SSENSE sales are always the sales I look forward to the most each year. Their selections are huge and there’s always steals to be found.
This sale season I’ve been spending less given I rarely step foot outside, and when I do comfort has taken precedence over style. However, I’ve still been living vicariously through this sale, regularly scrolling through dozens of pages of discounted items.
I’ve always been curious about how SSENSE (and other retailers, for that matter) decide how to discount their items. Do they have a team of data scientists running statistical analyses determining the optimal price for each item? It seems like an interesting problem…
While I have no idea what the process looks like, I thought it might be worthwhile to analyze all of the sale items to see if any patterns pop up. I wrote a simple script to go through all of the items for both men and women - here’s what I found:
Note: Since the data was scraped right when the 80% sale began, some of the original sale items have already sold out. Any sold out items will not be reflected in this analysis.
To start, some overall stats:
|# of items||30,198||17,324||12,874|
|Average original price||$540||$503||$590|
|Average sale price||$247||$228||$272|
I wasn't too shocked to see that Women's pieces are significantly more expensive than Men's pieces. However, I was surprised that there were nearly 40% more Men's items than Women's.
For the rest of the post, the toggle at the bottom of the page can be used to filter graph and table data. For the graphs, hover or tap to see detailed counts.
This graph shows the distribution of original prices. As expected, it's skewed right, with the bulk of items priced $200-$500. The Women's items seem to be slightly less skewed, with more expensive items.
However, the spikiness of the graph shows that there tends to be more items priced with the last 2 digits >= 50. Maybe there's some sales psychology going on here?
Next, let's look at the distribution of discounts. We can see that the majority of discounts is between 60 and 65%. It's also interesting to note that the distribution of discounts is nearly identical between Men's and Women's. This might suggest that whatever method used to generate discounts does not take gender into account.
Now that some initial exploration has been done, let's dig a little deeper.
Explore the chart below - search for your favorite brand, sort the different columns, and set the item count threshold.
We can see that Unravel, Etro, and Neil Barett were the heaviest-discounted brands with at least 100 items, while Alyx, Phillip Lim, and A-COLD-WALL* were the least.
While it’s interesting to see these brands, it’s hard to draw any conclusions. One personal hypothesis is that discounts are negatively correlated with brand performance, in which case this data may give an indicator of which brands are in highest demand. However, that would mean that for Men, Maison Kitsune, Carhartt WIP, and Y-3 are some of the best selling brands, which would definitely surprise me.
Next, we can look at a breakdown by clothing type.
Many of the most discounted types were winter-related types (coat, parka, scarf, gloves, etc.), which is to be expected for the summer sale. The most heavily discounted types were pet leashes, with an average discount of 62%. I guess people's sartorial choices don't trickle down to their pets... This is pure speculation, but I wonder if the reason why blazers were also in the top 10 most discounted items for men was because of the current global COVID-19 WFH situation.
Interestingly, the least discounted items were all non-clothes (i.e. accessories, shoes, etc.).
|Clothing Type||Average Discount||Count|
Another interesting cut of the data is to analyze the sale by color.
Reversible, burgundy, taupe, and purple were the heaviest discounted items - maybe an indicator that some of these colors are becoming less popular?
We see that silver, gold, rose, bronze, etc. are some of the least discounted colors, mainly due to jewelry/accessories being relatively less discounted.
Lastly, while looking at the data I noticed that there were a handful of "SSENSE Exclusive" items. Since these items are presumably more limited than other items, I wondered if they might be discounted differently than normal items.
However, it turns out that these items on average were more discounted than non-exclusive items. While the sample size is small (fewer than 1000 exclusive items), this seems to be an indicator that SSENSE exclusive items don’t tend to be anything significantly more desirable than normal items.
|SSENSE Exclusive?||Average Discount||Count|
Hopefully you found this analysis interesting! While I still don't know how the SSENSE sale process operates, it's been fun to do a bit of digging to try to understand the sale more fully. There's a lot of data to play around with in this post, so hopefully you can find some takeaways I missed. Please reach out if you did!