Teachers can't easily prove a parent, sibling, or other human (not AI) wrote a student's essay — current tools focus on AI detection and web plagiarism, not authorship mismatch.
Builds a per-student writing profile from in-class assignments (vocabulary level, sentence structure, common errors). New submissions are scored against the profile; outliers are flagged with a confidence score and specific deviations highlighted.
Freemium SaaS — free for individual teachers (up to 30 students), paid per-school or per-district license ($3-5/student/year)
The pain is real and visceral — 1,132 upvotes on a teacher subreddit confirms broad resonance. Teachers KNOW parents ghost-write but can't prove it. However, this is a 'chronic annoyance' not an 'existential crisis' for most teachers. It rarely has career or legal consequences for the teacher, which limits urgency. The pain is higher for administrators dealing with grade disputes and academic integrity policies.
TAM is meaningful but bounded. US K-12: ~25M students in grades 6-12. At $3-5/student/year, US TAM is $75-125M. International could 2-3x this. However, realistic near-term SAM is much smaller — only English/writing teachers who assign essays regularly, and only districts willing to pay for yet another integrity tool. Serviceable market is likely $10-20M in first 3 years. Not venture-scale alone, but solid for a bootstrapped SaaS.
This is the weakest link. Individual teachers have near-zero budget authority and are already fatigued by tool overload. District purchasing is slow (6-18 month sales cycles) and budget-constrained. Turnitin already takes the 'integrity tool' budget line at most schools. Convincing administrators to pay for a second integrity tool — specifically for a problem they may not quantify — is a hard sell. The $3-5/student/year price is right but getting the PO signed is the challenge.
Stylometric analysis is well-researched (decades of NLP literature). Modern LLMs and embeddings make building a writing fingerprint engine significantly easier than 5 years ago. A solo dev can build a working MVP in 6-8 weeks: ingest baseline essays, extract features (vocabulary richness, sentence length distribution, syntactic patterns, error patterns), compute similarity scores, flag outliers. However, ACCURACY is the hard part — false positives (student improved!) and false negatives (subtle ghost-writing) require significant tuning. FERPA/COPPA compliance for K-12 data adds real engineering overhead. Cold-start problem (need 3-5 baseline samples per student) complicates onboarding.
This is the strongest dimension. Literally zero commercial products do per-student stylometric fingerprinting for ghost-writing detection. Every competitor focuses on AI detection or copy-paste plagiarism. Turnitin's authorship features require their proprietary writing environment and track process metadata, not writing style. The gap between 'teachers know this is a problem' and 'tools that solve it' is enormous. Blue ocean.
Natural annual subscription aligned with school years. Writing profiles improve with more data, creating genuine lock-in — switching costs increase over time as the fingerprint library grows. Per-student pricing scales with enrollment. District-level contracts are inherently multi-year. The product gets MORE valuable the longer a school uses it (longitudinal data), which is the best kind of retention moat.
- +Genuine blue-ocean gap — no product does this. You'd be first-mover in human ghost-writing detection.
- +Strong emotional resonance with teachers (proven by Reddit engagement). The problem is viscerally felt even if not formally measured.
- +Natural data moat — per-student profiles improve over time and create high switching costs.
- +Complementary positioning to Turnitin (not competitive), enabling co-existence or acquisition potential.
- +Recurring revenue aligned with school-year purchasing cycles.
- !Willingness-to-pay is uncertain — schools may see this as 'nice-to-have' not 'must-have,' especially when budgets are tight and Turnitin already takes the integrity line item.
- !False positive sensitivity is extreme — flagging a student whose writing genuinely improved (or who got legitimate tutoring) could create parent backlash, lawsuits, or PR disasters for the school.
- !Turnitin could ship a 'writing consistency' feature within 12 months if the category proves viable — they have the data, the distribution, and the brand.
- !FERPA/COPPA compliance for storing detailed writing profiles of K-12 minors is non-trivial and adds legal/engineering cost.
- !Cold-start problem: tool is useless until you have 3-5 baseline samples per student, meaning teachers won't see value for weeks after adoption.
Tracks document creation metadata
AI content detection and plagiarism checker. Markets 'writing style analysis' but this is AI-vs-human classification, not author-vs-author verification.
AI content detection tool that shows stylistic metrics and 'writing analysis' breakdowns per document. Popular with individual teachers.
Open-source academic stylometric analysis tool from Duquesne University. True authorship attribution using function words, character n-grams, and other linguistic features. Used in forensic linguistics research.
Identity-verification proctoring tools that use video, biometrics, and lockdown browsers to ensure the right person is taking an exam.
Web app where a teacher creates a class, uploads 3-5 in-class writing samples per student (or pastes text), and the system builds a writing fingerprint. New submissions are scored 0-100 for consistency with the student's profile. Dashboard shows flagged submissions with specific deviations highlighted (e.g., 'vocabulary complexity jumped 3 grade levels,' 'average sentence length doubled,' 'zero spelling errors vs. baseline average of 8 per essay'). No LMS integration in MVP — just copy-paste or file upload. Free for 1 class of up to 30 students.
Free tier (1 class, 30 students) to prove value with individual teachers → paid teacher plan ($8/month for unlimited classes) → school license ($3/student/year, admin dashboard, bulk import) → district license ($2/student/year at volume, SSO, LMS integration, API access) → potential Turnitin/PowerSchool acquisition target once you prove the category.
8-12 weeks to MVP and first free users. 4-6 months to first paying individual teachers. 9-15 months to first school/district contract (due to purchasing cycles — you need to catch back-to-school budget planning in spring for fall deployment). Meaningful revenue ($5K+ MRR) likely 12-18 months out.
- “student admitted the work wasn't their own”
- “teacher had to manually recognize the writing didn't match the student's level”
- “parent wrote essay but couldn't even address the prompt correctly”
- “bragging about how his mom writes his essays”