B2B SaaS SEO operated on a stable playbook for nearly a decade. Publish educational content, build topical authority, convert readers into leads. The funnel worked because Google sent traffic to informational content and users clicked through.
That model is breaking. AI answers informational queries directly. Zero-click searches dominate. Traffic to educational content is declining while the cost of producing it stays constant. Meanwhile, transactional and branded queries still convert, but competition there is brutal.
The result: SEO is splitting into two distinct games with different rules, different tactics, and different ROI profiles. Most SaaS companies are still playing the 2020 game, producing TOFU content that no longer drives pipeline. This guide explains what changed, why it matters, and what to do instead.
The Query Eligibility Problem Nobody Talks About
Before discussing content strategy, understand a structural constraint that determines whether you can rank at all.
Google classifies every query by intent and assigns eligibility to specific site types. For “best project management software,” the eligible site types are editorial and review platforms: G2, Capterra, PCMag, independent blogs. SaaS vendor sites are structurally ineligible because Google considers them biased sources for evaluative queries.
This is not a content quality issue. You can write the most comprehensive, well-researched comparison page on your domain. It will not rank for “best X software” because your site type is disqualified from that query’s eligible set.
How to identify your eligible query space:
Search your target keywords. Look at the top 10 results. Categorize each result by site type: editorial, UGC/forum, vendor, aggregator, video. If vendor sites appear, you are eligible. If the results are entirely editorial and UGC, you are structurally excluded regardless of content quality.
Practical implications:
Vendor sites can rank for branded queries (“Asana review,” “Slack pricing”), navigational queries (“Asana login”), and transactional queries with commercial modifiers (“Slack enterprise pricing”). They struggle with pure informational and evaluative queries where Google prefers neutral sources.
This means content strategy must start with eligibility mapping, not keyword volume. A 10,000 monthly search keyword where you cannot rank is worth zero. A 500 monthly search keyword where you are eligible and can convert is worth pursuing.
Why TOFU Content ROI Is Collapsing
The traditional B2B content playbook: publish educational blog posts targeting problem-aware searches, capture emails, nurture through the funnel. This worked when Google sent traffic to informational content and users had no alternative.
Two shifts broke this model.
First, AI provides direct answers. When someone searches “how to improve sales team productivity,” Google’s AI Overview synthesizes an answer from multiple sources. The user gets what they need without clicking. Your blog post might be cited, but you receive no traffic, no email capture, no attribution.
Second, informational content is commoditized. Every competitor publishes similar educational content. The marginal value of another “complete guide to X” is near zero. Users skim, bounce, forget. Brand recall from generic TOFU is minimal.
The math has changed:
A TOFU blog post costs $500-2000 to produce. In 2020, a ranked post might generate 1000+ monthly visits, 20+ signups, 2-3 SQLs. In 2025, the same post generates 200 visits (AI captures the rest), 3 signups, near-zero SQLs. Cost stayed flat, output dropped 80%.
What still works in educational content:
Original research with proprietary data. Survey 500 customers about workflow challenges and that data cannot be synthesized by AI because it exists nowhere else. Research earns backlinks, gets cited by LLMs, builds genuine authority.
Contrarian takes that challenge consensus. “Why we stopped using OKRs” generates engagement because it contradicts the standard narrative. Commodity content (“OKRs matter, here’s how”) adds nothing.
Deep technical content for practitioners. A guide to implementing webhook authentication serves users doing the work. They cannot get this from AI because they need your product’s specificity.
Transactional Queries: Where Conversion Actually Happens
If informational content is declining, where should investment shift? Toward queries where users are ready to act and your site type is eligible.
Branded queries are your highest-value SEO asset.
When someone searches “[Your Product] vs [Competitor]” or “[Your Product] pricing,” they are in buying mode. These queries have lower volume but conversion rates 10-50x higher than TOFU.
Most SaaS companies underinvest here. They have one pricing page, no comparison pages, no alternative pages. Meanwhile, competitors and review sites own this space.
Look at how Notion handles this. Search “Notion vs Coda” and Notion ranks with a dedicated comparison page. They control the narrative on their own brand queries instead of ceding that ground to third parties.
Build comparison pages even when uncomfortable.
Many SaaS founders resist writing “[Competitor] alternatives” or “vs” pages because it feels aggressive. This is a mistake. Users search these terms. If you do not rank, someone else controls how your product is positioned.
The tactic: Create honest comparison pages that acknowledge competitor strengths while highlighting your differentiation. Google rewards balanced content here. One-sided attack pages underperform because they trigger the “biased source” classification.
Pricing and feature pages need SEO treatment.
Pricing pages often get zero SEO attention. No schema, thin content, poor internal linking. Yet “[Product] pricing” is a high-intent query that converts.
Add context to pricing pages: what each tier includes, who it serves, how it compares to alternatives. This gives Google something to index beyond a pricing table and improves your eligibility for pricing-related queries.
The Topic Cluster Myth (And What Actually Works)
Content marketers love topic clusters. Build a pillar page, surround it with supporting content, interlink everything, watch rankings compound. The theory is elegant. The reality is messier.
Google does not rank clusters. Google ranks pages.
A cluster’s value comes from three mechanics:
- Crawl efficiency: Internal links help Google discover new pages faster
- PageRank flow: Link equity from pages that earn backlinks distributes to linked pages
- Topical signals: Anchor text in internal links signals what pages are about
None of these require the rigid hub-and-spoke structure that cluster frameworks prescribe. A well-interlinked site with strong content performs the same whether you call it a “cluster” or not.
The compounding effect comes from domain authority, not cluster architecture.
When you publish 100 pieces, some earn backlinks organically. That link equity flows to other pages through internal links. Over time, new pages start with higher baseline authority and rank faster. This is real. But it happens because of accumulated backlinks, not because you organized content into clusters.
What actually matters for internal linking:
Link from high-authority pages to pages you want to rank. Check which of your pages have the most backlinks and ensure they link to your priority targets.
Use descriptive anchor text. “Learn more” passes less signal than “project management software comparison.”
Create logical paths for users. If someone reads about a problem, link to your solution. This improves both SEO signals and conversion paths.
Skip the elaborate cluster documentation and visual maps. Just link contextually and strategically.
Programmatic SEO: When It Works, When It Fails
Programmatic SEO promises scale. Template-based pages targeting long-tail variations. Integration pages, location pages, use-case pages. When it works, it captures thousands of queries efficiently. When it fails, it creates thin content that tanks site quality.
Programmatic works when you have genuine data advantage.
Wise creates currency conversion pages because they have real exchange rate data. Yelp creates location pages because they have real business listings. The template adds value because the underlying data is unique.
If your programmatic pages are just keyword variations with swapped terms, you have no data advantage. “CRM for dentists” vs “CRM for lawyers” with identical content except the industry name is thin content. Google recognizes and penalizes this pattern.
Programmatic works for integration pages when integrations are real.
Zapier ranks for thousands of “[App A] + [App B] integration” queries because those integrations exist. The pages document real functionality. SaaS companies can replicate this with their own integration libraries if the integrations are substantive.
Template quality determines ceiling. A good template has unique data per page, addresses specific user intent, and provides information unavailable elsewhere. A bad template is keyword stuffing with dynamic insertion.
Common failure patterns:
Thin location pages with no local data. “Software development in Chicago” with generic content and a swapped city name.
Comparison pages with no real comparison. “X vs Y” where the Y section is scraped feature lists with no analysis.
Use-case pages with no use-case specificity. “CRM for healthcare” that reads identically to “CRM for finance.”
Before launching programmatic, validate that each page passes a simple test: does this page contain information a user could not get from the template alone?
LLM Visibility Is a Different Game
Optimizing for LLM citation (ChatGPT, Perplexity, Claude) follows different rules than Google optimization. Conflating them leads to wasted effort.
LLMs use retrieval-augmented generation (RAG).
When you ask ChatGPT a question requiring current information, it searches an index, retrieves relevant chunks (typically 500-1000 tokens), and synthesizes a response. Your content gets cited if your chunks score high in semantic similarity to the query.
What drives retrieval:
Chunk-level semantic match. LLMs do not retrieve full pages. They retrieve chunks. A page with one highly relevant section outperforms a comprehensive page where relevance is diluted.
Assertive, quotable statements. “The average SaaS churn rate is 5-7% annually” gets retrieved. “Churn rates vary depending on factors” does not. LLMs prefer specific claims because they are easier to synthesize.
Numerical specificity. Concrete figures survive summarization. Vague prose gets dropped.
What does not drive retrieval:
Schema markup. RAG systems work on text content, not structured data.
Entity optimization. Unless you are a major entity in Knowledge Graph, entity signals do not affect LLM retrieval.
Topic clusters. LLMs retrieve chunks, not site structures.
Practical optimization:
Write self-contained statements under H2 headers. Each section should be independently quotable.
Front-load specifics. Put concrete claims in the first sentence of each section.
Avoid hedge words. “Typically,” “generally,” “it depends” reduce retrieval probability.
Include data. Original statistics, benchmarks, and percentages are retrieval magnets.
AI Overviews: Prerequisites Before Optimization
Google AI Overviews select sources from existing top 10 results. If you do not rank organically, you cannot appear in AI Overviews. AIO optimization without ranking is pointless.
The selection criteria:
Existing ranking. AIO pulls from organic results, not the full index.
Source diversity. AIO prefers synthesizing multiple sources over citing one extensively.
Factual consistency. Claims that contradict other top results get deprioritized.
Statement extractability. Prose that can be quoted directly, without modification, is preferred.
What this means tactically:
First, rank. Use traditional SEO to reach top 10.
Then, structure for extraction. Clear H2s with direct answers. Specific claims. Avoid meandering introductions.
Monitor AIO appearance. Search your target queries in incognito. Note whether AI Overviews appear and who gets cited. Reverse-engineer what cited content does differently.
Do not sacrifice Google ranking for AIO optimization. They are not trade-offs. AIO eligibility requires ranking first.
Reddit and Quora: What They Actually Do for SEO
Reddit threads rank for thousands of SaaS-related queries. Quora answers appear in AI Overviews. This has led to advice about “Reddit SEO” and “Quora distribution.” Most of this advice misunderstands the mechanism.
Reddit ranking is not your ranking.
When a Reddit thread ranks for “best CRM for startups,” Reddit gets the traffic. You appearing in that thread does not mean you rank. You get visibility within Reddit’s traffic, but no direct SEO value flows to your domain.
The links are nofollow. The brand mention does not register as entity signal. You are borrowing Reddit’s authority, not building your own.
Actual value of Reddit presence:
Keyword research. Reddit surfaces real phrasing and pain points. How users describe problems differs from how marketers imagine they describe them.
Referral traffic. Direct clicks from Reddit to your site. Small volume but high intent.
Brand awareness. Users see your name in context of solving problems.
Quora has different dynamics.
Quora answers get cited in AI Overviews more frequently than most sources. A well-written Quora answer on a topic can appear in AIO even when your main site does not.
This is useful for visibility but not for traffic. AIO citations often do not result in clicks. The user gets the answer, sees your brand mentioned, and moves on.
How to use these platforms:
Participate genuinely. Answer questions where you have expertise. Build reputation within communities.
Do not treat as link building. The links do not pass SEO value.
Use for research. What questions recur? What language do users use? What objections appear repeatedly?
Track brand mentions. If your product gets discussed organically, monitor sentiment and respond to misconceptions.
Measuring What Matters (Traffic Is a Vanity Metric Now)
SEO reporting traditionally centered on traffic. Rankings, visits, pageviews. These metrics are becoming disconnected from business outcomes.
Why traffic misleads:
AI Overviews suppress clicks. You can rank #1 and get minimal traffic if AIO answers the query.
Informational traffic does not convert. 10,000 visits to a TOFU post generating zero pipeline is worse than 500 visits to a comparison page generating 5 demos.
Bounce rate masks quality. A user who finds their answer and leaves is not a failure for informational content, but it is also not business value.
Metrics that connect to pipeline:
Organic conversions by page. Which pages generate signups, demo requests, trial starts? Double down on what converts.
Assisted conversions. Organic sessions that touch the funnel before converting through another channel. This reveals SEO’s role in multi-touch journeys.
Ranking for transactional keywords. Track position on “[Product] vs [Competitor],” “[Product] pricing,” “[Product] reviews.” These keywords have low volume but high value.
Brand search volume. Is branded search growing? This indicates whether overall marketing (including SEO) builds recognition.
Revenue influenced by organic. Requires CRM integration. Tag leads by first-touch and multi-touch channels. Calculate pipeline and revenue where organic participated.
What to stop measuring:
Total organic traffic without segmentation. Useless as a single number.
Keyword count. Ranking for 10,000 keywords means nothing if none convert.
Domain authority. A third-party proxy metric with no direct relationship to your outcomes.
The Actual Strategic Decision
Every B2B SaaS company faces a resource allocation question: where to invest SEO effort given these shifts?
Option A: Double down on informational content and accept lower ROI.
This makes sense if you have strong brand and can convert awareness to pipeline through other channels. HubSpot still publishes TOFU content because their brand is strong enough that awareness translates to revenue even without direct conversion paths.
If you are not HubSpot, this is burning money.
Option B: Shift to transactional and branded queries.
Invest in comparison pages, alternative pages, pricing content, integration pages, use-case landing pages. Lower volume, higher conversion. This works for most SaaS companies, especially those below $50M ARR where brand is not yet a moat.
Option C: Invest in LLM visibility as a separate channel.
Create content optimized for LLM retrieval. Original research, data-driven insights, quotable statistics. Measure brand mentions in LLM outputs. Accept that traffic may be minimal but visibility influences perception.
This is experimental. Measurement is difficult. But early movers may build advantage.
Option D: Reduce SEO investment entirely.
If your category has weak search volume, brutal competition, or short sales cycles where SEO cannot influence, allocate elsewhere. Paid acquisition, partnerships, outbound. SEO is not mandatory for every SaaS company.
The worst choice is continuing the 2020 playbook without adjustment. Producing commodity TOFU content, measuring traffic, and wondering why pipeline is flat.
What to Do in the Next 30 Days
Concrete actions, not abstract principles.
Week 1: Audit query eligibility.
Take your top 50 target keywords. Search each. Record what site types rank. Categorize your eligibility. Identify where you can rank and where you are structurally excluded.
Week 2: Analyze current content ROI.
Pull organic traffic by page. Pull conversions by page. Calculate traffic-to-conversion rate. Identify top converters (usually transactional and branded content). Identify traffic-heavy zero-converters (usually TOFU). Reallocate production resources accordingly.
Week 3: Build transactional content gaps.
List all comparison queries (“[You] vs [Competitor]”) you should own. List all alternative queries (“[Competitor] alternatives”) where you should appear. Create a production queue prioritized by competitor size and query volume.
Week 4: Establish LLM baseline.
Search your product name in ChatGPT, Perplexity, Claude. Note how you are described, what is accurate, what is wrong, whether you are recommended. Repeat for category queries (“best X software for Y use case”). This is your baseline. Now you can measure whether LLM visibility improves.
Final Note
SEO is not dying. It is fragmenting into distinct games: traditional search, AI Overviews, LLM visibility, UGC platforms. Each game has different rules and different winners.
The companies that struggle are those applying uniform tactics across all games. The companies that win are those who understand which games they can play and concentrate resources accordingly.
For most B2B SaaS companies, the shift is clear: less commodity educational content, more transactional and branded content, more original research, more structured data for specific retrieval.
The playbook changed. Adjust or get left behind.