Favouritism in performance reviews damages trust and loses good people. Here is how to build a system where ratings reflect work, not relationships, in African companies.
Marketing Lead
April 9, 2026
•
5 Mins Read
Every HR leader in Africa has heard a version of this complaint. It sounds like: "He always gets the highest rating, and everyone on the team knows the manager just likes him."
Sometimes the employee being complained about is genuinely performing. Sometimes they are not. But the complaint is almost always evidence of the same problem: the team does not believe the review process is fair.
Favouritism in performance reviews is one of the fastest ways to lose your best people. When high performers see average colleagues rated the same or higher simply because they have better access to the manager, they start asking why they should keep delivering.
This article explains how favouritism enters the review process, why it is hard to see from inside a management relationship, and the practical steps HR can take to build a system where ratings reflect work, not proximity.
Favouritism in reviews is rarely explicit. It does not look like a manager saying "I will give Emeka a higher rating because he is my friend." It looks like this:
The SHRM has documented how the halo effect, combined with personal familiarity, creates a pattern where certain employees consistently receive better evaluations regardless of their actual performance relative to peers. In African companies with flat structures and informal management relationships, this pattern is amplified by the social closeness between managers and their teams.
In many Nigerian, Kenyan, and South African organisations, the line between professional relationships and personal ones is thin. Managers and direct reports often socialise together, share backgrounds, and know each other's families. That closeness creates warmth and loyalty that is genuinely valuable, but it also creates conditions where favouritism is hard to separate from genuine appreciation.
Two additional factors make it harder to call out:
The most effective single change is a rule: any rating of above expectations or below expectations requires a written, specific example before it is accepted. Not a character statement ("she is very committed"), but a work example ("she delivered the Lagos market expansion report three days early and the board used it in a strategic planning decision").
When managers know they must justify non-standard ratings with specifics, they self-censor the most obvious favouritism before it gets to calibration.
If a manager's rating is the only input, the most favoured employee in that manager's team will receive the most favoured rating. If the same employee receives a manager rating plus peer feedback from three colleagues plus a self-assessment, the full picture becomes harder to distort.
360-degree feedback does not eliminate favouritism, but it makes it much harder to sustain. A manager can rate their favourite employee highly on all dimensions; they cannot prevent that employee's peers from giving more mixed feedback. The contrast creates a conversation.
In calibration training, name the pattern directly: "A manager who spends a lot of time with a particular employee will know them better and like them more. That is natural. Our job today is to check that the rating reflects the work, not the relationship."
Ask calibration participants: "If this employee had worked for a different manager who did not know them socially, would the rating likely be the same?" That question forces the distinction.
Favouritism often appears as a pattern across multiple cycles. A manager who consistently gives their top ratings to the same one or two employees, while those employees' peer ratings are average, is a signal worth investigating. HR should review rating distribution data across managers at the end of every cycle.
Calibration is the accountability layer that favouritism cannot easily survive. When a manager has to present their rating for a favoured employee in front of other managers and HR, and defend it with evidence, the social cost of protecting a weak rating increases.
In lean African HR teams where calibration can feel like an optional extra, make it non-negotiable. A 90-minute calibration session for a team of 20 is worth more to the fairness of the system than any other single investment.
When an employee raises a favouritism concern, HR's response determines whether the system retains trust or loses it.
Do not dismiss the concern as "personalities" or "misunderstanding." That signals to the employee that HR protects the manager, not the process.
Do not immediately investigate the manager. That creates defensiveness and does not usually produce honest information.
Instead, do three things:
Talstack's Performance Reviews module gives HR visibility into rating distributions, the ability to track review completion, and access to the documented rationale behind each rating. When favouritism complaints arise, having that documentation is the fastest way to assess whether the concern is valid.
The halo effect is an unconscious cognitive bias: one positive trait generates positive ratings across all dimensions without the manager realising it. Favouritism can be conscious or unconscious, but it is driven specifically by personal affinity or relationship rather than by a single attribute. In practice, they overlap significantly. Both are reduced by the same interventions: evidence requirements, 360 feedback, and calibration.
This is the most important question to answer in calibration. If the favoured employee is also the best performer, the evidence should confirm it. If the evidence is thin, the rating needs to be revisited. The process should be able to produce a fair rating for a genuinely high performer without relying on personal relationship as a proxy.
Employees should know that calibration exists and that their ratings are reviewed against a consistent standard before they are finalised. They do not need the details of what was said about them in the calibration session. Transparency about the existence of the process builds trust. The content of the session is appropriately confidential.
Favouritism does not disappear through policy. It disappears through process. A performance review system where every rating requires evidence, where peer feedback is collected before ratings are finalised, and where calibration is a non-negotiable step, makes favouritism structurally harder to sustain.
Your best people are watching. They will tolerate many imperfections in a performance system. What they will not tolerate is a system that rewards relationships over results. When that happens, the people who have the most options leave first.
