Sale series correlation
Let’s examine sale series of two items during a 31 day period.
If you add each row, you will find that:
The sum of sales for item 1 came to 31.
The sum of sales for item 2 came to 28.
How can we use this information to calculate the average daily value of sales of the two items? Everyone understands the logic – take the number of days in the period (31 days), and then divide it by the number of sales (31 and 28 respectively), as follows:
V1 = 31 items / 31 days = 1 item per day
V2 = 28 items. / 31 days = 0.9 items per day
Using this logic to predict average daily value seems correct. Upon first glance, it appears that the average daily value of both items was close. After all, there is little difference between 1 and 0.9. However, if you look closely at the two rows and count how many days each item was sold, the story changes. These numbers are not comparable. The first item was sold on 14 different days; the second on 9. The only conclusion we can draw at this point is that item 1 sells on more days and item 2 is purchased in larger quantities per day.
Why did this happen? There are many reasons. These may be related to demand, cyclicality, or product availability.
If all of these are constants to both items, we can conclude that average daily sales were indeed proportionate, and that any deviation which we observe would be atypical.
However, these three things are not constant, and there are three major occurrences that account for the discrepancy.
- Item 2 is new and not available for purchase before the 15th day of the month. Therefore, of course no sale could be recorded. In this situation, meaningful analysis of sales would begin on the date of the first sale of item 2, the 15th of the month.
- Item 2 is not new, but, from days 1 to 12 it was out of stock. In this situation, the correct analysis would have to be done beginning on the12th.
- This product is not new, is sold regularly, but, we experienced a kind of “drawdown” of demand from the 1st through the 12th. This can be determined by examining stock at the beginning of the period. Here, we can take day 1as the starting point for our analysis.
Let’s suppose that option 2 is the case and it is therefore impossible to do a meaningful analysis of sales until day 12. We then review days 12 to 31 (19 days). By doing the same math we started with, we now find that the average for item 2 is 1.5 items per day (28 / 19 days = 1.5 pcs), which is very different from our initial findings.
Because we get different averages depending on real information about the item, and because any lack of information on the availability of goods at different times can significantly skew the overall analysis of turnover, it becomes clear that a simple one-step analysis will never provide accurate results. Because the rate of sale for each item during the 31 day period involves a variety of factors, a lot of information must be analyzed to determine actual item value
Mycroft Assistant automatically detects the beginning of a period of calculation, and takes into account all nuances of the movement of SKUs. It provides management with accurate sales analysis for the purpose of facilitating the forecasting of necessary resources.
— Demand forecasting – information about it you can find HERE
— Sales analysis – information about it you can find HERE ;
If you want to get started with the Mycroft Assistant – where we have implemented everything necessary for you to work effectively, simply sign up HERE and follow the provided instructions.