A newsagents I was speaking with recently expressed concern about sales. I took a look at their data and saw healthy numbers. Whereas they were looking at data for two weeks, I was looking at data for three months.
Comparing trading periods with less than two weeks data is dangerous. You need at least three months, preferably longer.
In my own businesses I look at three months and then compare over a longer period.
Anything shorter than three months can skew your perspective. Seasons, for example, can impact and this is heightened with a smaller data set.
Comparing the first six months of a year with the year before provides a cleaner, smoother analysis and this is what you need for uncovering real trends being experienced by the business. Looking at two weeks is dangerous.