# How to Build a Zhihu Topic Portfolio for Scaled AI Publishing

> Avoid chasing viral threads alone. A 2:5:3 mix of defensive, search-intent, and pain-point topics balances instant reach with evergreen search traffic.

Source: https://www.starboxai.cn/en/insights/ai-marketing/2026-11-04-zhihu-puliang-xuanti-strategy/ (Starbox AI Insights, updated 2026-11-04)

## Viral Questions vs. Long-Tail Queries: How Distribution Works on Zhihu

Publishing 5 to 10 answers daily using generative AI often creates a common dilemma: chasing trending questions with tens of thousands of upvotes yields zero traction, while answering obscure threads produces no immediate page views. 

To resolve this issue, you must separate Zhihu (China's largest Q&A platform) into two distinct distribution mechanisms: the recommendation feed and search engine indexing.

The recommendation feed rewards high early engagement velocity, including upvotes, saves, and comments within the first few hours of publication. Viral questions already have established answers with thousands of upvotes. A fresh AI-assisted answer posted to a high-traffic thread enters at the bottom of the stack, where real users rarely scroll.

Search distribution follows a different logic. Zhihu holds significant domain authority across Chinese search engines as well as generative AI platforms like DeepSeek and Kimi. These platforms index structured answers to specific, problem-oriented queries. When users search for operational solutions or software comparisons, Zhihu threads frequently rank near the top. 

Chasing viral threads alone wastes content volume on low-probability impressions. Winning search visibility requires targeting specific user intent where existing answers are outdated, incomplete, or absent.

## The 2:5:3 Topic Portfolio Strategy

Instead of relying solely on viral topics, structure daily output using a 2:5:3 portfolio model: 20 percent defensive questions, 50 percent long-tail search questions, and 30 percent scenario pain points.

| Question Type | Target Share | Traffic Source | Primary Objective | Selection Criteria |
| :--- | :--- | :--- | :--- | :--- |
| **Defensive High-Attention** | 20% | Platform Feed & Browse | Maintain brand presence on primary category terms | Over 1,000 followers on the question; active discussions within the past 30 days. |
| **Search Long-Tail** | 50% | In-platform Search & External Search | Capture durable search traffic and search engine citations | Clear intent modifiers (e.g., "how to", "best tool for", "vs"); low answer count. |
| **Scenario Pain Points** | 30% | Search & Related Question Feeds | Build authority in underserved niche use cases | High search intent, zero high-quality answers, specific operational bottlenecks. |

### 1. Defensive High-Attention Questions (20%)
These are industry-defining questions where your brand or viewpoint must be present. You are not expecting these answers to reach the top position overnight. Instead, they serve as defensive brand anchors:
- Questions covering broad category definitions or major platform trends.
- Topics where potential enterprise buyers research standard market options.
- Threads where competitors have already staked out positioning.

### 2. Search Long-Tail Questions (50%)
Long-tail queries represent the core engine of your daily publishing volume. These questions match precise search queries typed by users looking for specific solutions:
- Comparison queries (e.g., comparing two specific workflows or technical stacks).
- Implementation and how-to inquiries with narrow operational scopes.
- Questions that incorporate modifier keywords such as industry names, budget limits, or role designations.

Because long-tail questions usually feature fewer than 20 competing answers, a well-structured, fact-dense AI answer can secure a top-three ranking quickly and maintain that visibility for months.

### 3. Scenario Pain Points (30%)
Pain-point questions address emerging operational friction that broad industry questions overlook. These are blue-ocean topics where user interest exists, but high-quality answers are missing:
- Frustrations with existing tools or software workflows.
- Questions addressing recent regulatory changes, platform rule updates, or new operational standards.
- Detailed tactical problems faced by specific job titles.

Producing content for these questions establishes early ownership over niche topics before competitors notice the search volume.

## Daily Step-by-Step Topic Screening Checklist

To produce 5 to 10 quality answers per day without bottlenecking on topic discovery, follow this three-stage screening checklist every morning.

```
[ ] Step 1: Harvest Core Search Queries
    - Enter category seed keywords into Zhihu's search bar.
    - Export autocomplete suggestions and "People Also Search" queries.
    - Check external search suggestions for matching phrase patterns.

[ ] Step 2: Filter by Competition and Recency
    - Eliminate questions with over 100 answers unless user activity has surged in the last 48 hours.
    - Filter for questions with at least 5 to 10 followers to verify baseline interest.
    - Prioritize questions where the top answer is over 12 months old or under 300 words.

[ ] Step 3: Match Daily Allocation Ratios (Target: 10 Answers)
    - Select 2 Defensive Questions (Broad industry scope, high followers).
    - Select 5 Search Long-Tail Questions (Specific operational terms, low answer counts).
    - Select 3 Scenario Pain-Point Questions (Direct practical problems, weak existing content).
```

## How to Validate Your Topic Mix: 3 Core Metrics

Tracking total views alone will mislead you, because long-tail search traffic accumulates gradually over weeks rather than spiking in hours. Evaluate topic performance using three functional metrics:

1. **Search Indexing Rate:** Measure what proportion of your answers appear on the first page of Zhihu search results for their exact title query within 14 days of publishing. A healthy rate is 40 percent or higher.
2. **Evergreen Reading Ratio:** Check the proportion of impressions generated 30 days after publication versus the first 48 hours. If more than 60 percent of views occur after day 30, your search interception strategy is working.
3. **Save-to-Reading Ratio:** In informative B2B topics, readers save content they intend to use as a reference. A save rate above 2 percent indicates that the answer solves an operational pain point rather than offering surface-level commentary.

### When to Pivot the Allocation Model

The 2:5:3 formula is an operational baseline, not a permanent rule. Adjust the balance when specific conditions emerge:

- **When to increase defensive share (to 40%):** If your account has accumulated high topic badges and domain authority, your answers can credibly compete in higher-traffic feeds. Shift capacity to higher-visibility industry questions.
- **When to increase long-tail share (to 70%):** When managing a newly registered account with zero platform authority, drop defensive questions almost entirely. Focus exclusively on unanswered long-tail searches to accumulate platform weight and initial crawl citations without algorithmic resistance.