Module 2: AI-Powered Niche Research

What changed about niche selection post-HCU

Before Google's Helpful Content Update series (2022–2024), the niche selection strategy was simple: find a keyword cluster with search volume, low domain authority sites in top 10, and at least one affiliate program. Write 80 articles. Wait.

That still works in some niches. But the failure rate has increased substantially because Google is now better at identifying and demoting "written for search engines, not people" content — and entire sites, not just individual pages, get classified this way.

The two biggest changes you need to account for:

1. YMYL adjacency: Niches adjacent to Your Money or Your Life (health, finance, legal, safety) now require significantly more E-E-A-T signals to rank. A new site with no author credentials competing in supplement advice or stock picks is essentially starting from a disadvantage that content quality alone won't fix.
2. Saturated "best X" niches: Niches where every query returns roundup posts from high-DA sites (Forbes, NerdWallet, Wirecutter variants) are no longer accessible to new sites without a genuinely differentiated angle. The 2019 technique of "write better content than the current results" doesn't work when the current results have 10,000 backlinks.

The winning move in 2026 is to find niches where:

• The top results have low-to-moderate domain authority (DA 15–40)
• There's software or SaaS with affiliate programs (higher commissions, recurring revenue)
• There are "jobs to be done" queries — how-to, comparison, best-for — not just informational
• The niche is specific enough that you can become genuinely authoritative on it

Step 1: AI shortlisting — the prompt

I use this prompt with Claude (Sonnet or above) as the first pass. It's designed to return actionable angles, not generic category names.

The goal of this prompt is to get specific angles, not broad categories. "Productivity software" is a broad category. "Task management apps for remote engineering teams" is a specific angle with identifiable buyer intent and real products to promote.

Run this prompt for 2–3 different background/interest combinations. You'll get 20–30 candidates. From these, select your top 5 based on your gut interest and the AI's competition assessment. Then validate them properly.

Step 2: Keyword validation with free tools

AI shortlisting is fast but not data-backed. You need real keyword data before committing to a niche. Here's the full free-tools validation sequence:

Tool 1: Google Keyword Planner (free with a Google Ads account)

Enter your 3 candidate keywords from the AI output. Look for:

• Average monthly searches: ideally 500–5,000 for core "best X" keywords. Too high (100k+) and you're in saturated territory. Too low (under 100) and the niche may not have enough volume to sustain a site.
• Competition level: "Low" or "Medium" in the paid column is a proxy for organic competition. "High" indicates a lot of advertisers — can mean good commercial intent but also saturated organic results.

Tool 2: Ahrefs Webmaster Tools (free with domain verification)

Sign up for AWT and verify your domain (you can verify a new blank domain just to get access). Ahrefs' free Keywords Explorer allows a limited number of searches — use them on your top 3 niche candidates.

The key metric: Keyword Difficulty (KD). For new sites:

• KD 0–20: Realistic target for new sites
• KD 21–40: Possible but will take 6+ months of consistent content
• KD 41+: Requires either domain authority or significant link building investment — avoid for new sites

Tool 3: Google Search itself

Search your top candidate keywords in an incognito window. Specifically look at:

• Who's ranking? If positions 1–5 are Forbes, NerdWallet, The Wirecutter, and Healthline — don't enter this niche without a serious differentiation angle.
• AI Overview present? If Google shows an AI Overview for the query, your organic result is pushed further down. This doesn't make the keyword worthless, but factor it in.
• Age of content: Click into one of the ranking pages. Check when it was published. If the top results are 4–6 years old and not recently updated, that's a signal that the incumbents aren't actively defending the niche.

Step 3: Affiliate program availability check

A niche with no real affiliate programs forces you into Amazon Associates (3–8% commission, 24-hour cookie) as your primary monetization. That's a fragile business. Aim for at least 3 affiliate programs in your niche that pay 20%+ recurring commission or $50+ per conversion.

The fastest research method:

1. Search "[niche keyword] affiliate program" and "[product name] partner program"
2. Check PartnerStack.com — it aggregates B2B SaaS affiliate programs in one place
3. Check Impact.com and ShareASale for consumer software and services
4. Look at competitor affiliate sites in the niche: examine their footer for "affiliate" or "partner" links, and check which tools they link to repeatedly in their content

Step 4: Competition audit

Take your top 2 niche finalists and run a quick competition audit:

1. Search the primary "best [niche]" keyword (e.g., "best project management software for small teams")
2. Click into the top 3 results
3. Paste each URL into Ahrefs' free backlink checker or SimilarWeb's free tier
4. Note the domain's approximate authority and backlink count

If the top 3 results are DA 50+ sites with thousands of backlinks, you need a more specific angle — a sub-niche where those sites aren't as dominant. If you're seeing DA 15–35 sites with a few hundred backlinks, you have a realistic path to compete.

The niche scoring framework

Score each niche candidate across these 5 dimensions (1–3 points each, max 15):

Target niches scoring 11+ out of 15. Below 8, move on. Between 8–10, look for a tighter sub-angle that might score higher.

Niche red flags that kill new sites

Some niches look attractive on paper but have structural problems that prevent new sites from ever getting traction:

• VPN / web hosting / best WordPress theme: These are dominated by authoritative sites with massive link budgets and often self-serving reviews. The top 10 results for most queries in these spaces are established affiliate media companies. Almost impossible to crack without a very niche angle.
• "Best credit cards" / loan comparison: Full YMYL, requires extensive E-E-A-T, FTC compliance considerations, and you're competing against NerdWallet and Bankrate with multi-million dollar SEO budgets.
• Weight loss / supplements: YMYL health content with serious ranking challenges for non-credentialed sites post-HCU. Any niche where people's physical health is the primary topic.
• Make money online / MLM reviews: Ironic, but "make money online" is one of the most saturated affiliate niches. Every guru product review, MLM review, and "how to make money from home" site is competing here.
• Tech news: High volume but no buyer intent and constantly refreshing — you'd need to publish daily to stay relevant. Not an affiliate model, it's a media company model.

Worked examples: three niche evaluations

Example 1: AI writing tools for e-commerce

Score: 13/15. Good buyer intent (Shopify sellers actively evaluating tools), multiple SaaS affiliate programs (Jasper, Copy.ai, Writesonic all have programs), moderate competition (DA 25–45 sites ranking for most terms), and software like product description generators have clear commercial queries. Main limitation: space is moving fast and tools are commoditizing, which can hurt content freshness.

Example 2: Project management software for construction

Score: 14/15. Strong vertical-specific angle that national "best PM software" roundups don't serve well. Tools like Procore, Buildertrend, and CoConstruct have affiliate or referral programs. The audience (construction project managers) has high purchase intent and the software is high-ticket. Competition is mostly generic business software review sites not tailored to construction. This is a classic "competitor ignores the vertical" opportunity.

Example 3: Podcast editing software

Score: 9/15. Decent buyer intent and some affiliate programs (Descript, Audacity plugins). Low average KD. But: the niche is fairly small (limited total search volume), most keywords have been picked up by established podcasting sites like Buzzsprout and Podbean who cover tools as part of their broader authority. You'd need a very specific angle — "podcast editing software for non-technical creators" or similar — to carve out a defensible position.

Module 2 action steps

1. Run the AI shortlisting prompt for your background/interests
2. Select 5 candidate niches from the AI output
3. Validate each with Google Keyword Planner and Ahrefs Webmaster Tools
4. Check affiliate program availability on PartnerStack and Impact
5. Score each niche using the framework above
6. Pick your winner — the one scoring 11+ with the highest personal interest