q4, you can book 30 minutes of consulting. The most useful conversation is still figuring out what’s being measured well and what isn’t, before touching bids.