We’re now going to look at getting a little more advanced, for which you may need stronger analytical skills within your forecasting team. We’re going to look at customer propensity to call according to their position in their customer lifecycle with your business.
First, we need to draw out the major customer lifecycle touch-points of an average customer of your business. These are often some or all of the following:
- Sell: You sell something to a prospect
- Provide: You provide the goods and services
- Bill: You bill the customer
- Collect: You receive and process payment (and chase non-payment)
- Fix: You fix customer problems / equipment / items
- Maintain: You conduct maintenance activities
- Change: You respond to a customer’s request to make a change
- Leave: You process a customer who wishes to leave you
Your own business may only have one or two of these, or might have more. What matters is that you can identify appropriate customer touch-points in your own business.
If your organisation has a Customer Relationship System (CRM) that you use to log customer interactions with the call centre, it should provide a rich source of information about when and why customers call you, particularly if your agents are required to log every interaction and not just those that require some system change to be made.
Let’s pick one of these touch-points: Provide. It should be possible to analyse the CRM data to work out how many customers contact the centre on the day of Provide (day 0), the day after (day 1), and so on. You can then build up a profile of when customers call you in relation to the day of Provide; you’ll usually find some spikes in the profile relating to particular customer activity, and also that at some point there ceases to be any discernable difference between one day and another (it might be after day 3, day 7, day 20 – it really depends on your business and your processes). In many businesses there is a very noticeable propensity for customers to call within the first few days, which means that the overall call volumes will be fairly sensitive to the number of new customers your business acquires. That means that if your sales figures fluctuate quite a lot then you could well see that mirrored in your call volumes. Analysing your data in this way means that you can more rigorously build your forecasts around planned fluctuations in sales demand.
This process can be repeated for other customer touch-points. For example, consider the behaviour of a customer who receives a bill and then calls you up. It should be possible to see what those customers are doing by looking at the records in the CRM system. Some customers will be wondering where their bill is, so in this case it’s necessary to look at the calling profile from, say, 5 days before they should receive their bill to around 5 days after. So, if you know what your billing profile is in terms of the days or weeks of the month you should be able to predict specifically the calls likely to come in in response those bills, and when they are likely to come in.
Other customer touch-points like Change are harder to predict because your organisation may be less in control of the decision that led to the customer’s call being made. A customer can phone you at any time to change their product, address, or bank details or other parameter in their relationship with you. Being less in control means that it is harder to be so scientific in building a prediction system when compared with things like Provide and Bill above. In many cases the modelling for certain contact touch-points can’t really get much beyond the basic system we looked at earlier.