Partners often come to us needing a specific installation done in a fixed, short window: two days on-site, sometimes less, often at a location that's several hours from Kisumu. Getting a deployment like that right depends far more on what happens before the crew leaves the workshop than on anything that happens once they arrive. This is roughly the process we run.
Start with a real site survey, not a description of the site
A written description or even photos from a partner are a starting point, not a substitute for direct confirmation of the things that actually determine equipment choices: power availability and type at the site, connectivity (cellular signal strength, or whether LoRaWAN gateway coverage exists), physical mounting options and their condition, and site access (can a vehicle reach the exact point, or does equipment need to be carried the last stretch). Wherever a remote survey (a video call walkthrough with someone on-site, if a prior visit isn't practical) can substitute for our own visit, we use it, specifically to catch anything that would change our equipment list before we've already loaded the vehicle.
Build the equipment list against failure, not just against success
A field kit for a short deployment window should assume something won't go to plan, because on a fixed two-day schedule there's rarely room to make a second trip for a missing part. That means: spares for the components most likely to have a manufacturing defect or get damaged in transit, a broader toolkit than the specific job seems to require, because the actual site conditions sometimes differ from what the survey suggested, and enough contingency in the schedule itself (built into how the two days are planned, not just hoped for) to absorb a problem on day one without losing day two entirely.
Confirm the site-specific unknowns before departure, not on arrival
A short list of things we insist on confirming before a crew leaves, because discovering them on-site costs the most expensive kind of time: exact GPS coordinates or a clear physical description of the mounting location, who has authority to grant physical access on the day (a farm owner absent on the scheduled day is a solvable problem if known in advance, and a serious one if discovered on arrival), and any local permissions or notifications needed (a county office, a community leader, a landowner) so the crew isn't the first to raise them on-site.
Sequence the work so day one derisks day two
For a two-day window, we sequence deliberately: the highest-uncertainty task (the one most likely to reveal a problem, like a mounting surface that turns out to be different from what the survey suggested, or a connectivity reading that doesn't match expectations) happens as early as possible on day one, not late on day two when there's no time left to adapt. Day two is planned around confirming the installation is genuinely working under real conditions, not just physically mounted, and training whoever on-site will be the first point of contact for basic troubleshooting.
Leaving the site with a working handover, not just working hardware
A deployment isn't complete when the hardware is mounted and reporting data. It's complete when whoever is on-site day-to-day (a farm manager, a school's ICT teacher, a county officer) knows what normal looks like, knows the simplest thing to check if something seems wrong, and has a direct way to reach us if it's beyond that. We build this handover into the schedule explicitly, rather than treating it as something to fit in if time allows, because a deployment that works perfectly but leaves nobody on-site able to recognise a problem early tends to generate the most expensive kind of maintenance callout later: the one that could have been caught on day one of the fault, not week three.
If you're planning a field deployment and want help scoping it properly before committing a crew and budget, get in touch. You can also see how we take hardware from prototype to the field.
