Best advice I can give you: use column or bar chart instead of your pie.
Some reasons:
- pie charts take up more space to convey the same information
- in CRM, pie charts are always drawn to fill the space available, so there is no sense of relative scale between one pie and another, or one day and the next.
- you have to show a legend for a pie chart, taking up more space and relying on working memory, because your colours have no inherent meaning (if you were charting only cars by colour, this would be different, but there is nothing in the human brain to equate cars with red, trucks with green etc. In fact what I learnt at school is that all lorries are either red or yellow, according to the old rhyme...)
- if your chart one day has no buses (for example) for the data set being shown, then all your colours shift along so that bikes are shown in yellow and there is no blue segment. So now you have confused anyone that did learn the arbitrary colour scheme, and reinforced the need to keep referring to the legend. I refer to this problem as "palette shift". It can be fixed in CRM by using multi-series charts, but pie charts can only have one series so you cannot avoid this problem.
- using multiple colours assumes everyone can discriminate them clearly. Since about 8%* of the adult male population suffers from deuteranopia (typically red/green "colour blindness") you might not meet your country's disability discrimination laws such as DDA in UK, 508 compliance in the US(?) if some of your workforce can't use the software you provide them to do their jobs. (*varies by ethnicity, but a reasonable average)
- our brains are not wired to compare arbitrary angles very well (we can compare angles of slope above the horizon and tell if they are different, but that is a slightly different trick). We are amazingly good at comparing even very small differences in height (eg column chart), and only slightly less good at doing this with horizontal lengths / widths (eg bar chart).
So in summary: a column chart only needs one colour, does not need a legend, labels the columns directly (and does not matter if one is missing in today's data), saves on space, requires very little working memory, can be compared to other similar column charts (eg vehicles in two regions, if you use a fixed axis for both), and can be interpreted pre-attentively, requiring far less brainpower and much less likelihood of mis-interpretation.
