At the national tallying center in Bomas of Kenya, presidential results initially appeared on a large electronic screen as they arrived from polling stations. Kenyans could also follow the figures on television, radio and online. Then the transmission system stalled after receiving results from about 14,000 of the country’s 32,613 polling stations. The screen meant to make the election visible became an advertisement for a system that had stopped working.
Technology carried unusual expectations into the election held on 4 March 2013. Kenya was voting under a new Constitution and through a new electoral commission after the 2007 post-election crisis. This election rested upon the promise that biometric registration would clean the voter register, electronic devices would identify voters, and polling officers would transmit provisional results directly from polling stations. Each system promised to close an opening through which previous elections had been questioned and the country thrown into crisis.
Three technologies, three different jobs
The three technologies did not perform the same task in the election process. Biometric voter registration captured fingerprints and personal details while preparing the register. Electronic voter identification devices were supposed to confirm voters at polling stations. The results transmission system was meant to send provisional figures after votes were counted manually. Failure in one system did not prove that another had been manipulated, although public discussion often combined them into one story about election technology.
Also Read: The 2007 Presidential Election No One Could Prove Who Won
Biometric registration produced a register containing about 14.35 million voters, but the commission had only thirty days to conduct the exercise. IEBC chairman Ahmed Issack Hassan later estimated that approximately three million eligible Kenyans remained unregistered. The European Union Election Observation Mission reached a similar conclusion and found that some marginalized communities still lacked the national identity documents required for registration.
The more visible breakdown occurred during voter identification. Electronic devices were either unavailable, unused or not working in roughly half the polling stations observed by the European Union mission. Hassan later said that only about 40 per cent worked during the election. Officials returned to printed registers, lists of voters whose biometric details had not been captured and, in some places, the green books used during registration. These alternatives allowed voting to continue, but they left no single method for recording who had been identified and permitted to vote.
The results screen that stopped working
The transmission system then failed to provide the promised second layer of public verification. It had been designed to send provisional presidential results from polling stations through configured mobile phones over a closed network. Those figures were never meant to replace the signed paper returns used for the official tally. They were supposed to let the public compare what polling stations reported with what the commission later announced.
Its collapse mattered because the provisional system had become IEBC’s main public window into the count. A programming error made rejected votes appear eight times higher than the figures being entered, briefly pushing their displayed proportion above five per cent. The transmission platform later stopped after receiving results from fewer than half the polling stations. IEBC then concentrated on manually tallying the official constituency returns while candidates and citizens waited for figures they could no longer follow independently.
What observers and the Court actually found
That failure did not prove that Uhuru Kenyatta had been awarded votes he did not receive. The European Union observers reported that counting was conducted properly in the polling stations they attended and that recorded results reflected the votes cast there. They found minor discrepancies during tallying, but almost all were below one per cent and showed no overall pattern favoring either leading candidate. Technology had failed, yet the available observer evidence did not establish that the failure changed the winner.
The official result gave Kenyatta 6,173,433 votes, representing 50.07 per cent of valid votes cast, while Raila Odinga received 5,340,546. That narrow passage above the constitutional threshold made every dispute concerning rejected votes, the register and the tally politically significant. Odinga and other petitioners challenged the election before the Supreme Court, arguing that the register was unreliable and that the collapse of electronic identification and transmission had damaged the result.
The Court unanimously upheld Kenyatta’s election after hearing the consolidated presidential petitions. It found that electronic technology was intended to support the election rather than replace the legally recognized manual process, and that failure of the electronic systems did not invalidate the paper-based tally. It also ruled that rejected ballots should not be included when calculating whether a candidate had obtained more than half of all votes cast. The Court concluded that the petitioners had not supplied sufficient proof that the irregularities overturned the result declared by IEBC.
The lesson Kenya still needs before 2027
The Court upheld the result, but its judgment did not erase the documented failure of the election technology. The European Union mission found that IEBC introduced some systems only months before polling, leaving too little time for procurement, testing, deployment and practical training. After the election, Hassan described the mistake through a proverb: “Never test the depth of a river with both feet.” By then, IEBC had already stepped into the river carrying three new technologies and the credibility of the first presidential election conducted under the 2010 Constitution.
Also Read: Kenyans Warned of Media Blackout and Internet Shutdown
The mistake was not simply buying and deploying machines that failed. IEBC allowed public confidence to depend on systems whose breakdown was not matched by an equally visible fallback. Manual voting and tallying can produce a valid election, but citizens must know beforehand which record becomes authoritative when a device refuses a fingerprint or a server stops receiving results. The fallback cannot appear after failure as an explanation offered by officials under pressure.
Before 2027, IEBC must publish the purpose, testing record and failure procedure for every election technology it intends to use. Polling officials should know which document controls when a device fails, while agents and observers should be able to follow the replacement procedure without waiting for instructions from Nairobi. Under the principle later affirmed in the Maina Kiai case, constituency presidential results are final and cannot be altered at the national tallying center, although the national process still has the necessary work of verification, aggregation and declaration. Kenya does not need another electronic screen that works beautifully until the presidential figures become close. It needs a count that remains visible after the screen goes blank.
Follow our WhatsApp Channel and X Account for real-time news updates.
This article was written by George Nyongesa, a lecturer in philosophy and logic at the University of Nairobi and Chuka University.
