Multi agent systems are becoming pervasive in modern industrial and service applications. Either being autonomous robotic agents or human beings performing a collaborative task, relative and/or absolute localisation of the team agents is a mandatory task to be efficiently solved for any meaningful application. In this paper, we will analyse the cooperative localisation problem from the perspective of the number and quality of the available measurements, rooting this analysis to a practical example of ranging measurements for robot localisation. We will make use of a known distributed Kalman filtering technique to highlight a trade off between number and quality of measurements, which is a known effect in estimation theory but never analysed for distributed estimation problems. What is surprising is that the net effect is quite similar to what happens in opinion formation for social groups, which is sometimes referred to as majority effect.