The quiet crisis: your search safety net is fraying
Look at those headlines and a pattern jumps out:
- News publishers expect search traffic to drop 43% by 2029.
- Google downplays GEO while AI-generated SERPs get noisier.
- How much can we influence AI responses? How to choose a link-building agency in the AI SEO era.
- Generative Engine Optimization tools. AI citations up 642% with “simple semantics.”
Translation: the organic and paid search environment you built your growth model on is being quietly rewritten. At the same time:
- Google is opening Olympic inventory to biddable CTV.
- Streaming games are setting audience records.
- Social-first ranking strategies and short-form video are where attention is actually happening.
Most teams are reacting tactically: “optimize for SGE,” “test more short-form,” “try CTV.” That’s not enough. If you’re a CMO, performance lead, or media buyer, the real job now is to rebuild your media system so you’re not dependent on a channel (classic Google search) that is structurally de-prioritizing you.
The structural shift: from “10 blue links” to AI answers and attention feeds
Three shifts are colliding:
1. Search is becoming an answer engine, not a traffic engine
Between SGE-style AI answers, “People also ask,” and endless SERP features, Google is working harder than ever to keep users on Google. Publishers see it clearly, but brands are exposed too:
- AI answers compress the funnel. Users get “what” and “why” without clicking.
- Generic queries (“best X for Y”) are increasingly mediated by AI and affiliates.
- Branded and high-intent queries are getting more expensive (see the “100 most expensive keywords” lists creeping up every year).
Your “efficient” search machine is being taxed from both ends: less organic surface area, more paid competition, more automation that hides the underlying mechanics.
2. Automation is replacing understanding
Search Marketing’s “insight gap” is real:
- Smart Bidding + broad match + RSAs + Performance Max = performance, but also opacity.
- Teams are optimizing to platform-reported ROAS instead of customer reality.
- AI copy, AI landing pages, AI bid strategies: lots of output, thin understanding.
You can hit short-term targets while losing the signal you need to make big, strategic calls: which messages actually move people, which segments are worth defending, which channels are compounding.
3. Attention is moving to feeds and streams, not queries
While search gets weirder, attention is consolidating elsewhere:
- Social-first ranking strategies and short-form video “what’s working now” guides are everywhere.
- Streaming sports and CTV are mainstream, addressable, and biddable.
- Brands are suddenly proud to say “we commit to human-generated content” because AI sludge is flooding feeds.
The center of gravity is shifting from “capture intent when they search” to “shape intent where they already spend time.”
The real risk: performance addiction in a changing market
None of this means search is dead. It means the model of “search as the reliable, infinitely scalable profit engine” is dying.
The teams most at risk share a pattern:
- 60-80% of incremental revenue is still coming from Google Ads + branded search + shopping.
- Organic traffic is overly reliant on a handful of commercial-intent pages.
- Media mix models are really just “search + everything else as a rounding error.”
- Brand budgets are the first thing cut when CAC creeps up.
That model works until:
- AI answers start absorbing your comparison queries.
- Affiliates and marketplaces outrank you on your own category terms.
- Broad match starts spending into fuzzy intent that looks good in-platform but doesn’t hold on LTV.
By the time you see it clearly in your dashboards, the compounding damage (to brand search volume, direct traffic, and word-of-mouth) is already done.
The job now: rebuild your media mix around owned demand
You can’t control Google’s roadmap. You can control how dependent you are on it. The practical move is to shift from a “search-centered” system to a “demand-centered” system that uses search as one of several capture mechanisms.
Four moves matter.
1. Separate “captured demand” from “created demand” in your planning
Most budgets are still organized by channel. That hides the real question: are we buying demand someone else created, or are we creating it?
Reframe your plan into two buckets:
- Captured demand: paid search, shopping, retargeting, affiliate, marketplaces, some SEO.
- Created demand: social video, influencers, CTV, upper-funnel display, PR, brand content, email, community.
Then:
- Set explicit targets: e.g., “By Q4, at least 40% of new customers sourced from created-demand channels.”
- Report separately: one CAC for capture, one for creation. Stop averaging them into one fuzzy number.
- Protect creation budgets from short-term cuts by tying them to leading indicators (more on that below), not just last-click ROAS.
2. Treat AI search and SGE as a distribution channel, not a black box
The AI SEO / “Generative Engine Optimization” conversation is noisy. Ignore the hacks; focus on being the kind of source AI systems like to cite.
That means:
- Entity clarity: Make your brand, products, and people machine-legible. Structured data, consistent naming, clean internal linking, and clear “about” pages.
- Topical depth: Instead of 50 thin pages, build 5-10 deep, updated resources that actually answer the full problem space.
- Evidence-rich content: Original data, case studies, clear methodologies. AI systems are more likely to cite content with concrete proof, not fluffy listicles.
- Semantic alignment: The “simple semantics increased our AI citations by 642%” type results usually come from aligning language with how users and models talk about the topic, not keyword stuffing.
Operationally, assign someone to own “AI surfaces” the way you once assigned someone to own “featured snippets.” Not as a magic growth lever, but as insurance: if AI is going to answer, you want your brand inside the answer, not erased by it.
3. Build a social-and-CTV “demand creation spine” that isn’t a brand vanity project
Social-first ranking strategies, short-form video, creator programs, and biddable CTV aren’t “brand fluff” anymore. They’re your hedge against a world where people see you before they ever search.
The key is to make them commercially accountable without forcing them into last-click performance logic.
A workable spine looks like this:
- Anchor formats: Pick 1-2 primary content formats that you can ship weekly for a year (e.g., TikTok/Reels-style short-form + a recurring YouTube or CTV-friendly series).
- Audience-first, not algorithm-first: Define 2-3 sharp audience POVs (e.g., “the brutally honest CFO’s guide to SaaS waste,” “the no-BS skincare scientist”). Algorithms change; POVs travel.
- Paid amplification with guardrails: Use paid social and CTV to guarantee distribution to your ICP, but cap frequency and watch assisted conversions, branded search volume, and direct traffic as your success metrics.
- Creator and influencer as media, not PR: Treat creators as a media line item with clear reach and frequency goals, not just “content we can repost.”
Then connect it to search:
- Track branded search volume and direct traffic by geo where you’re heavy on social/CTV vs. control geos.
- Watch how those geos perform on non-branded search CAC and conversion rate.
- Use that delta as your argument to protect and scale demand creation spend.
4. Fix your measurement so automation doesn’t blind you
The insight gap isn’t inevitable; it’s a measurement choice. If you’re flying blind, it’s usually because:
- Everything is judged on platform ROAS.
- Attribution is either “last click” or “whatever GA4 says by default.”
- No one is looking at post-click behavior by channel and creative.
You don’t need a seven-figure MMM project to fix this. You need a few simple, durable practices:
- Source-of-truth events: Define 3-5 events that matter (qualified lead, first purchase, second purchase, subscription month 3) and track them consistently across all channels.
- Holdout tests: Regularly run geo or audience holdouts for big line items (brand search, CTV, major social campaigns) to see what actually moves incremental sales.
- Channel-specific LTV curves: Stop treating all customers as equal. Measure 6-12 month LTV by first-touch channel and major creative themes.
- Creative-level diagnostics: For automated campaigns (PMax, Advantage+, etc.), pull post-click stats by asset where possible: bounce rate, scroll depth, time on site, downstream conversion. Great automation with bad creative is just an efficient way to waste money.
The goal is not perfect attribution; it’s decision-grade clarity. Enough signal to say, “We can afford to reduce search by 15% and move that into CTV and creators without blowing up CAC over 6-12 months.”
What to do in the next 90 days
This doesn’t have to be a multi-year transformation project. You can start rewiring your system in a quarter.
Step 1: Audit your dependency on classic search
- Pull the last 12 months of new customers by first-touch channel.
- Calculate the share coming from:
- Paid search (brand + non-brand + shopping).
- Organic search.
- Everything else.
- If search (paid + organic) is over 50% of net-new, you have concentration risk. Over 70% and it’s a red flag.
Step 2: Design one meaningful demand-creation bet
- Pick a single, high-potential combination like:
- Short-form video + creator partnerships in your top 3 markets, or
- Biddable CTV + social retargeting for a flagship product, or
- A recurring editorial series amplified via LinkedIn + YouTube for B2B.
- Fund it by trimming the least efficient 10-15% of your search and retargeting spend.
- Set leading metrics: branded search volume in target markets, direct traffic, assisted conversions, and LTV of exposed cohorts.
Step 3: Assign ownership for AI surfaces
- Give one person (or squad) explicit responsibility for:
- Monitoring how your category appears in SGE / AI answers.
- Identifying 10-20 “must-win” topics where you want to be cited.
- Updating or creating content that is deep, evidence-based, and entity-clean for those topics.
- Review quarterly, not weekly. This is infrastructure, not a growth hack.
Step 4: Add one hard piece of measurement discipline
- Pick one of:
- A brand search holdout test (turn off brand search in a small geo and watch impact).
- A CTV or social holdout (control vs. exposed geos).
- A basic channel-level LTV analysis.
- Use the results to reset one budget decision. Don’t let it die in a slide deck.
The platforms will keep changing. AI answers will get better (and more extractive). CTV will get more crowded. Social algorithms will keep doing social algorithm things.
The teams that win the next few years won’t be the ones with the cleverest SGE hack. They’ll be the ones who stopped treating search as a safety net and started treating it as one channel in a broader, intentionally designed system for creating and capturing demand.