How Lumipad identifies, vets, and prepares cohort trainees from rural Mindanao communities. Six phases that walk from "we want to recruit a cohort" to "twelve people are ready to start Week 1." The program's most important decisions happen before training begins — getting recruitment right makes every subsequent week easier.
Lumipad runs cohorts of six to twelve trainees over six weeks. Selection happens before they arrive — and the quality of selection determines whether the curriculum lands or struggles. A cohort with the right people is a fast, generous, mutually-supportive training experience. A cohort with poor selection is a slow grind for everyone in it. The pages that follow describe how we get this right.
Recruitment for this program is not like recruiting for a corporate role. The candidates we want — rural Filipinos with mechanical aptitude, community standing, and the commitment to be away from farming work for six weeks — don't have LinkedIn profiles or polished CVs. They're identified through cooperatives, barangay-level trust networks, and graduates recommendations. The selection criteria favour people who can build, not just people who test well. The single most predictive recruitment signal we have is whether someone has fixed something mechanical in their life.
The full recruitment cycle takes about eight weeks. The first three phases run sequentially — identify communities, then conduct outreach, then collect applications. The final three phases overlap somewhat — assessment runs while applications are still trickling in, and pre-program preparation begins for selected trainees while final selection decisions get made for marginal cases.
Each phase below is a self-contained module. Click a phase to see its objectives, day-by-day work, materials needed, and decision criteria. Partner organisations adapting this for their context should run phases in order — the social sequencing matters at least as much as the operational sequencing.
Mindanao has hundreds of agricultural cooperatives. Not all are good fits for Lumipad. The first phase narrows the wide field to two or three target communities for the upcoming cohort — close enough geographically that training travel is reasonable, structurally sound enough that recruitment goes through real institutional channels, and with the kind of agricultural challenges that NDVI surveys actually address.
Three signals indicate a community should be removed from the shortlist, even if everything else looks good:
Don't skip site visits to save time. Phone calls and emails reveal what people will tell you over a phone call — which is often the polished version. Day-on-the-ground reveals the real cooperative dynamics.
How candidates actually hear about Lumipad. Three parallel channels: through partner cooperatives (the formal track), through graduate communitys (the highest-yield track), and through direct community events (the discovery track). The mix matters because each surfaces a different kind of candidate. Cooperatives surface organisationally-vetted candidates; graduates surface mechanically-skilled ones; community events surface the people the first two miss.
Outreach is also where we manage expectations. Three points we make explicit at every info session:
Be honest about all three from the start. Candidates who join with mistaken expectations drop out mid-cohort — disrupting the cohort cell, wasting program resources, and damaging the relationship with their home community.
Candidates submit formal applications. We deliberately keep the form short (one page front-and-back) and paper-based — many candidates don't have reliable internet, and a too-elaborate form filters out good candidates whose only barrier is unfamiliarity with formal paperwork. The application is not the primary filter. Phase 4's hands-on assessment is.
Equally important: we don't ask about three things common in corporate recruitment that aren't relevant here.
The single most important phase of recruitment. A half-day visit with each candidate — typically conducted at the cooperative office or barangay hall in their home community — that combines a hands-on mechanical task, a literacy and arithmetic check, a structured conversation, and an honest discussion about whether they're ready. This is where most decisions actually get made.
Each candidate gets a numerical score (1–5) on four dimensions, plus written notes:
A candidate scoring 5/5/3/4 is competitive. A candidate scoring 5/2/5/5 is borderline — strong mechanical and commitment but might struggle academically. A candidate scoring 2/3/3/3 is unlikely to be selected this cycle. Don't reduce assessment to a single average; the dimension distribution matters.
Final selection happens in a two-day team review with all assessors present. The output is not just "who's in" but "who's in with whom" — trainees get assigned to cohort cells of 3–4 people who'll support each other through training and become a small business cell post-graduation. Cell composition is its own optimisation problem.
The cohort cell is borrowed from cooperative organisation in Latin America (the Mondragón cooperative network's "primary cooperatives") and adapted for our context. The reasoning: a single graduate running a microenterprise alone is fragile — one drone breakdown, one bad month, one personal emergency, and the business stops. A cell of 3–4 graduates working semi-collectively is robust. They can cover for each other, pool capital for spare drones, and divide work between technical and customer-facing roles.
Cell formation during selection (rather than after graduation) means cell members spend the entire 6-week training together — building trust, learning each other's working styles, working through small conflicts in the safe environment of training before the stakes get higher.
Don't form cells purely by skill complementarity — geographic clustering matters more. A perfect skill mix that lives 6 hours apart cannot collaborate in practice; an okay skill mix in adjacent barangays can.
Two weeks between selection notification and program start. Two parallel tracks running simultaneously: trainees prepare to leave home for six weeks, and Lumipad staff prepare the workshop, materials, and logistics. The phase ends when all twelve trainees are in Davao, lodging is sorted, materials are unpacked, and Cohort 0X is ready to begin.
Across most Mindanao communities, a person leaving for an extended period (especially for education or work) is marked by a small ceremony — often a meal, sometimes a blessing from a community elder. Lumipad actively supports these ceremonies, attending where possible and ensuring trainees have time/resources to participate properly before departure.
This isn't just cultural respect (though it is that). It's also operationally important: communities that send members to Lumipad with proper ceremony are invested in their success, expect them to return with skills, and become recruiting grounds for future cohorts. A trainee who left without ceremony is more likely to lose connection with their home community during the program.
Where local customs differ — Indigenous community ceremonies, Muslim community traditions, others — we adapt and follow the community's lead.
The principles below emerged from running cohorts and watching graduates careers unfold. Some are intuitive; others are counterintuitive enough that we've watched first-time partner orgs make the opposite call and regret it. Treat these as defaults — adapt them only with awareness of why they exist.
Partner orgs running parallel programs in other regions of the Philippines (or beyond) should adapt this playbook to their context — not copy it verbatim. Three areas where adaptation usually matters:
Mindanao agricultural cooperatives are the recruitment anchor here. In other contexts: indigenous community councils (Cordillera, Visayas), Muslim community religious organisations (BARMM), urban-poor parishes (Metro Manila satellite areas), or small-farmer associations elsewhere. The principle is "recruit through institutions with standing in the community"; adapt the specific institution to your region.
The four-dimension scoring (mechanical / literacy / commitment / fit) holds across contexts, but the weighting can shift. In communities with stronger formal education, mechanical aptitude becomes more differentiating (since literacy is widely shared). In communities with weaker access to formal education, commitment becomes most predictive — candidates who travel further to attend assessment are signaling something real.
The cell structure (3–4 trainees, geographic clustering, complementary skills) is robust across contexts. What changes: the post-graduation business model the cells are heading toward. In contexts where the graduate community's primary work is something other than NDVI surveys (e.g. infrastructure inspection, fisheries monitoring, emergency response), cell skill mix should optimise for that work, not ours.
Recruitment ceremonies, departure rituals, family-coverage norms, gender expectations all vary across regions and communities in the Philippines. We follow the community's lead in each. If you're unsure how a particular custom interacts with the program, ask the community before assuming. Particularly important in indigenous communities, BARMM areas, and conservative Catholic-rural regions.
Twelve is the upper limit for a single trainer; below six the cohort dynamics suffer. The first cohort was 6 (small for our taste, dictated by initial budget); a later cohort was 9; the current cohort will be 12. The cohort cell structure (3–4 per cell, 2–3 cells per cohort) imposes the discrete sizes — 6, 8, 9, 10, or 12 are natural; 7 and 11 require uneven cells which we avoid.
For partner orgs starting their first cohort: aim for 8 or 9. Big enough that cohort dynamics work; small enough that one trainer can manage. Scale up to 12 once the second cohort confirms the operational pattern.
Carefully. A strong individual candidate without an institutional landing pad is more likely to graduate, return home, and find no client base for the microenterprise — a worse outcome for them and for the program's reputation than not training them at all.
Two paths: (1) defer their selection by one cycle, use the time to establish institutional relationships in their community (cooperative outreach, barangay meetings); accept them in the next cohort with a real local landing. (2) Accept them now, but invest extra effort in helping them either build new institutional ties at home post-graduation, or relocate to a community where graduates already have a network.
The right answer depends on the candidate's specific circumstances. A later cohort had two such cases — one we deferred (and accepted in the next cohort), one we accepted with extended post-graduation support. Both worked out.
Cautiously, and with full disclosure to their current employer. An agricultural extension worker who learns drone-based survey skills is genuinely valuable — both for their employer and for the graduate community. But: someone with a stable existing job has different incentives than someone for whom Lumipad is their primary livelihood path.
If we accept such a candidate, we ask their employer to acknowledge the program in writing. We also have a candid conversation about whether they intend to stay employed (and use Lumipad skills there) or transition out (using Lumipad as launchpad). Both paths are fine; the candidate just needs to be honest about which.
One pattern that works well: agricultural extension workers from cooperatives we already partner with. They become bridges between Lumipad and their cooperative; the cooperative gains an in-house drone capability without the operational cost of building one from scratch.
Yes, with case-by-case assessment. The work involves hands-on building, soldering, and field operations — but the specific physical demands are narrower than they might first appear. A later cohort had a deaf trainee (excellent mechanical hands, supported by sign-language interpretation during instruction) and a graduate with limited mobility in one arm (built drones one-handed faster than most two-handed builders).
The questions we ask: (1) Can the candidate physically do the work, with reasonable accommodation? (2) Will the cohort training environment work for them — visual learning, hands-on tasks, group dynamics? (3) Will they be able to do the post-graduation work effectively in their home community?
Where any of these is uncertain, we have a Phase 4 conversation specifically about accommodation. Candidates with disabilities almost always have clearer self-knowledge than we do about what they can and can't do.
A later cohort was 4 of 9 women (44%); the current cohort will target 5 of 12 (42%). Active outreach happens via the women's auxiliary committees that most agricultural cooperatives have, which surface candidates the male-dominated coop board sometimes overlooks.
Mechanical aptitude is gender-blind. The bigger barrier for women candidates is family/farm coverage during the 6-week program — gendered expectations around childcare and household work make the time commitment harder for women than men. We're explicit about this during outreach: if you can arrange coverage, we want you in the program. Cooperatives sometimes help organise community-level childcare during cohorts.
One pattern emerging: women graduates become particularly effective recruiters of women candidates from their own communities. The Phase 2 graduates-as-recruiters channel skews toward women candidates roughly in proportion to women graduates — so each cohort's gender mix tends to compound from the previous.
Three trainees in earlier cohorts dropped out — one due to family medical emergency (Week 2), one due to mismatched expectations (Week 1), one due to homesickness (Week 3). Each was handled differently.
The medical emergency trainee was held for the next cohort with no penalty. The expectations-mismatch trainee was redirected to the local agricultural extension program (better fit for their skills and goals). The homesickness trainee was given a long weekend home and a structured conversation about whether to continue or step away — they chose to continue and graduated.
Operationally, having a 2–3 person reserve list (Phase 5) lets us bring in a substitute within Week 1 of the program if needed. After Week 1, integration costs become too high; the slot remains empty and we adjust cell composition.
Mindanao communities have real political, religious, and family conflicts. They sometimes affect recruitment in ways we have to navigate. Three patterns we've seen:
The general principle: treat candidates strictly on individual merit; be transparent about how decisions are made; respect community norms while keeping the program's standards consistent.
Roughly ₱8,000–12,000 per trainee for the recruitment cycle alone (not including the program itself). Breakdown: site visits and travel ~₱3,500/trainee, info sessions and outreach materials ~₱1,500, Phase 4 assessment travel ~₱2,500, pre-program logistics ~₱1,500, graduates-as-recruiters honorariums ~₱1,000.
For a 12-trainee cohort, the recruitment cycle costs roughly ₱100,000–145,000 — about 12–15% of the total program cost. Partner orgs running parallel programs typically come in slightly higher (₱130,000–180,000) due to less mature graduate communitys and more first-time outreach work.
The cost is well-spent: a strong recruitment cycle pays for itself many times over in avoided drop-outs, better cohort dynamics, and stronger post-graduation outcomes.