A Ward By-Election in Samburu Says More About 2027 Than Most Polls Do

UDA's Daniel Lelgojine was declared winner of the Poro Ward by-election in Samburu West, and the party attributed the win, along with other recent by-election results, to growing public confidence in President Ruto and the Kenya Kwanza administration's development agenda.

That is what every party says after every by-election it wins, and it is almost never the right reading.

What a ward by-election actually measures

Three things, in descending order of reliability.

Local organisation. Who got their voters to a polling station on a Tuesday in a contest most of the ward barely registered. This is the dominant factor and it is not national.

The candidate personally. In a ward of a few thousand voters, whether the candidate is known and liked is a larger effect than any party brand.

National mood, faintly. Somewhere behind the first two, and usually swamped by them.

Why turnout distorts everything

By-election turnout is low. Very low. In a contest where a third of the register participates, the result describes the third that showed up, and the third that shows up is disproportionately the organised, the mobilised and the paid-for.

That produces a specific bias: by-elections favour whoever has machinery, which usually means whoever is in government and can move resources and people at short notice.

So "we won the by-election" frequently means "we are better at by-elections", which is a real capability and not the same thing as being more popular.

The honest read

A single ward result is a data point about one ward. A series of them, across different regions, moving in the same direction, is worth something. One is not.

And the counter-evidence tends to arrive within months. Kenyan by-elections in this cycle have gone in more than one direction, which is what you would expect if local organisation rather than national mood is doing the work.

The lesson that is actually transferable

Here is the useful part, and it applies to every campaign reading this.

By-elections are won by turnout operations. Not by rallies, not by posters, and not by national polling. By identifying your voters individually, knowing which polling station each of them votes at, and getting them there on the day.

That capability is exactly as valuable in a general election as it is in a by-election. The difference is that in a general election everybody else is also mobilising, so the advantage is smaller — which means it has to be built up over a much longer period to matter.

A campaign that starts identifying supporters household by household in mid-2026, records where each one is registered, and reviews the count weekly by ward, has a turnout operation by August 2027. A campaign that reads by-election results as omens has a narrative.

We built BallotWinners for the first kind: every household visit recorded at the door, counted by ward and by the volunteer who made it, so the ward that is behind is visible while there is still time to do something about it.

What a by-election does tell you, reliably

Three things, and they are worth extracting because they are actionable.

Which side can move resources fast. A by-election is a logistics test conducted at short notice. Whoever wins it demonstrated the ability to deploy people, transport and money into one constituency within weeks. That capability is real and it transfers.

Whether local structures still exist. Parties discover in by-elections that branch offices they believe they have do not function. It is the cheapest possible way to find that out.

How the security environment is trending. Kenya's recent by-elections have been marked by violence, intimidation and bribery allegations, and each one is a data point about what a general election day will look like.

What it does not tell you

The national mood, the state of the presidential race, or how the same ward will behave when 60 per cent of the register turns out instead of 30. Those are different questions and a by-election answers none of them.