Tutoring Marketplace Development: Costs, Features, and What Actually Ships

You built a subject tutoring service. You have 40 tutors, a growing list of families paying for sessions, and a Notion doc to coordinate scheduling. Now the cracks are showing: double bookings, payment disputes, tutors not showing up, and parents calling you directly when something goes wrong.

You looked at Tutor.com and Wyzant. They are not white-label products. You cannot run your brand on them. Preply is built for language learning, not STEM or test prep niches. You are back where you started, except now you know the answer is custom tutoring marketplace software.

Here is what that costs, what it includes, and where these projects go wrong.

Cost table: what you're actually buying

Monthly running costs after launch: $1,200-$3,500 (hosting, Stripe fees, video infrastructure, search indexing, background check API calls).

The online tutoring market was valued at $8.7 billion in 2023 and is projected to reach $18.7 billion by 2030, growing at roughly 11.5% annually. The economics work for operators who own the platform. They do not work for operators paying 25% commission to someone else's marketplace.

Tutor.com, Wyzant, and Preply: where they stop being enough

These three platforms serve most of the general tutoring market well. The problem is none of them are software you can license.

Tutor.com is a managed service. They own the tutor supply, the pricing, and the student relationship. You cannot embed it in your product or white-label it for your school district clients.

Wyzant is a consumer marketplace. Tutors pay 25% commission to list their services. Students search a shared directory. If you want to compete in the same space under your own brand, you are paying Wyzant to build their audience while they take a quarter of every session.

Preply is optimized for language learning. Their matching, progress tracking, and session structure are built around language acquisition, not subject tutoring, test prep, or specialist niches like AP Physics or coding for kids.

Custom tutoring marketplace software makes sense when:

• Your commission model is below 25% and you need to attract tutors away from Wyzant by giving them better economics

• You are targeting a specific subject niche (SAT prep, STEM K-8, coding, music) where general platforms have weak tutor supply

• You need credential verification specific to your niche (teaching licenses, subject certifications, DBS checks for UK-based services)

• Your matching logic is more specialized than "pick a subject and see who's available"

• You have B2B clients (school districts, corporate training programs) who need white-label deployment or API access

• You are in a geography where English-language platforms have thin supply - Australia, Canada, Gulf markets

"In two-sided marketplaces, the matching algorithm is the product. The platforms that win surface the right fit, not just any fit." - Andrei Hagiu, Professor at Boston University Questrom School of Business, from his published research on platform competition and marketplace design.

Do not build a general tutoring platform to compete with Wyzant across every subject and grade level. That is a $300K investment to enter a market where a $100M competitor has supply, brand, and SEO you cannot replicate in year one. Build in a niche where you can own the supply side.

Who actually builds custom tutoring software

Four types of operators commission custom tutoring marketplace development. Each has a different reason.

The niche subject specialist. You have 60 certified AP Chemistry tutors and a waitlist of 200 students. Right now you coordinate via email. You want a platform where students can browse verified tutor profiles, book sessions, pay online, and track their progress across 10-week programs. A general marketplace doesn't fit because your credential requirements (college-level subject certifications) and your session format (problem sets, not open-ended tutoring) need custom matching logic.

The school district tech team. Your district serves 12,000 students across 8 schools. You want a platform where teachers can refer struggling students to verified tutors, parents can book sessions through a district-approved portal, and administrators can track usage and outcomes. Wyzant doesn't offer B2B contracts. Tutor.com doesn't give you data. You need custom software you own.

The international EdTech founder. You are building a tutoring marketplace for students in Southeast Asia being matched with North American tutors for English-language STEM support. Existing platforms have weak supply in your target geography and don't support your currency or payout structure. You need custom infrastructure.

The operator adding tutoring to an existing product. You run a learning management system (LMS) for homeschool families. You want to add a tutoring marketplace as a feature, not a separate product. Custom development lets you embed the booking and session experience inside your existing platform so students don't leave to find a tutor.

According to a 2024 survey by HolonIQ, 63% of EdTech operators who scaled past $5M ARR built proprietary matching or scheduling infrastructure rather than relying on third-party tools. The flexibility compounds as you grow.

V1, V2, V3: what to build in each phase

V1: the booking loop ($80K-$180K, 14-20 weeks)

V1 proves that students can find a tutor, book a session, pay, and show up. Everything else is iteration.

What ships in V1:

• Tutor profiles: subjects, grade levels, hourly rate, bio, availability calendar, sample session video (optional)

• Student search: filter by subject, grade level, availability, price range

• Basic matching: filter-based, sorted by review score (no weighted ranking yet)

• Booking engine: student picks a slot, system checks availability, creates booking record

• Stripe Connect payments: student pays platform, platform holds, releases to tutor after session

• Embedded video: Daily.co rooms provisioned on booking confirmation, join button active 15 min before session

• Post-session review: star rating and text from student, session notes from tutor

• Admin panel: tutor approval queue, booking overview, payout history

What does not ship in V1: background checks, credential document verification, whiteboard, progress reports, mobile apps, weighted ranking algorithm. Ship V1 with 20-30 tutors in a single niche. Prove the booking loop before building verification infrastructure.

V2: verification and retention ($60K-$90K added cost, 10-14 weeks)

V2 makes the platform trustworthy and adds the retention mechanics that keep students booking.

What ships in V2:

• Background check integration via Checkr API: tutor completes profile, Checkr runs report (1-3 business days), webhook updates verification status, unverified tutors hidden from search

• Credential document upload: tutors upload degree scans or certifications to S3, staff review queue in admin panel, manual approval flips credential status

• Weighted tutor ranking: composite score across subject expertise depth, availability overlap, price fit, review score (Bayesian average), and response rate - stored as precomputed column, updated nightly

• Session notes structured format: topic covered, homework assigned, next session focus - parent-facing progress timeline builds automatically

• Parent portal: view upcoming sessions, session notes, payment history, re-book with same tutor

• Automated reminders: 24-hour and 1-hour pre-session emails via SendGrid, post-session note prompts to tutor

Cost breakdown for V2: Checkr background checks run $30-$60 per tutor. Credential verification staff cost depends on your review volume. Budget one staff hour per 5-10 applications at launch.

V3: scale and mobile ($120K-$200K added cost, 16-20 weeks)

V3 adds the infrastructure that matters when you have 500+ active tutors and 2,000+ monthly sessions.

What ships in V3:

• iOS and Android apps: React Native codebase shared with web API

• Collaborative whiteboard: Tldraw embedded alongside Daily.co video, shared by booking ID

• AI-assisted matching: session completion data and re-booking rates feed a model that weights signals differently per subject category

• B2B portal: school district or corporate client accounts, bulk booking, reporting dashboards, invoicing

• Automated re-booking: after session notes are submitted, system suggests next available slot with same tutor and sends parent a one-click booking link

• Advanced analytics: tutor utilization rates, subject demand by geography, conversion funnel from search to first booking

Where tutoring marketplace projects fail

Two failure modes end more projects than technical complexity.

Launching without supply. You spend $120K building the platform, launch with a clean onboarding flow, and have 3 tutors available in your target niche. Students search, find nothing that fits, and leave. The platform is technically sound but commercially empty. Fix: spend 60 days before launch recruiting and onboarding tutors before students arrive. Run the first 20 sessions manually if needed to prove demand before automating the flow.

Ranking by recency instead of fit. The default sort for most MVP marketplaces is "most recent review" or "joined recently." This surfaces new tutors with 1-2 reviews above experienced tutors with 50 sessions completed, just because they joined last week. Students book based on ranking position. Poor ranking quality produces poor first sessions. Poor first sessions produce poor reviews. The marketplace degrades from the inside. Fix: implement the weighted scoring function in V2, not V3. It is the core retention mechanic on the supply side.

A 2023 study by Andreessen Horowitz on marketplace liquidity found that two-sided marketplaces with poor matching quality had 3x the churn rate of those with high-quality ranking, regardless of supply volume. You can have 1,000 tutors and still fail if the matching is weak.

How RaftLabs builds tutoring marketplace software

We scope tutoring marketplace projects in three conversations: one to understand the operator's niche and commission model, one to map the matching logic and credential requirements, and one to agree on V1 scope and what gets deferred.

The projects that go well share three traits. The operator knows their niche. They have tutor supply relationships before development starts. They treat V1 as a proof of concept, not a finished product.

The data model we use starts with four core entities: tutors, students, bookings, and sessions. Relationships between them carry the matching logic. The Stripe Connect three-party escrow pattern handles all payment routing. Daily.co handles video without custom WebRTC infrastructure. Checkr handles background checks via API. Algolia handles subject text matching at query time; our ranking function runs on the filtered result set.

We build in Node.js with PostgreSQL. Redis handles session locks and availability caching. React on web, React Native on mobile (one shared API). We host on Vercel (frontend) and AWS (API + database).

If you are scoping tutoring marketplace development, we can walk you through the data model, cost breakdown, and V1 feature list in a 30-minute call. Request that call here.

If you already have tutors onboarded and need to move quickly, tell us your niche, your commission target, and your planned tutor count at launch. We will tell you what V1 needs to include to make that viable.