How to Research a Kenyan County Before You Campaign in It
Anybody can learn a county. Most campaigns do not, and then wonder why a message that works two counties away lands badly here.
What follows is a method rather than a summary. It works for a candidate moving into unfamiliar ground, for a party official assigned a region, and for a journalist who has been told to cover somewhere new.
Step one: get the units right before anything else
Find out exactly how many constituencies the county has, how many wards, and what each of them is called. Do not estimate this and do not rely on memory.
Then get the gazetted polling station list for the county, with the codes. That single document gives you the full hierarchy from county down to station, and it is the spine every other piece of research hangs on. Related: how the units nest.
Getting this wrong is the fastest way to be dismissed by people who live there. Naming a ward that is in the neighbouring constituency, in public, is a mistake nobody forgets.
Step two: read the county's own documents
Counties publish more than most people realise, and almost nobody reads it.
Start with the county integrated development plan, which sets out what the county says it intends to do and where. Then the annual budget documents and the fiscal strategy paper, which show what it actually funded. Then the Auditor-General's reports, which show what happened to the money afterwards.
Reading those three in that order tells you the difference between what a county government promised, what it paid for and what it can account for. That difference is the substance of most county-level politics, and it is available in writing.
Step three: read the county assembly
The assembly is where county law is made and where the budget is approved, and its record is the closest thing to a scorecard for every MCA in the county.
Ask the clerk's office for order papers, committee reports and voting records. Find the committees that matter locally, usually finance, health, and lands or agriculture depending on the county. Find out which motions passed and which quietly did not.
Most of this is unreported, which means the campaign that reads it knows things the incumbent assumes nobody knows. Related: what an MCA actually does.
Step four: past results, at the smallest level available
County totals tell you who won. Station-level results tell you where.
Work through the last cycle station by station where the data allows it. You are looking for the wards that swung, the wards that never move, the stations where turnout collapsed, and the places where the party vote and the personal vote clearly separated.
That last pattern is the useful one. A ward that voted one way for President and another way for MCA is telling you that local reputation beats national alignment there, and that is a completely different campaign from a ward that voted the same way down the whole ballot.
Step five: the calendar and the geography of daily life
Market days. Harvest and planting seasons. School terms. Religious calendars. The days when the roads flood. The times of day when people are actually home.
A campaign that schedules against the county's real calendar reaches people. A campaign that schedules against its own convenience holds meetings for empty chairs. This information is not published anywhere. It comes from asking people who live there, which is also how you find the local organisers worth knowing.
Step six: understand the regional story without importing it
Every part of Kenya has a political narrative attached to it, and outsiders arrive holding a simplified version. Related: how one regional label hides two different politics.
Treat the narrative as a hypothesis. Counties inside the same region diverge, and wards inside the same county diverge further. The moment you repeat a regional generalisation to a room that knows better, you have told them you are a visitor.
What to do with the finished picture
Turn it into one page per ward.
Each page: the ward's polling centres and stations, the current MCA, the last results, the two or three issues residents raise, the gathering points, and the name of one person who lives there and will take your call. That is a document a campaign can actually run on.
Doing it for every ward in a county is weeks of work. Starting it in the middle of 2026 makes it possible. Starting it in 2027 does not.
Frequently Asked Questions
Where do I find a Kenyan county's budget and spending records?
Counties publish an integrated development plan, annual budget documents and a fiscal strategy paper, and the Auditor-General reports on how county money was accounted for. Reading them in that order shows the gap between what was promised, what was funded and what can be accounted for. The county assembly clerk's office is the practical starting point for anything that is hard to find online.
What can past election results tell a campaign?
At station level, far more than county totals do. You can see which wards swing, which never move, where turnout collapsed, and where voters split their ballot between a national choice and a local one. A ward that votes one way for President and another for MCA is telling you that local reputation outweighs party alignment there.
How long does researching a county properly take?
Weeks, if the goal is one usable page per ward covering stations, the sitting MCA, past results, local issues, gathering points and a named local contact. That is why the work belongs well before an election year rather than inside it. Started early it is a background task; started late it is impossible.