An analysis of INEC’s IReV portal performance during the 2026 Osun governorship election, highlighting transmission progress and identifying critical transparency gaps that must be addressed to ensure electoral readiness for the 2027 general elections.
The performance of the Independent National Electoral Commission’s Result Viewing Portal (IReV) during the 2026 Osun governorship election offers an opportunity to evaluate the platform’s preparedness ahead of the next general election cycle.
Its usefulness, therefore, depends not only on whether results eventually appear on the portal but also on how quickly they become available, how consistently transmission occurs across locations, and how transparently INEC accounts for polling units whose results are not displayed.
The Osun State governorship election involved 3,763 polling units across 30 Local Government Areas and 332 Registration Areas/Wards. CJID monitored IReV throughout the election day and into the following morning, alongside its broader observation of voting, counting, and collation across the state. The pattern showed relatively rapid transmission during the first several hours after voting, a period of slower activity later in the evening, and another substantial increase in uploads overnight.
Results began appearing on IReV while voting and counting were still concluding in parts of the state. At approximately 4:12 p.m. on Saturday, 437 of the expected 3,763 polling-unit results had been uploaded, representing about 11.6 per cent. The number subsequently increased to 849, or 22.56 per cent, and later to 2,463 results, representing 65.45 per cent.
By approximately 9:30 p.m., 3,292 polling-unit results were available, putting statewide coverage at 87.5 per cent and leaving 471 results outstanding. Considering that the official close of polls was 2:30 p.m., and that voting continued beyond that time in polling units where voters remained in the queue, a substantial proportion of the state’s polling-unit results had therefore become publicly accessible within several hours of voting and counting.
Transmission did not progress at the same pace throughout the evening. CJID observed a period when the number of uploaded results remained at 3,292, with the last upload at that stage recorded at approximately 8:48 p.m. Uploading subsequently resumed, with CJID recording 3,482 results at approximately 11:42 p.m., 3,537 shortly afterwards and 3,5,57, shortly after midnight. By 6:55 a.m. on Sunday, 16 August, the figure had reached 3,681, representing 97.82 per cent, and at approximately 7:42 a.m., 3,690 of the expected 3,763 polling-unit results were available, taking statewide coverage to 98.06 per cent.
The progression from 87.5 per cent at 9:30 p.m. to 98.06 per cent the following morning shows that transmission continued substantially overnight,even after the slowdown observed earlier in the evening. It also means that most polling-unit results were publicly available while collation was still underway, which is particularly important for a platform whose principal contribution to the electoral process is the ability it gives the public to independently examine polling-unit records as results move through the collation process.
The statewide figures, however, conceal differences in transmission across LGAs. At one point when statewide uploads were approaching 95 per cent, Irepodun and Ila were already at approximately 99.2 per cent, Ede South and Obokun at about 99 per cent and Boluwaduro at 98.6 per cent. At the other end, Ilesa East stood at approximately 90.9 per cent, Atakumosa West and Ejigbo at about 89.5 per cent, Orolu at 89.4 per cent and Ola-Oluwa at approximately 80.6 per cent.
The gaps narrowed as transmission continued, but their presence suggests that statewide upload percentages alone are insufficient for understanding IReV performance. Differences between LGAs could reflect the timing of poll closure and counting, network availability, BVAS-related issues, geography, security conditions or the conduct of election officials. Establishing which of these factors accounted for slower transmission would require more detailed operational data from INEC, but identifying where delays are concentrated is useful for determining where improvements may be required before subsequent elections.
By 7:42 a.m., 73 of Osun’s 3,763 polling units did not have results displayed on IReV. The final governorship result was announced while the portal remained at 98.06 per cent.
IReV is a public result-viewing platform rather than the statutory collation mechanism through which election winners are determined, so the significance of the 73 outstanding polling units lies principally in the completeness of the public record. Once an election has been concluded, citizens examining IReV should ideally be able to understand what happened at every polling unit, including those from which there is no result to display.
CJID’s field reports from Osun illustrate why this requires more information than an upload percentage can provide. In Ife Central, observers reported that voting did not take place in one polling unit following a disagreement over the number of ballot papers supplied, while the result from another polling unit was cancelled following over-voting and the destruction of ballot papers. In circumstances such as these, the absence of a result on IReV may originate from what happened at the polling unit rather than from the transmission system.
The current presentation of IReV provides limited information for making these distinctions. A polling unit where voting was cancelled can appear in much the same way as one where a valid result exists but has not yet been transmitted. Providing a status for polling units without displayed results — whether voting was not conducted, the poll was cancelled, transmission remains pending or another documented circumstance applies — would make the portal considerably more useful for public scrutiny.
This becomes particularly important as IReV approaches complete coverage. At 60 or 70 per cent, the principal question is understandably how quickly the remaining results are being uploaded. At 98 or 99 per cent, attention begins to shift towards the smaller number of polling units that remain unaccounted for, where they are located and why their results are unavailable.
Recent off-cycle governorship elections provide some context for assessing the Osun figures, although the available records differ in what they allow us to measure. CJID tracked the Osun portal repeatedly during election night, providing a progression of uploads over several hours, while the current IReV records for Anambra and Ekiti provide clearer evidence of where those elections eventually settled than of their complete election-night trajectories.
For the 20 June 2026 Ekiti governorship election, IReV currently shows 2,390 results from 2,445 polling units, representing 97.75 per cent of the expected total. The portal records its last upload at 9:19 a.m. on 21 June, the morning after the election. Almost two months later, 55 polling units remain without results displayed on the portal.
The Ekiti LGA figures show that the gap is concentrated more heavily in some in areas than others. Several LGAs are recorded as complete, while Ado Ekiti has 339 results from 344 polling units, has Ikole 183 from 189, has Ijero 138 fr,om 145 and Ekitihas East 95 from 112. The statewide figur,e of 97.75, per cent therefore masks a more pronounced gap in particular parts of the state, with Ekiti East alone accounting for 17 of the polling units without displayed results.
The 8 November 2025 Anambra governorship election is closer to complete coverage. IReV currently shows 5,711 results from 5,720 polling units, representing 99.84 per cent, with nine polling units without displayed results. Most LGAs have complete records, while the remaining gaps are distributed across a limited number of LGAs. The portal also records the most recent upload on 23 November 2025, more than two weeks after election day, showing that the public record continued to change well after the election had concluded.
These records bring out different aspects of IReV performance. Osun gives a clearer picture of transmission speed because CJID followed the portal throughout election night and recorded its movement from 87.5 per cent at 9:30 p.m. to 98.06 per cent the following morning. Anambra shows how close the portal can eventually come to complete coverage, while Ekiti shows how a relatively small statewide gap can persist and become concentrated in particular LGAs.
They also point to the value of separating timeliness from eventual coverage when assessing IReV. Anambra’s current 99.84 per cent is higher than the 98.06 per cent recorded in Osun on the morning after the election, but the current portal does not tell us what proportion of Anambra’s results was available at equivalent points on election night. Similarly, the date of Anambra’s most recent upload tells us that the record continued changing after the election, but not the scale of those later additions. A credible comparative framework therefore requires more systematic publication of upload timestamps and performance data across elections.
The unresolved entries across the three elections also strengthen the case for better polling-unit status information.
Anambra currently has nine polling units without displayed results and Ekiti has 55, while 73 were outstanding in Osun at the point CJID recorded the portal on Sunday morning. The figures themselves cannot explain whether the gaps arose from transmission delays, cancellations, non-conduct of elections or other circumstances. As IReV increasingly records statewide coverage in the high nineties, explaining these remaining entries becomes part of assessing the completeness of the transparency system.
The public significance attached to these questions is inseparable from the experience of the 2023 presidential election, when IReV became one of the most contested aspects of the results process. INEC subsequently acknowledged that its inability to upload presidential polling-unit results as expected generated widespread criticism and affected public perception of the election, while the Supreme Court observed during subsequent litigation that the malfunction contributed to reduced public confidence.
INEC’s subsequent explanation showed that the problem was more specific than a general collapse of IReV. According to the Commission, National Assembly results were uploading while presidential results encountered a configuration problem related to how the different elections had been mapped within the application.
National Assembly contests were associated with states and constituencies, whereas the presidential election was configured as a single nationwide constituency, creating difficulties when the backend attempted to organise presidential result images according to polling-unit locations.
After technical fixes were introduced, INEC encountered a further problem with a backlog of queued uploads, while some BVAS devices had already been switched off and had to be reconnected before pending transmissions could proceed. The Commission later reported strengthening IReV resilience and introducing additional quality-assurance checks and end-to-end testing.
The recent off-cycle elections suggest a more stable operational environment. Across Anambra, Ekiti and Osun, IReV has been able to publish the overwhelming majority of expected polling-unit results, while the detailed monitoring undertaken in Osun shows that more than 98 per cent were available by the morning after voting. The kind of widespread presidential-result upload problem experienced in 2023 was not observed in Osun.
The scale of a general election, however, is considerably different. Osun involved 3,763 polling units, while INEC’s 2023 post-election review records 176,846 polling units nationwide. A general election also requires electoral infrastructure to operate simultaneously across 36 states and the Federal Capital Territory, under widely varying telecommunications, security and logistical conditions, while supporting multiple elections and a much larger number of devices and election personnel.
The experience of 2023 is relevant here because it showed that successful performance under one electoral configuration does not guarantee the same outcome under another. The fact that National Assembly uploads were proceeding while presidential uploads encountered difficulties demonstrates why the performance recorded in state governorship elections cannot by itself establish how IReV will behave during a nationwide general election.
