The retreat industry has a scaling problem. The economics of hospitality push operators toward larger groups: more guests, more revenue, lower cost per head. The result is retreats that feel like conferences. Eighty people in a ballroom doing yoga. Fifty people at a silent dinner that is not actually silent. The experience is diluted to the point where it becomes content for social media rather than a genuine intervention in a person's life.

Umnya caps every retreat at eight to fourteen guests. This is not a marketing decision. It is a design decision rooted in the science of group dynamics and the logistics of operating in remote environments.

Research on group cohesion consistently shows that meaningful connection, the kind that changes behaviour, occurs in groups of eight to twelve people. Larger groups fragment into subgroups. Smaller groups lack diversity. The work of Robin Dunbar on social brain theory identifies roughly 12 as the number of individuals with whom a person can maintain genuine mutual trust at any given time. The sweet spot for transformative group experience is remarkably narrow, and it is not eighty.

There is also a practical dimension. The Sahara is not a resort. Erg Chigaga is reached by a four-hour 4x4 drive from the last paved road. Every litre of water, every meal, every piece of equipment must be transported across open desert. Operating at scale here is not just undesirable. It is impossible without compromising the environment and the experience. The logistics of a 14-person camp are already complex. The logistics of a 40-person camp would require infrastructure that would destroy the reason for going.

The same constraint applies in the Atlas. Traditional gites in Amazigh mountain villages were built for families, not tour groups. They have four or five rooms, a central hearth, a terrace facing the valley. Bringing 30 people into that space would end the quality that makes it worth visiting. The architecture of the High Atlas dictates the retreat size as much as any design philosophy does.

In practice, the group size affects every element of the retreat. The guide knows every guest's physical limitations and personal history by the end of day one. The programme can pivot when a guest is struggling, adding rest time, modifying a session, choosing a different route. The dinner conversation has no subgroups. The silence on the dunes is actually silent.

The social dynamic within a small group is also categorically different from that of a large one. Research on peer accountability in behaviour change, published across sports psychology and health behaviour journals, consistently shows that small groups create conditions for genuine mutual investment that larger cohorts cannot sustain. After day two of an Umnya retreat, guests are typically talking to each other about things they do not discuss at home. Not because the retreat is therapeutic in a clinical sense, but because a small group in an extraordinary environment produces a quality of honesty that most participants describe as rare in their adult lives.

The economics deserve examination too, because the question of scale is often framed as purely commercial. It is not. A 14-person retreat can be designed and guided with genuine quality by two or three specialists. A 50-person retreat requires a management layer, a scheduling infrastructure, a standardised programme that cannot deviate, and a set of physical facilities designed for volume rather than experience. By the time you have built the infrastructure required to serve 50 people well in a remote environment, you have built something that serves none of them as well as a small operation could.

Guests sometimes ask whether the small group format means limited social diversity, whether they will spend eight days with the same person. The reality is the opposite. A group of eight to fourteen people, carefully assembled from a shared interest in movement and environment, produces more varied and substantive conversation than a large retreat ever could. There are no cliques because there are no sub-groups. There is no retreat persona, the performance of wellness that large groups invite, because there are not enough people in the room to perform for. What remains, after a day in the desert, is simply what you actually are.

The format also allows the programme to be genuinely responsive to the group. If three guests arrive with a shared interest in breathwork that the itinerary had not foregrounded, the programme can pivot. If the group's energy on day five calls for stillness rather than movement, the afternoon can be restructured. This is not improvisation. It is the design advantage of small-scale operation: the feedback loop between guest and guide is short enough to be useful, and the programme is loose enough to respond.

The result is a retreat where every guest is known by name. Where the programme can adapt in real time to the group's energy. Where dinner is a conversation, not a seating plan. Where the silence is actually silent.

Mass retreats sell transformation. Micro-retreats deliver it.

The format is not for everyone. Some guests arrive expecting the anonymous comfort of a large resort, where the scale means they can opt out of connection and move through the week unobserved. A small group makes that impossible. By day two, everyone knows your name, your pace, and your preferred coffee temperature. For guests who find that intrusive, we are honest about it in advance. For the majority, it is the aspect they remember longest. The retreat experience fades. The relationships, formed in extremis, in a landscape that reduces everyone to the same size, tend not to.