In anyone's book, that's a pretty sweet deal.
If you are already a customer of their organic bazaar then it was a no-brainer. Even if not, the free money concept it offered, I'm sure, was compelling enough to persuade non-regular clients to take the offer and go out of their way to visit the chain (after all, everyone has to eat).
From the outset I knew it was going to be a fast selling offer. What I was interested in knowing, however, was just how rapidly the coupons would sell. I decided to track it to find out. A similar offer back in January 2011 with Amazon.com resulted in a sale of 1,301,296 coupons $20-for-$10. Would the Whole Foods deal break this record?
The offer went live at 8:00 am PST and by 8:15 am I'd started to hear about it through my facebook friends.
I had to scramble to write code to grab the page and write the results to a database because I did not want to miss datapoints. About 10 minutes later I was up and running.
(Writing code under intense time pressure is an exercise in compromise. The "Good" side of you wants to write neat code with error trapping and no assumptions about any returned values, the "Dark" side of you wants to simply hack it out and get it live knowing full well all the limitations. I'm proud to say that the Dark side won today's battle and MacGyver code was used successfully to generate the data for this blog post.)
I sampled the page every minute, but it appears the refresh rate that livingsocial updates the statistics of coupons sold is once every 15 minutes. I continued at my higher frequency sampling throughout the day incase they adjusted, but their updates remained constant.
Below is a chart showing uptake of the offer. Time is shown on the x-axis, starting at 8:00 am PST on the left. The y-axis shows cummulative sales.
The offer was really popular.
It took just over an hour (78 minutes) to sell half a million coupons! This volume of traffic, understandably, stressed their servers. Early on in the morning I was receiving error pages instead of sales figures (I wish I'd grabbed a screen shot of the error page, if I recall it contained a picture of an ice-cream cone and words to the effect "oops, that should not have happened!"). This is the reason I'm missing a few data points on the left of the graph.
But let's not be nasty, credit has to be given for their sustained through-put. To issue 500,000 coupons in 78 minutes required an average of 107 transactions per second. Hats off to their development and operation staff!
As the morning wore on, the transaction rate reduced slightly, and at 4:20 pm (according to the clock on my PC) the magical cap of 1,000,000 sold vouchers was reached. (I have to admit, this was a bit of a shock to me. It was only after seeing the sales peg that I saw the offer was over. I can see, obviously, why they put a cap on it, but I was secretly wanting to see just how high it would get!)
There's no argument that offers like this generate a huge amount of viral traffic; I'm sure everyone reading this got notified of the deal from a least two friends in their Facebook feeds, or Tweets. I'll wager that some of you also got sent details of the offer via email from your Luddite friends, but is it a cost effective way of spending a marketing budget?
The few million dollars that Whole Foods invested in this scheme is no small chunk of change and could have purchased a considerable volume of traditional adverts, keywords and airtime, but nothing beats word-of-mouth. I'm pretty sure well over a hundred million people have been touched by today's campaign and that's starting to look like an efficient CPM. Plus, everyone with the voucher will spend at least $20 in the store, and there's a fair chance they'll spend considerably more, or like what they see and come back again if they were a first time visitor this time.
(Also, don't forget that when Amazon did their un-capped 1.3 million coupon sales, just the month before they had invested $175 million in LivingSocial, so exactly what funny-money went on about the deal is probably not widely known.)