The Death of TOFU: Content Strategy in the AI Era

What We Tracked

6 B2B and B2C clients. All had content programs built on the traditional TOFU/MOFU/BOFU model. All had been running for at least 3 years with enough historical data to compare.

We classified every tracked keyword by funnel stage:

• TOFU: Informational intent. "What is X," "how does X work," "X explained," "benefits of X"
• MOFU: Consideration/evaluation intent. "Best X for Y," "X vs Y," "X reviews," "X pricing," "how to choose X"
• BOFU: Decision/transactional intent. "Buy X," "X agency," "X services," "hire X," "[brand] X"

Then we compared traffic from each funnel stage: current 6 months vs the equivalent period 18 months prior.

``` Funnel stage | Traffic change (18-month comparison)

• ------------|--------------------------------------

TOFU | -51% average MOFU | -12% average BOFU | +8% average ```

TOFU traffic has been cut roughly in half. MOFU traffic is down but holding. BOFU traffic is up.

The pattern is consistent across all 6 clients regardless of industry. The direction of change is the same everywhere. The magnitude varies.

---

What's Happening to TOFU

Three forces are killing TOFU content traffic simultaneously:

1. AI Overviews answer informational queries without a click.

We documented this in the graveyard keyword research. When an AI Overview appears, CTR on informational queries drops to under 2% regardless of ranking position. The full analysis is here.

2. Users are starting informational research in AI chat instead of Google.

A growing share of "what is X" and "how does X work" queries now originate in ChatGPT, Perplexity, and Claude rather than Google. Those queries produce no GSC impressions. They happen in the dark. This is the AI dark matter problem.

3. Thin informational content is being actively deprioritized.

Google's helpful content system updates have consistently reduced rankings for content that exists primarily to rank rather than to genuinely help. If your TOFU content is essentially a rewrite of what's already ranking — which describes most TOFU content produced at scale — it's being pushed down.

All three forces act on TOFU simultaneously. That's why the traffic drop is as large as it is.

---

What's Replacing TOFU as the Funnel Entry Point

If TOFU content is no longer reliably bringing people into the funnel, what is?

Based on what's actually driving pipeline in our client data, three content types are doing the work that TOFU used to do.

1. Original Research and Proprietary Data

Content that contains data nobody else has — because you collected it yourself — cannot be summarized by an AI tool without citing you. It also cannot be ranked around by competitors who don't have access to the same data.

What it looks like:

• "We surveyed 400 small business owners about how they're using AI tools in 2026. Here's what we found."
• "We tracked 3,900 AI search queries for 90 days. The citation patterns tell a different story than the conventional SEO wisdom."
• "We analyzed 180 AI responses to the same query. 6 different answers. Here's the breakdown."

This is the content that builds awareness in 2026 the way informational content did in 2020. When you publish original data, you give journalists, practitioners, and AI tools something to cite. The citations drive awareness without requiring the user to start on Google.

Pipeline mechanism: Original research gets shared, linked, and cited — driving awareness through distribution rather than through search ranking.

2. Case Studies With Real Numbers

Generic case studies ("we helped a client improve their results") produce nothing. Specific case studies ("we deleted 60% of a client's blog posts and organic traffic increased 23% in 90 days") circulate.

The specificity is not just for credibility. It's because AI tools can answer the generic version ("how do content audits help SEO?") but they can only point to you for the specific version ("what happened when Digital Ingenuity deleted 60% of a client's blog posts?").

Specificity creates content that AI tools must cite you for — because the data only exists on your site.

Pipeline mechanism: Case studies attract people at the consideration stage. They're not asking "what is content marketing" — they're asking "does this approach work?" Case studies answer that question at exactly the right moment.

3. Contrarian Takes With Evidence

"The conventional wisdom about [topic] is wrong, and here's the data that proves it."

This format drives pipeline because it attracts the decision-stage buyer who's already done their research and is looking for a more sophisticated perspective. They're not asking "what is featured snippet optimization" — they're asking "is our featured snippet strategy actually working?" And a post titled "Position Zero Doesn't Mean What It Used To" answers that exact question.

Contrarian content also gets shared more than confirmatory content. People don't share "Featured Snippets Are Great." They share "We Won 23 Featured Snippets and CTR Dropped on 19 of Them."

Pipeline mechanism: Contrarian evidence-backed takes attract practitioners who are further in their thinking — the ones who've already moved past basics and are looking for what the data actually shows.

---

The New Funnel

The funnel hasn't disappeared. It's compressed and reoriented.

```

Old funnel (2018–2022):

TOFU content → Google search → awareness → newsletter → nurture → conversion

New funnel (2025–2026):

AI/social → brand awareness → direct search or direct navigation → MOFU content → conversion

Entry point has moved. The middle is similar. The close is faster. ```

The practical implication: your content program should produce less generic TOFU and more:

• Original research (new awareness engine)
• Specific case studies (new MOFU)
• Contrarian evidence posts (new MOFU + late-stage trust builder)

The "consideration" and "decision" stages of the funnel are more important than ever. The buyer still exists. They're just entering the funnel later and expecting more when they arrive.

---

What to Do With Your Existing TOFU Content

This is the question most teams ask first. You have 200 posts. What happens to them?

The answer depends on the post.

Keep and upgrade: Any post with original data, specific examples, or a unique angle that can't be easily replicated by AI. Add more data. Deepen the specificity. These can survive.

Consolidate: Multiple thin posts on the same topic should be merged into one comprehensive resource. One 3,000-word post on a topic outperforms five 600-word posts in both Google and AI citations. We ran this experiment and documented the results.

Delete: Posts that are generic, thin, outdated, or that exist purely to capture a keyword and have no unique substance. Deleting low-quality content has been shown to improve the overall quality signals of a site. Our 60% deletion experiment resulted in a 23% traffic increase.

Redirect: For any deleted post that has inbound links, redirect to the closest relevant live page.

---

The Measurement Shift

If TOFU content is no longer your primary pipeline driver, your reporting metrics need to catch up.

Stop reporting:

• Organic sessions from TOFU keywords as a lead indicator of pipeline health
• Total impressions across all queries (impressions for AI-answered queries are noise)

Start reporting:

• Organic sessions by funnel stage separately (TOFU, MOFU, BOFU)
• Leads and trials attributed to organic MOFU content specifically
• Brand search volume trend (proxy for awareness being built through non-click channels)
• Direct traffic trend alongside branded search (early signal for AI dark matter brand lift)

These metrics tell you whether your content is actually building pipeline — not just generating sessions.

---

Related Experiments

Graveyard Keywords: High Impressions, Dying CTR — the data behind TOFU traffic collapse. 847 queries with growing impressions and dying clicks.

We Deleted 60% of Blog Posts and Traffic Increased — the content audit case study that shows what happens when you cut TOFU content and let quality content breathe.

AI Search Optimization: Strategic Moves — the strategic framework that puts these experiments in context.

90 Days Tracking AI Citation Volatility — how AI tools are citing content differently than Google ranks it, and what that means for what you create.

---

Key Takeaways

• TOFU content traffic is down an average of 51% across 6 client sites measured over 18 months. The direction is consistent regardless of industry.
• Three forces are acting simultaneously: AI Overviews answering informational queries without clicks, users starting informational research in AI chat instead of Google, and Google deprioritizing thin informational content.
• The funnel still works. It starts at MOFU now. Original research, specific case studies, and contrarian evidence posts are the new awareness and consideration drivers.
• What to do with existing TOFU content: upgrade what has unique substance, consolidate thin posts on the same topic, delete posts that are generic and outdated.
• Reporting metrics need to change. Total organic sessions is no longer a useful pipeline indicator. Sessions by funnel stage, MOFU-attributed leads, and brand search volume trends are.

---

We're publishing a follow-up next month with the content audit methodology we use to classify and prioritize existing posts. That post will include the exact scoring rubric and the decision tree for upgrade vs consolidate vs delete. It will be based on having run this process on 8 client content libraries over 12 months.

If your content program is still primarily TOFU-driven and you want to understand the scope of the issue before making any changes — run the funnel-stage breakdown on your GSC data first. The numbers will tell you whether this is a minor adjustment or a fundamental reorientation.