Gap report · TecPal

The verdictBenchmarked against 2026 LMS standards, EdTech security frameworks and retention research

TecPal's front-of-house is ahead of the market it competes in: transparent pricing, verifiable certificates, live-class recordings with offline and audio options, an assignments loop with AI review, community, gamification and a booking system are all things most Nigerian competitors do not have. The real gaps cluster in three places: retention machinery (the numbers say community and first-week engagement decide completion, and nothing yet watches for the learner who is quietly disappearing), financing (the single biggest enrolment barrier in this market is the upfront fee, and TecPal offers instalments but no financing partner), and operational proof (status page, incident runbook, SOC 2 when B2B deals arrive). Nothing found invalidates what is built. Twenty-one findings follow, each tagged with who can act on it.

Learning experiencecontent, delivery, access

Assessmentcredentials, integrity

Student successretention, completion

Securitycompliance, trust

Paymentsfinancing, access

Growthmarketing, sales

Operationsstaff tooling

Data & AIPal, analytics

Student success · the highest-stakes domainCourses with active community complete at 65.5% against 42.6% without; first-week engagement is the strongest single predictor

At-risk early-warning queueThe research is blunt: 30% of dropouts happen in week one, and a learner who passes the 30% mark has a 75% chance of finishing. TecPal records attendance, assignment submissions and feed activity, but nothing watches those signals. An admin queue that flags "no class, no submission, no login for 5 days" and prompts a call, not an email, is the single highest-return build left.

Built · admin-atrisk.html

First-week engineeringImmediate hands-on projects cut week-one dropout by about 22%, and first-week community participation predicts completion better than anything else. The Bootcamp already ships a day-one project; the gap is making the first cohort-feed post a required, structured step of onboarding rather than an invitation.

Policy + small build

Small stable study groupsRetention research favours small fixed teams over open forums. The cohort feed and study rooms exist; assigning every learner to a named 4-person pod in week one, with the pod visible on their dashboard, is the missing structure.

Built · dashboard pod

Post-class pulseOne tap after each class: was today useful, yes or no, optional line of text. Feeds the tutor and catches a failing cohort in days rather than at the end-of-course survey.

Built · classroom rail

Payments and financingThe upfront fee is the enrolment barrier; 87% of surveyed students would not consider a programme without flexible financing

Education financing partnerBNPL for education is live in Nigeria (CredPal, EasyBuy on Jumia; Lipa Later runs education BNPL regionally) and AltSchool already competes on income-share agreements. TecPal has instalments but no third-party financing, which caps enrolment at people with the cash. A CredPal-style partnership puts "from ₦X/month, no upfront" on every Diploma card. Commercial negotiation only you can start.

Founder to-do

Sponsor a learnerA distinctly Nigerian pattern the market underserves: diaspora relatives paying school fees. A checkout variant where someone abroad pays for a named learner, gets the receipt and progress updates, and the learner gets the seat. Card-in-dollars via Paystack already works.

Built · sponsor.html

Income-share agreementsDeliberately skipped for now, despite AltSchool doing it. ISAs in Nigeria mean collections risk, income verification you cannot do cheaply, and a regulatory grey zone. Revisit when the outcomes data is two cohorts deep and verified.

Deliberate skip

Security and complianceEducation is a top ransomware target and institutional buyers now ask for audit evidence before contracts

SOC 2 Type II, timed to B2BInstitutions and employers increasingly require SOC 2 from EdTech vendors before contracts. Cost runs roughly $40k to $70k with a 6 to 12 month observation window. Not worth starting today; worth starting the quarter the first For Teams or sponsorship deal passes ₦50m a year. NDPA registration comes first regardless.

Founder to-do, timed

Public status pagestatus.tecpal.ai on Better Stack's free tier. When the site or checkout has a bad hour, a status page converts panic and scam-suspicion into "they know, they are on it". For a trust-first brand this is marketing as much as operations.

Founder to-do, 1 hour

Incident response runbookNDPA expects breach notification within 72 hours. Nobody has written down who declares an incident, who calls the NDPC, who emails learners, and what the email says. One page, written before it is needed. Added to DEPLOYMENT.md §16.7 as a template.

Written this pass

Already in place and ahead of peersRLS on every table, Turnstile on the money endpoints, rate limits, non-enumerable certificate codes, consent-gated analytics with PII scrubbing, 2FA mandate for staff, audit trail, signed video URLs. This section exists because most local competitors have none of it.

In place

Learning experienceEngagement drops half after six minutes of video; quizzes every few minutes lift retention about 40%

Six-minute rule for recorded contentWhen full courses are produced, cut lessons into segments of six minutes or less with a check question every few minutes, rather than publishing the class recording as the course. The live recording stays as revision; the produced course is a different artifact. This is a content-production standard, now written into the CMS guidance.

Standard written

Install-to-home-screen (PWA)Manifest plus a small service worker makes the dashboard installable with cached shell and downloaded recordings reachable offline. On mid-range Androids this is the difference between an app and a bookmark, without app-store cost.

Built · manifest + sw.js

Adaptive pathsBranching content based on quiz results is the 2026 premium standard. Right sequencing: after real course content exists, not before.

Deliberate skip for now

Assessment, growth, operations, AIThe shorter lists

Shareable skill badges (Open Badges)Micro-credentials carry real employer recognition now. TecPal has internal achievements; the gap is a per-module badge a learner can put on LinkedIn that deep-links to the verify page. Rides on infrastructure that already exists.

Built · achievements share

Assessment integrity statementFinal assessments are human-marked, which is right. Publish one paragraph on how identity is confirmed for virtual assessment (camera on, ID shown, project defended live), because employers ask exactly this before trusting a certificate.

Built · course page

WhatsApp ChannelWATI covers 1-to-1; a public WhatsApp Channel is free organic reach in the one app every Nigerian learner opens daily. Post cohort dates, free events, one tip a week.

Founder to-do

Google Business Profile + reviews engineAlready specified in DEPLOYMENT.md §17.4 and still the highest-return marketing hour available. Every graduating cohort gets asked for a campus review.

Founder to-do

Tutor workload viewThe timetable detects room and tutor clashes; it does not total weekly teaching hours per tutor, which is how burnout hides. Small addition to the existing panel.

Built · timetable panel

Pal grounded on live dataPal's answers are authored; production Pal should read the catalog, timetable and the learner's own record through the Edge Functions, with the same PII rules the analytics layer enforces. Already specified for the developer; restated here because it is the AI feature learners will actually judge.

Developer

Outcomes data pipelineThe survey templates exist in Messaging; the missing piece is the discipline: no published stat without a survey response behind it, and the Outcomes page updates quarterly. This is what keeps "verified, not claimed" true a year from now.

Founder policy