By Gigi Ngcobo, Venture Scale Lead at Injini

The rollout of an EdTech pilot programme is often measured by visible signals. Thousands of tablets deployed. Active user dashboards trending upward. “Time-on-app” statistics climbing. Positive press coverage following shortly after.

But do these signals tell us whether learning has improved?

Across Africa, the EdTech sector has grown rapidly. Yet the continent’s core education indicators remain deeply concerning. In South Africa, one of the continent’s most industrialised economies, the evidence is sobering. The 2023 Progress in International Reading Literacy Study (PIRLS) found that 81% of Grade 4 learners could not read for meaning in any language, up from 78% in 2016.

At the other end of the schooling pipeline, the 2025 Matric pass rate reached a historic 88%. Yet high dropout rates persist. These figures sit uneasily together: improved terminal pass rates alongside significant early learning deficits and attrition. They suggest a system where headline performance and foundational attainment do not necessarily move in tandem.

Education outcomes are multi-causal and slow-moving. No single intervention can shift them overnight. System-wide metrics are also blunt instruments; they cannot easily isolate the contribution of any individual tool or programme.

Even so, if EdTech were consistently delivering large, scalable learning gains under real classroom conditions, we would expect clearer signals over time.

Breaking through the busywork trap

Part of the challenge lies in how the sector is structured. The African EdTech market is often optimised for observable activity rather than verified attainment. In my work with early-stage EdTech founders across the continent, this pattern is familiar. Teams are often asked to demonstrate adoption, engagement and pipeline growth long before they are resourced to demonstrate verified learning gains.

This is not because founders, investors or funders disregard impact. It is because activity is easier to measure, contract and price.

Adoption rates, logins and usage minutes are immediately visible. They can be reported quarterly. They support investment memos and procurement decisions.

Learning gains, by contrast, are difficult to attribute, slow to verify and expensive to measure. Demonstrating whether a tool improves literacy or numeracy requires careful study design, access to classrooms, ethical oversight, substantial data and, ideally, independent researchers.

High-quality impact evidence is costly. It also behaves like a public good. Once rigorous evidence is produced, its benefits extend beyond the originating firm to competitors, policymakers and the broader ecosystem. Individual companies cannot easily capture the full return on that investment. Underinvestment in rigorous evaluation is therefore not primarily a moral failure. It is a predictable market outcome.

When markets pay for reach and engagement, they receive reach and engagement.

Infrastructure is not instruction

The sector’s data-rich environment further complicates the picture. Every technology product generates metrics. For early-stage teams, upward trends can feel like validation.

Yet in resource-constrained environments, logins and downloads are weak proxies for learning. Hardware maintenance gaps, connectivity constraints and classroom management realities can absorb time and attention without necessarily improving comprehension or mastery.

Treating “EdTech” as a single category obscures an important distinction between infrastructure and instructional tools. Devices, connectivity solutions and platforms can scale rapidly. That scale does not guarantee instructional impact.

Kenya’s Digital Literacy Programme illustrates the point. Launched in 2013 with the ambition of providing a laptop to every primary school learner, it signalled a bold commitment to digital inclusion.

The programme was later scaled back amid infrastructure gaps, limited teacher readiness, weak prioritisation and insufficient maintenance and security.

This was not a failure of technology alone. It exposed a broader constraint: implementation capacity is itself a financed input into learning outcomes. Teacher training, timetabling, monitoring, maintenance and support systems form part of the production function of learning. Without sustained investment in that capability bundle, even well-designed tools are likely to underperform.

The new investor thesis: The "unit economics of impact"

If the sector is serious about attainment, the investor and funder conversation must evolve. Instead of asking only, “How much does the app cost?”, we should also ask, “What is the fully loaded cost of generating a verified learning gain under real classroom constraints?”

That question includes software, training, implementation support and the cost of credible evaluation.

This is what I describe as the “unit economics of impact.”

In conventional venture logic, unit economics measures the marginal cost of acquiring and serving a customer relative to revenue. In education, the meaningful unit is not merely a user account. It is a demonstrable improvement in learning.

Yet learning gains are not easily contractible at scale. They are context-dependent, noisy and slow to verify. Standardised benchmarks are imperfect. Attribution is contested. Measurement requires time.

As a result, the financial architecture of the sector rarely prices verified attainment directly. Instead, it prices adoption, engagement and coverage.

Without funding structures that underwrite evaluation infrastructure and implementation capacity, EdTech in Africa will continue to face a paradox: high activity alongside ambiguous system-level impact, and promising products that remain financially fragile.

What disciplined implementation looks like

Evidence from the Nord Anglia Education Metacognition Project, conducted across 27 schools globally, underscores a consistent finding in education research: technology’s impact depends less on novelty and more on its integration into proven learning strategies.

Digital tools that embed structured reflection and “thinking routines” can support measurable gains in critical thinking and curiosity. The mechanism is not screen time. It is pedagogical coherence.

Technology becomes productive when it is embedded within a clear instructional model, supported by teacher training and used consistently. Implementation is not an afterthought. It is part of the learning production process.

From return on investment to return on instruction

Reframing EdTech through the unit economics of impact shifts the focus from short-term visibility to long-term instructional return.

If investors and funders treat evidence as shared infrastructure, akin to public health research or transport systems, and support independent evaluation, theory-of-change rigour and sustained outcome tracking, the sector can begin to price what matters.

This does not mean abandoning growth. It means aligning growth with verified attainment.

Ultimately, if African EdTech is to leapfrog global standards, it must stop counting who opened the door and start measuring what changed once they walked through it — and we must build financing and procurement systems that can pay for that change.