Need help with approach to gift card optimization


#1

Apologies in advance, statistics and data analysis isn’t my speciality. Currently doing commercial strategy with a small, private retailer to expand the business and looking at gift cards as one of those levers. My challenge is I’m trying to determine the optimal discount % to offer on the gift card depending both the overall value of the card and how many are purchased. I can offer gift cards at any value (i.e. $100, $200, $500, $1000) and as the value gets higher, there will obviously be fewer that get purchased. In theory the discount % on the lower value cards would be less than the higher value cards but I’m not sure that’s true (i.e. would I rather have 1,000 cards worth $100 purchased at 30% off or 10 cards for $1,000 at 50% off). The inverse is true for volume as the discount % goes up (for any value card) the number of cards purchased would increase. There really isn’t any constraint on the overall number of gift cards I can offer. So I’m trying to determine a relatively simple way of proposing the optimal discount % for each value to help the sales agents use as a guideline with the overall objective being maximized profit. I’m not even sure there is an easy way to model this out since there is no constraint on the number of gift cards I can offer? Really appreciate any help, suggestions or guidance…so out of my depths!!!


#2

@feldozz

You can use mathematical optimization for your problem and if there is linear relationship in the model and constraints then you can use linear optimization by using linear programming.
https://web.stanford.edu/group/sisl/k12/optimization/#!index.md
This is the link for your reference, you can pursue different free MOOCs available on net for better clarity.


#3

Hi Ryan (@feldozz), an interesting question indeed.

So here is how I’d approach this problem. Let’s say you have a particular region A where you offer multiple gift cards (SKU from hereon) of different values such as $100, $200 or so. As per your question, my understanding is that you need to find out that whether a current discount on this SKU is optimum or not, if not what should be the optimum discount percentage? (Hope I’ve understood it correctly).

If I were you I’d have tried to create a data sheet that would’ve included the current Region-SKU combination matched with past discounts and sales volume of SKU that had been offered on this particular combination. For example, let’s say “A” region currently has $100 SKU with 30% off, and it’s sales volume in last month was 100. Now, I’d have tried to place this data against the discount percentage (from past data ofcourse) that had fetched me highest sales volume. Let’s say in last March when I offered the same $100 SKU at about 20%, this had fetched me about 120 in sales volume.

Doing so, will give you a data sheet with all your SKU-Current %offer-Sales volume, placed against past best discount %age and sales volume. From this exercise, you can decide whether your current SKU offer is doing right or not.

If you can do it for different regions, you’d also realize that there are few regions that are insensitive to the discounts, and increasing discounts from past (in hope of increasing sales) had actually no effect on your sales.

Please note that you need to match the current SKU with all the past discount percentage that had been offered and sales volume corresponding to that discount.

Let me know if this will helps. Feel free to ask any question that arises.

BTW you can also predict the sales volume once you have this data ready :slight_smile:

Peace.