Shirt.Woot

Fourteen weeks in the new system of picking prints out of the Derby, it seems time to take a look at sales. To compare apples-to-apples as much as possible, this data begins after the first day price increase and includes only shirts that printed on AA in all sizes or on Anvil in all sizes on its first day.

First day sales averages:

First Place Shirts

• Old Derby Rules, Winner, AA: 3,468
  
• Old Derby Rules, Winner, Anvil: 2,702
  
• New Derby Rules, Winner, Anvil: 2,696
  

Second Place Shirts

• Old Derby Rules, Saturday 2nd place, AA: 1,856
  
• Old Derby Rules, Saturday 2nd place, Anvil: 1,218
  
• New Derby Rules, Saturday Editors' Choice, Anvil: 1,240
  

Third Place Shirts

• Old Derby Rules, Sunday 3rd place, AA: 1,978
  
• Old Derby Rules, Sunday 3rd place, Anvil: 1,402
  
• New Derby Rules, Sunday Guest Editor, Anvil: 837
  

So this goes some of the way towards confirming what most people guessed. The new rules haven't had near as much effect on sales as the switch to Anvil. It's possible that the guest editor is selling worse than third would but it's impossible to really gauge that without confidence intervals on those estimates. I'd wager sales of the guest editor pick are pretty variable.

Thanks for the stats - love it!

It's interesting that, under the old derby rules, the third place shirt outsold the second place shirt (on average).

In this case, maybe it has more to do with the day of the week and first day sales - more people online Sundays (home in the evening, getting ready for the workweek) than on Saturdays (at the beach, out with friends, etc.).

TOTAL sales for second place shirts may be higher than total sales for third place shirts, but you then have to take into account the price increase on subsequent days, as well, which will lessen demand.

mrwednesday wrote:So this goes some of the way towards confirming what most people guessed. The new rules haven't had near as much effect on sales as the switch to Anvil. It's possible that the guest editor is selling worse than third would but it's impossible to really gauge that without confidence intervals on those estimates. I'd wager sales of the guest editor pick are pretty variable.

It shouldn't be difficult to pull confidence intervals on the averages, but I confess that I don't know how to do it easily in google spreadsheets (and the data isn't currently in the right format to pull into R - and I don't have the patient to do it by hand at midnight!), so it will have to wait until tomorrow.

I'm also curious if there's a relationship between the placement of a shirt that was the guest editor's pick and its sales (i.e., if the GE's pick was third in the popular vote, did it sell better?). Again, will pull the rest tomorrow. NYCBrent, you have a great point; I have total sales data from any shirt Reckoned before the website relaunch; I'll see if I can find the time to pull total sales for more recent shirts to see how overall sales compare.

Sales of guest editors' pick:
806
629
1876 (Curiosity)
981
637
1201
636
589
127
1431
413
1177
382

ETA: I just realized that I originally included the second and third place shirts from the most recent Double-Take Derby as EC/GE choices. I've reclassified them as "Old Rule" 2nd/3rd place Anvil shirts and corrected the averages in my initial post.

Are you dropping the top and bottom numbers?

I'm taking off my artist hat and putting my "Veteran Wooter" hat back on...

I think the Woot.com redesign has a lot to do with the drop in sales.

First (on the strictly shirt side of things)--we have more side sales, which are going to affect first day sales (in theory, folks only have so much money). Not sure the effect is all that significant though.

Second, and really the focus of this post--I'm going with my gut and wager a large percentage of shirt sales come from people who click over to shirt.woot after first going to woot.com. Woot.com's traffic has been dropping pretty steadily since their redesign, based on site analytics [percentage edited because I can't verify accuracy].

Now when we say "redesign," we're really talking about a change to allow many more sales than "one a day," which I think works awesome for shirt.woot, because it offers unique, original products. But Woot.com built its empire on a model where someone could come in, check Woot, say yes/no to the one thing offered, then proceed (hopefully checking wine, shirt, etc. too). It was quick and there wasn't too much work involved. We saw tons of copycat sites based on this model.

Then the redesign came, and that model kind of went away. Suddenly buyers were faced with 8 different Woot sites, all with a main sale and multiple side sales featuring multiple products. Suddenly Woot became work, and it didn't really feel like a sale-a-day site anymore. And a lot of people who used to quickly click over to Woot to see what was up, stopped bothering.

So in the end, I think the loss of Woot traffic has had a large affect on shirt.woot sales through loss of click-through traffic.

Ha! Providence!
I was running my own numbers on this.
Factoring all the impactful changes (price, supplier, redesign, derby rules), getting correlation data for votes vs. first day sales. Will post when I get to a computer.

Maybe.

ETA: Corellian Freighters!

So I ran some numbers, and will probably get skewered by the stats folks over this, but here goes anyway:

For the period between Dec 25 and The price hike on Jan 17, votes vs sales had correlation coefficients as follows:
1st place: 0.396936131
Second place: 0.94373863
Third place: 0.853031345
Solid numbers, but the sample size was small.

For the period between the price hike and the supplier change, Votes vs sales had correlation coefficients as follows:
1st place: 0.753775014
Second place: 0.770846686
Third place: -0.811301459

For the period between the supplier change and the site redesign, Votes vs sales had correlation coefficients as follows:
1st place: 0.439008415
Second place: 0.353260048
Third place: 0.439294231

For the period between the site redesign and the Derby change, Votes vs sales had correlation coefficients as follows:
1st place: -0.131998924
Second place: -0.083024469
Third place: 0.216202201


For the period between the Derby change and the end of October, votes vs sales had correlation coefficients as follows:
1st place: 0.625493206
Second place: 0.454959914
Third place: 0.39091613

Total all everything since Dec 25 2011 correlation coefficient for votes vs sales is:
0.560799846

I have no idea what it would take to calculate this, but does anyone have the numbers to see if kevlar is right? Something like a comparison of daily tees sold, daily + reckoning + side sales now versus daily + reckoning sales previously?

I agree that the side sales have cut in to the reckoning and daily sales. I suspect that the price increase has spread out the daily sales, since a $3 difference isn't as deciding as a $5 one.

Mavyn wrote:I have no idea what it would take to calculate this, but does anyone have the numbers to see if kevlar is right?

See my updated post.
be gentle.

Sales numbers are first day only.

I'm wondering if there's been a decline in the total number of shirts sold per day, or if the shirt sales are steady, just spread out with the side sales picking up the drop shown in the Reckoning/Daily numbers.

kevlar51 wrote:I'm taking off my artist hat and putting my "Veteran Wooter" hat back on...

I think the Woot.com redesign has a lot to do with the drop in sales.

First (on the strictly shirt side of things)--we have more side sales, which are going to affect first day sales (in theory, folks only have so much money). Not sure the effect is all that significant though.

Second, and really the focus of this post--I'm going with my gut and wager a large percentage of shirt sales come from people who click over to shirt.woot after first going to woot.com. Woot.com's traffic has been dropping pretty steadily since their redesign, based on site analytics [percentage edited because I can't verify accuracy].

Now when we say "redesign," we're really talking about a change to allow many more sales than "one a day," which I think works awesome for shirt.woot, because it offers unique, original products. But Woot.com built its empire on a model where someone could come in, check Woot, say yes/no to the one thing offered, then proceed (hopefully checking wine, shirt, etc. too). It was quick and there wasn't too much work involved. We saw tons of copycat sites based on this model.

Then the redesign came, and that model kind of went away. Suddenly buyers were faced with 8 different Woot sites, all with a main sale and multiple side sales featuring multiple products. Suddenly Woot became work, and it didn't really feel like a sale-a-day site anymore. And a lot of people who used to quickly click over to Woot to see what was up, stopped bothering.

So in the end, I think the loss of Woot traffic has had a large affect on shirt.woot sales through loss of click-through traffic.

Being an English Major and writer I have to concur with this analysis. The price hike and Anvil switch offered big blows to the core community. However, site traffic as a whole feels down and is down since the redesign. I remember the days when the comments section of any given woot site was several pages and filled with frequent commenters and newbies alike.

The ease of commenting and finding stuff on the site has become so cluttered and eye-tiring that it's not quite as enjoyable as before. I love the new side-sales, but admittedly often am short of cash some weeks when I would grab something like for example the Jas book bag. I wish they would last longer than just a week. It's a little too high pressure for me. Heck the Reckoning is all the high-pressure and drama the site needs.

Short version: It feels like things have become too clutter and too rush-rush around here. I don't know...

Bumpity.