I wanted to share some feedback and suggestions about the Account Analytics beta page, which exists for a while and is intensively used by me, even that I have a single website only .
First of all, it is great to get a quick overview of the most important metrics a check for success, attacks and errors, based in peaks and evolution. It’s two minor visual glitches that I suggest to adjust:
1. Use a consistent amount of decimal digits
Obviously the precision is two decimal digits, but trailing zeros are omitted. It’s a question of taste as well, but in science it’s common to show trailing zeros to indicate the precision of the number and align them visually as well in e.g. tables. The
6.5% here along could mean anything between
6.54%, while it actually is something between
6.504%. For the
0% of course it’s worse.
2. Fix interval starts and ends
Depending on the time and day of month, either the start or the end or both of graphs point downwards. In case of last week view this is especially true after midnight and gets better during the day, which means it’s related the the length of the first and last day of the view, that hast only just started. For the 24h view it seems to be the same for part of the first and last hour that has passed, although strangely sometimes only the start or only the end is affected. To not give a wrong impression and fix this, the time span for each point of the graph needs to be equal. It should be pretty easy, when it’s
00:05, to have each point of the 24h view showing the numbers from
X:05 and each point of the weekly and monthly views from
00:05 of the previous and
00:05 of the next day.
Here another 24h view at 20:37, which shows that strangely the relative % number graphs are affected as well. It looks like the first point get’s an absolute number of zero:
It it would be only the graphs, that would be less of an issue, but the numbers are affected as well of course. E.g. after midnight my weekly view shows typically a decrease around 15% of all absolute numbers, which is only due to the one day missing in the interval.