As we take a look at the monthly total forecast and the distribution by day, there are a couple of things we now need to take into account:
- Weekend days. Many contact centres (especially business-to-business) have a significant drop-off in inbound volume at the weekend. That means that the total monthly volume of a 30-day month with 10 weekend days will have noticeably fewer calls than a 31-day month with only 8 weekend days. So, once we have calculated a monthly total forecast it’s often necessary to make an adjustment for the number of days (28, 29, 30 or 31) and the number of weekend days in the forecast month (8, 9 or 10).
- Public Holidays. An adjustment needs to be made for the number of public holidays in the month, usually decreasing the monthly total forecast. (Christmas in the UK is an example of seasonal variation that is so significant we cover it separately.)
If you have a public holiday in the month you’re forecasting for then you should adjust that week’s shape according to the distribution of calls from corresponding public holidays in the past. If you have adjusted your total monthly volume then all you need to do is re-distribute your volume to accommodate the new shape; if you did not adjust your monthly total then you can simply make the adjustment directly in the relevant week’s daily volumes.
Combining our progress to date, we end up with 1, 2, or more distributions of calls for different customer lifecycles (new customers, billed customers, special events, other,…) which, when added together and adjusted for the number of weekend days and public holidays, will result in a generally good distribution. Daily accuracy is likely to be acceptable in many cases, depending on the accuracy of the original monthly forecast.