It's hard to answer this question without seeing or playing with the data. My suggestion is to start playing with the data yourself. Using Superstore (sample data included with Tableau), I started by creating a view that shows unique orders by customer. Filter this view to show only customers with infrequent orders (i.e. CNTD(Order ID) < 5). Filter the view to the customers that remain. Then, depending on what else you want to report, use order date to display when purchases occurred. I have never considered this question before and I built this (image below) in less than the time it took me to type up this response.
The best advice I have is try a few different ways to present the data before you settle on what works for you and your audience.
Thanks Kent for the impressive explanation.
The problem I am facing is how should I pull customers who don't purchase regularly(using what logic). Suppose last year customer ordered only in May this year he ordered in October.
Separating those customers from the heap is a task. Please revert if you could suggest some logic to separate these customers.
The logic I used in my above example was simple -- for any customer with less than 5 orders over the course of three years, consider this a target customer.
You need to decide on what constitutes an infrequent customer. Is it time between orders? Is it total orders?