Monday, February 17, 2014

Missing Flowers and Valentine’s Day Optimization

The four-day period leading up to Valentine’s Day is the busiest time of the year for the flower business. This year this period also brought the heaviest snowfall of the winter season. Everyone I know has stories of flowers that didn’t get delivered last Thursday and Friday.

Some flowers will get delivered late, though that somewhat misses the point of Valentine’s Day flowers. But many won’t get delivered at all, and will go bad sitting around warehouses and depots. The worst estimate I came across was that 50 percent of Valentine’s Day flowers would arrive late or not at all. More likely the rate of non-delivery is lower than that, maybe between 20 and 25 percent, but that is still a huge loss for the merchants involved. Worse, for the customers, it turns what was supposed to be a kind gesture into an aggravation.

The weather is being blamed, but the rate of missing flowers is so high that it hard to believe they all could have been delivered on schedule even with perfect weather. Recall how many Christmas deliveries were late in spite of relatively benign weather in the middle of December. I believe this is a case of optimization gone awry. There are customers who want to mark a holiday with a minimum of effort and attention. Combine that with florists and delivery services that want to minimize costs while delivering the maximum number of products and packages possible. With so many unrelated players pushing at the same narrow window of opportunity, any little problem will result in failures and lapses.

Optimization is one of the marvels of the late 20th century, but the cost of so much optimization is a general loss of robustness. Push optimization too far, and you find that little things go wrong and cascade out of control. We’re just discovering this in the last few years because of the economic squeeze brought on by recession. The financial pressure leads businesses to try to dial things a little tighter than they can reasonably go. This leads to an increased risk of catastrophic failure. That, I believe, probably happened here. The reputation of the flower business is damaged not just for the people whose flowers went missing, but for a much larger number of people who heard about the problems. After Valentine’s Day 2014, anyone who heard the same stories that I heard will think twice about leaving Valentine’s Day flower delivery as a task for someone else to do. With the compounding effect of word of mouth, future years’ business could decline by more than the 20 percent of orders that went missing this year.