Deep Niche Research — Round 3 + Methodology Refinements
Date: 2026-04-26 Source: Semrush API Guru tier API units used cumulative: ~6,100 / 50,000 Confidence: HIGH — every number below pulled live, KD scores confirmed for all top picks.
TL;DR — biggest new findings
- A new #1 winner: Alpaca Farms Atlas — KD 18, 8K volume, top SERP is single farms + tourism boards. Same playbook as UPick. Easier than Waterfalls.
- Real game-changer: Vegan Restaurants — KD 14 at 49.5K monthly volume. Top 10 SERP is mom-blogs + Reddit + Facebook. Massive gap.
- Personal Injury Lawyer is the highest-CPC niche I found ($126 CPC) but KD 46 + heavy ad spend by big firms → not a solo-operator play.
- My methodology had 4 holes I missed. New refinements at the bottom.
All untested niches, scored
NEW #1 candidate: Alpaca Farms Atlas ⭐⭐⭐
| Metric | Value |
|---|---|
| "alpaca farm near me" volume | 8,100/mo |
| KD | 18 ✅ |
| Competition | 0.04 |
| Top 10 SERP | Single farms + Facebook + Yelp + tourism boards |
| Inventory | ~250 alpaca farms in US (Alpaca Owners Association membership data is public) |
| Schema | LocalBusiness + TouristAttraction |
| Affiliate | Alpaca yarn/fiber (Etsy, AmazonAffiliate) + farm-stay booking |
Why this is exciting:
- KD 18 is the lowest "real volume" score in our entire research universe except
vegan restaurants near me - Top 10 SERP has ZERO national directory — literally just individual farms + a Facebook page + Yelp
- Inventory is small (~250) which means fully coverable in one weekend = full defensibility
- Same template as UPick Atlas (Next.js + ShadCN + Cloudflare Pages)
- Alpaca Owners Association (alpacainfo.com/farms) lists every member farm — free seed data
Score: 36/40 ⭐ — same as UPick.
Vegan Restaurants Atlas ⭐⭐⭐
| Metric | Value |
|---|---|
| "vegan restaurants near me" volume | 49,500/mo |
| KD | 14 ✅ (lowest I found anywhere with this much volume) |
| Top 10 SERP | wayfaringvegan.com, thevgnway.com, vegananj.com, theconscientiouseater.com, Reddit, Facebook — all mom blogs & community posts |
| HappyCow as defender | medium — they have apps + global focus, web SEO mediocre |
| Inventory | ~12,000 vegan/vegan-friendly US restaurants (HappyCow scrape + Yelp filters + manual) |
Why this is the sleeper of the round:
- KD 14 + 49.5K monthly volume = probably the best volume-to-KD ratio I've seen
- Top SERP is literally bloggers — the gap is screaming for a structured directory
- HappyCow exists but their web SEO is surprisingly weak (didn't appear top-3 for "vegan restaurants near me")
- Vegan/plant-based market is growing 10%+ annually
- Affiliate path: cookbooks, vegan products on Amazon, food delivery affiliates
Risk:
- Yelp local-pack often takes top of page for "near me" food queries — could cap traffic
- HappyCow could improve their SEO if they notice
Score: 35/40
Lavender Farm Directory
| Metric | Value |
|---|---|
| Volume | 8,100/mo |
| KD | 41 ⚠️ (higher than I'd hoped) |
| SERP | lavender-life.com (national directory exists!) + lavendertrails.com + thelavenderfarms.com + individual farms |
Verdict: Already has 3 incumbent directories. Not as open as alpaca farms. Score: 22/40. Skip — but lavender farms could be added as a crop to UPick Atlas.
Petting Zoo Directory
| Metric | Value |
|---|---|
| Volume | 49,500/mo |
| KD | 50 (high) |
| SERP | Mostly individual zoos + Facebook |
KD 50 is too high for a new site to crack. Score: 24/40.
Cat Cafe Directory
| Metric | Value |
|---|---|
| Volume | 49,500/mo |
| KD | 58 (high) |
| SERP | Individual cafes |
Volume is real but KD is rough. Smaller market (~150 US cat cafes). Score: 22/40.
IV Therapy Directory
| Metric | Value |
|---|---|
| Volume | 49,500/mo |
| KD | 25 ✅ |
| Top SERP | Single clinics (primeivhydration.com, livehydrationspa.com, etc.) |
| CPC | $0 reported but related lead-gen pays $50-150/lead |
Why it's interesting:
- Real KD 25 + 49.5K volume + clinic lead-gen affiliate = potential goldmine
- Top SERP is just individual clinics — no directory
- IV therapy is regulatory grey-zone but legal
Why I hesitate:
- Health/wellness is a YMYL (Your Money or Your Life) niche where Google holds quality bar high
- E-E-A-T (Experience, Expertise, Authority, Trust) requirements are severe — programmatic LLM content risks getting flagged hard
- Clinic data is hard to verify (licensing, legitimacy)
Score: 28/40. Doable but riskier than alpaca farms.
Halal Restaurants Directory
| Metric | Value |
|---|---|
| Volume | 40,500/mo |
| KD | 29 ✅ |
| Top SERP | zabihah.com (sole defender) + Grubhub + Yelp + Facebook |
Why interesting:
- KD 29 with 40K volume is rare
- Zabihah.com is the dominant directory but its UX is dated; high opportunity to displace
Why I hesitate:
- Audience is concentrated in major metros (NYC, Chicago, Detroit, LA, Houston) — geographic skew limits state-level long-tail
- Religious niches can have community-trust dynamics that pure SEO can't easily win
Score: 30/40. Worth more validation but not top-tier.
Personal Injury Lawyers (the high-CPC trap)
| Metric | Value |
|---|---|
| Volume | 74,000/mo |
| CPC | $126.74 ⚠️ |
| KD | 46 (high) |
| Top SERP | usa.gov + justia.com + forthepeople.com (Morgan & Morgan) + legalmatch.com |
Why it looks tempting:
- CPC of $126 means every click is potentially worth real money via Google AdSense
Why it's a trap for a new site:
- KD 46 = top SERP defenders have huge backlink profiles
- Lawyer firms spend MILLIONS on PPC — they have unlimited budget for SEO defenders
- usa.gov + justia.com are entrenched
- Lead-gen affiliate networks (Avvo, FindLaw) own most of the alternative paths
- Compliance (state bar regulations on legal advertising, attorney disclaimers) adds operational overhead
Score: 14/40. Skip — high CPC is a signal that big money already won this space.
Religious / Community Directories
churches near me = 246K volume! But here's the catch: Google's local pack dominates these queries. Maps results take top-of-fold; directories barely appear.
Score: 12/40. Don't even consider.
Seasonal Events
haunted houses near me (40K, KD 75), christmas lights near me (49K, KD 45), tulip festival near me (3.6K, KD not pulled).
Volumes are real but all heavily seasonal (3-6 weeks/year). Christmas lights is plausible at KD 45 but you'd be working 11 months for 4 weeks of traffic.
Christmas Lights Map could work as a side project but not a flagship. Score: 22/40.
Activity / Experience (axe throwing, escape rooms, paint and sip)
KD 55-59 across the board. Too high for new sites. Score: 18-22/40.
Educational / Kids (summer camps, swim lessons)
KD 46-53. Niche.com already covers schools, swim lessons SERPs are dominated by Goldfish Swim School and big-brand chains. Score: 20/40.
Updated full ranking (with everything from all 3 rounds)
| # | Niche | Score | Status |
|---|---|---|---|
| 1 | UPick Atlas | (built) | ✅ Already shipping, validated |
| 2 | Alpaca Farms Atlas | 36/40 | ⭐ NEW: lowest risk, fastest build |
| 3 | Waterfalls Atlas | 35/40 | Big volume, beatable, USGS data |
| 4 | Vegan Restaurants | 35/40 | KD 14 sleeper hit |
| 5 | Drive-In Theaters | 31/40 | Small but fully coverable |
| 6 | Halal Restaurants | 30/40 | Worth deeper validation |
| 7 | IV Therapy | 28/40 | YMYL risk |
| 8 | Disc Golf | 28/40 | UDisc defends |
| 9 | EV Charging | 26/40 | Government dominance |
| 10 | Scenic Drives | 26/40 | Pair with Waterfalls |
| 11 | Pickleball | 26/40 | Pickleheads defends |
| 12 | Self-Storage | 24/40 | $2.7M ad-spend defender |
| 13 | Petting Zoo | 24/40 | KD 50 too high |
| 14 | Cat Cafe | 22/40 | KD 58 |
| 15 | Lavender Farms | 22/40 | 3 incumbent directories |
| 16 | Christmas Lights | 22/40 | Seasonal limitation |
| 17 | Coworking | 22/40 | Smaller volume |
| 18 | Wineries/Breweries | 20/40 | Untappd defends |
| 19 | Educational/Kids | 20/40 | Goldfish Swim, Niche.com defend |
| 20 | Activity (axe/escape/paint) | 20/40 | KD 55+ |
| 21 | Dog Parks | 18/40 | BringFido fortress |
| 22 | Private Schools | 18/40 | Niche.com fortress |
| 23 | Boba | 16/40 | KD 69 |
| 24 | Personal Injury | 14/40 | Big-firm money fortress |
| 25 | Religious | 12/40 | Local pack dominance |
| 26 | Hiking Trails | 8/40 | AllTrails owns it |
Methodology refinements (4 holes I patched)
Hole 1: I was reading "Competition" instead of "KD"
Semrush's Competition (0.0-1.0) is paid-search competition — how many advertisers are bidding on this keyword. It does NOT measure organic difficulty.
Examples that diverged hard:
- "boba near me" — Comp 0.11 (looks easy), KD 69 (actually hard)
- "lavender farm near me" — Comp 0.04 (looks dead), KD 41 (moderate)
- "vegan restaurants near me" — Comp 0.16 (looks meh), KD 14 (very easy)
Refinement: Always pull both. If the gap between them is large (>30 points apart), the niche behaves differently than your first read. Trust KD for organic, Comp for ad-revenue potential.
Hole 2: I underweighted "near me" + local-pack reality
Many "near me" queries trigger Google Maps local pack at the top, which steals 60-80% of clicks before any organic result is seen. Categories where local pack dominates:
- Restaurants (any type)
- Lawyers / doctors / dentists
- Repair / service / contractors
- Religious centers (verified above)
- Storage facilities
- Gyms
- Salons / barbers / spas
Categories where local pack is weak or absent (organic still gets clicks):
- Outdoor recreation (waterfalls, scenic drives, alpaca farms — visit-once tourism)
- Specialty stores with low Yelp inventory (record stores, comic shops)
- Activity directories where the local pack hasn't fully indexed (escape rooms is mixed)
- Seasonal/event-based (haunted houses, Christmas lights — pack is weaker)
Refinement: Before scoring, do a manual SERP check on a real "near me" query. If the local pack has 3 results plus a "View all" link before any organic result appears, downgrade the score by 5-8 points.
Hole 3: I didn't account for YMYL + E-E-A-T
Google has explicit quality bar called "Your Money or Your Life" (YMYL) for niches affecting health, finances, legal, safety. AI-generated programmatic content gets flagged HARDER in YMYL niches.
Real list of YMYL niches (avoid programmatic SEO unless you have real expert credentials):
- Health / medical / supplements / wellness clinics (IV therapy, ketamine, TMS, etc.)
- Legal services
- Financial advice / mortgages / insurance
- Real estate
- Heavy appliances / safety products
- Childcare / education
The implication for our research: IV therapy at KD 25 looks tempting but the YMYL penalty makes it 2-3x harder than the KD score suggests.
Refinement: Apply a -8 point penalty to YMYL niches in scoring. Most won't be worth pursuing.
Hole 4: I underweighted seasonal compression
Some niches have huge head terms but are only searched 6-8 weeks/year. The annual revenue ceiling math has to account for the dead months — you can't earn $0 for 11 months and $X in month 12; ad networks won't approve you, affiliate accounts go dormant.
Highly seasonal niches:
- Pumpkin patches (Aug-Oct only)
- Christmas tree farms (Nov-Dec)
- Strawberry / cherry picking (4-6 week peak)
- Tulip festivals (3-4 weeks)
- Haunted houses (October only)
- Apple picking (Aug-Oct)
- Christmas lights tours (Nov-Dec)
Year-round / less seasonal:
- Alpaca farms (year-round farm tours possible)
- Waterfalls (peak spring runoff but 4-6 month strong window)
- Drive-in theaters (May-Sept primary)
- Dog parks (year-round)
- Vegan restaurants (year-round)
UPick Atlas point of vulnerability: The reframe we just did toward "pumpkin patches" weights the site toward 6-week annual peaks. We should also push the year-round terms (general "u-pick farms near me", crop-agnostic state pages) to keep traffic flowing in winter/early spring.
Refinement: Score column for "annualized search distribution evenness." If >60% of niche search volume happens in <12 weeks, downgrade by 4 points. UPick is borderline (fall pumpkin spike is huge but spring berry + summer peach + fall apple + winter Christmas trees give some balance).
What this means for next decision
Original recommendation (Round 2): Build Waterfalls Atlas
Revised recommendation (Round 3): Build Alpaca Farms Atlas
Reasoning:
- KD 18 (lowest of any candidate) = fastest path to ranking
- Inventory ~250 farms = fully coverable in 2 weekends, full defensibility
- Year-round traffic (no seasonal compression)
- Same template as UPick Atlas, can fork in <30 minutes
- AOA member directory provides free seed data
- "Alpaca farm visits" market is real — these farms actually want directory listings
Comparison:
| Alpaca Farms | Waterfalls | Vegan Restaurants | |
|---|---|---|---|
| Volume | 8K | 110K | 49.5K |
| KD | 18 | 21 | 14 |
| Inventory | 250 (full coverage) | 17,000 (incremental) | 12,000 (huge effort) |
| Build time | 1 weekend | 2-4 weeks | 4-8 weeks |
| Local pack risk | Low | None | HIGH (food queries dominated) |
| YMYL risk | None | None | None |
| Seasonal | Year-round | 4-6 month strong | Year-round |
| Defensibility | High (full coverage) | Medium (needs depth) | Low (HappyCow could pivot) |
| Revenue ceiling | $300-800/yr | $1.5-4K/yr | $1-3K/yr |
Alpaca Farms is the safest bet. Vegan Restaurants has higher upside but local-pack risk. Waterfalls is in between.
My actual single recommendation
Build Alpaca Farms Atlas as a 1-weekend test of the directory thesis.
Reasoning:
- We're spending 5+ hours of research time per niche. Cheaper to actually build the lowest-risk one and see if the SEO playbook works on a fresh domain in 60-90 days
- If alpaca site ranks → directory thesis is alive, then commit to Vegan Restaurants or Waterfalls (bigger upside)
- If alpaca site doesn't rank → directory thesis is dead, save $200-500 on dual-track infrastructure costs
It's the cheapest test of the most important question: does Omar's playbook still work in 2026 for a new operator?
Cost to test: 1 weekend + $10 domain + ~$5 Codex API. Worth it.
Where my analysis can still fail
- KD scores I pulled are Semrush's estimate; Ahrefs/Moz might disagree by 5-15 points
- Seasonal patterns can shift (pumpkin patch search peaked in October 2024, slightly later in 2025 due to weather)
- Local pack is changing — Google launched "Hotels-style filters" in late 2025 that affect food/service queries
- YMYL bar may relax for some niches (or tighten for more)
- Trump administration deregulation could affect federal data sources we depend on (USDA + USGS)
These are the asterisks I attach to every recommendation. Plan for being wrong 30% of the time.
API budget remaining
- Used cumulative: ~6,100 / 50,000 (12.2%)
- Plenty left for: deep dive on alpaca farm SERPs, Northstar Forge keyword analysis, competitor content gap analysis on PickYourOwn.org, pre-launch validation of any chosen niche