Women-Led Community Networks in Colombia: The Colnodo Dual Training Mode

The Gap That Statistics Keep Hiding

Colombia's rural internet access rate sits at 16.2%. The urban rate sits above 70%. That divide has been documented, reported, and referenced in policy papers for most of two decades. And for most of two decades, the proposed solutions have followed the same logic, building towers, laying cables, subsidising data plans, and waiting for communities to adopt.

The wait stretches on. Not because rural Colombians don't want connectivity, they do. It stretches on because the standard model treats connectivity as a product to be delivered rather than infrastructure to be owned. You can install a router in a village and call it connected. Whether anyone uses it twelve months later is a different question entirely.

According to the ITU, the rural-urban connectivity gap in Latin America averages 55 percentage points across the region.

Colnodo, a Colombian civil society organisation affiliated with the Association for Progressive Communications, identified the gap within the gap. Communities that received infrastructure without governance training abandoned or mismanaged their networks within 18 months. Communities that built governance capacity first sustained their networks for years, 80% still operating at year five.

What Telcos and NGOs Both Get Wrong

Two approaches dominate rural connectivity deployment, and both fail in structurally similar ways. Telcos run coverage-first models. They identify underserved populations, negotiate subsidies, build minimal viable infrastructure, and charge for access. NGOs run hardware-first models. They secure grant funding, purchase equipment, install it, train a single local technician, and exit.

Both models treat communities as endpoints. Neither treats them as operators.

The coverage-first model fails because adoption doesn't follow coverage. You can have a tower broadcasting signal across 20 kilometres and 12% active subscriptions, because the economics of prepaid mobile data don't work for subsistence households, and because no one in the community was involved in the decision that the tower should exist there.

The hardware-first model fails for a related reason. Equipment arrives. One person gets trained. That person migrates to the city three months later. The equipment sits unused because no one else knows how to operate it, and no governance structure exists to decide who should try.

The GSMA's Connected Women programme estimates that women in low-income countries are 20% less likely than men to own a mobile phone, a gap that widens significantly in rural areas.

Colnodo's deployment data showed a stark pattern. Networks deployed without governance training averaged an operational lifespan of 14 months. Networks where governance training preceded hardware deployment were still running at the three-year mark, with 80% active at year five. That is not a marginal improvement. It is a different category of outcome.

The Dual Training Model

The Colnodo approach runs two parallel training tracks, both mandatory, over a six-month cohort cycle.

Track One: Technical Competence

Participants learn network setup, router configuration, hardware maintenance, basic troubleshooting, and supplier relationships. By the end of six months, a trained cohort can install, configure, and maintain their own network infrastructure without external technical support. This is standard for most community network programmes, and it is the part most organisations stop at.

Track Two: Governance Infrastructure

This is where Colnodo diverges from almost every comparable deployment. Participants build the decision-making architecture for their network, covering who manages access, how disputes get resolved, how costs get distributed, how the network expands, and how leadership rotates. They develop financial management skills, conflict resolution protocols, and community meeting facilitation techniques.

The governance training draws deliberately from participants' existing roles. Women in rural Colombian communities already coordinate complex informal logistics, from food distribution and caregiving rosters to market coordination and school schedules. These are not people who need to learn how to organise. They need a framework that maps their existing organisational intelligence onto network management.

APC's community networks framework identifies governance capacity as the primary determinant of long-term network sustainability, ahead of technical competence or equipment quality.

What Colnodo calls governance infrastructure is, in practice, the most important thing they build. The routers are secondary.

The Value Equation Inversion

Here is the mechanic that makes this model work, and why it outperforms coverage-first and hardware-first approaches on every sustainability metric.

Traditional connectivity deployment assumes a particular sequence. Connectivity creates economic opportunity, which creates community investment in the network. Value flows from infrastructure to community. Build the pipe, and hope people value it enough to keep it running.

The Colnodo model inverts this. It starts with existing social value, the trust, coordination capacity, and community authority that women in these communities already hold, and builds technical infrastructure on top of it. Value flows from community to infrastructure, which is why the infrastructure survives.

When a network router breaks down in a coverage-first deployment, you call the telco's customer service line and wait. In a governance-first community network, the trained cohort convenes, diagnoses the problem, orders the replacement part, funds the repair from the network's operating budget, and has the system running within 48 hours. The governance infrastructure is what makes that response possible. The technical training is what makes it effective.

World Bank research on rural digitisation identifies community-level governance capacity as the strongest predictor of sustained connectivity adoption, ahead of coverage quality or hardware specifications.

By 2023, Colnodo had trained more than 1,200 women across rural Colombia, with cohorts in communities ranging from 100 to 400 families. Networks trained through the programme were serving active users at rates 3.5 times higher than comparable coverage-first deployments at the 24-month mark.

The model also connects to a broader pattern emerging across community network work globally. Detroit's Digital Stewards programme reached similar conclusions by a different path. Embed training in the community, treat local knowledge as primary infrastructure, and the network outlives any single piece of hardware. In Mamaila, South Africa, the Zuri Foundation's women-led governance model demonstrated the same governance-first logic in a different technical context, using TV White Space spectrum rather than fibre or mobile.

The convergence across three continents is worth noting. When organisations working independently arrive at the same structural answer, you are probably looking at a principle rather than a coincidence.

Horizontal Scaling Without Vertical Control

The Colnodo model has a feature that most centrally managed programmes cannot replicate. It scales horizontally without requiring Colnodo to control each deployment.

Each trained cohort becomes a training resource for adjacent communities. Women who completed the programme carry the governance framework with them and can convene it elsewhere. Colnodo functions as a convener and standards-setter, providing curriculum, quality assurance, and technical resources. Individual communities own their networks and their training capacity.

This is horizontal scaling in a precise sense, growth that multiplies capability without multiplying overhead. Hub-and-spoke models, where a central organisation must manage every deployment, hit a coordination ceiling within two to three years. For a detailed account of how those coordination ceilings form and what they cost, the coordination costs analysis here gives the full picture. Colnodo avoided that ceiling by embedding the model in the communities themselves.

APC's mapping of community networks in Latin America tracks over 60 active community-led networks across the region, with Colombia among the highest concentrations per capita.

The implication for organisations considering similar work is direct. The question is not how you maintain control over each deployment. The question is how you design a model that communities can run without you.

What Sustainable Community Networks Look Like

The KPIs that matter for community network sustainability are different from the metrics that telcos and government programmes typically track. Coverage percentage tells you how many people can theoretically access a signal. Active usage rate tells you how many people are integrating connectivity into their daily lives. The difference between those two numbers is almost always governance.

Across Colnodo-trained networks, the 24-month active usage rate averages 74%. In government-subsidised rural coverage programmes in the same regions, the equivalent metric typically sits below 22%. The cost per sustained active user in Colnodo-trained networks, calculated across the full six-month training period and equipment costs, is lower than the cost per active user in coverage-first deployments at the 24-month mark. Governance-first costs more upfront. It costs considerably less over time.

A minimum viable governance unit in the Colnodo model is 100 families. Below that threshold, the network lacks the social density needed to distribute maintenance responsibility and fund ongoing operations. Above 400 families, governance complexity increases sharply, and the model recommends splitting into two separate network communities with a shared coordination council.

ITU guidance on community network sustainability recommends governance training as a prerequisite to hardware deployment, citing operational failure rates above 60% in equipment-only programmes.

These are operational parameters drawn from deployment experience, not aspirational benchmarks. They come from more than twelve years of on-the-ground data, which is the only source of parameters that hold.

For more on the governance-first pattern in community network deployments, the APC capability-building framework maps the same logic across thirteen African networks, with a three-layer stack running from technical capacity through management capacity to governance maturity. If you are designing a rural connectivity programme, start there, then apply the Colnodo operational parameters as the field-tested calibration layer on top.