Wednesday, September 30, 2015

Amazon launches Uber-like delivery service

Courtesy Business Insider: Amazon has recently launched its own service, called Amazon Flex, where independent individuals can deliver packages for Amazon. All you need is a car and a phone, and Amazon will pay you $18 to $25 per hour.

Although I can't say I'm surprised, I don't expect this to be a big solution to Amazon's high operating costs. Consider once again the issue of peak traffic: a lot of Amazon's sales are in the holiday season. That's also when weather can be bad, and customers want their packages quickly. So, to make Amazon Flex a viable component of their strategy, they'll have to do some sort of surge pricing, which will increase costs. One thing in Amazon's favor, however, may be that more drivers are available (think college students on vacation); on the other hand more people will want to spend time with their family than schlepping around dropping packages. Let's see how it plays out.

Tuesday, September 22, 2015

Employment crunch

This news is about a week old, but still worth commenting on. Ecommerce logistics requires a lot of labor in multiple areas: warehouses and fulfillment centers, picking items from store shelves, and delivering items to customers. However, the demand for this labor is highly seasonal. Additionally, skills required are not particularly hard-to-find. As a result, the situation from the perspective of labor has been less-than-ideal, to say the least, something we've explored here and here.

Well, it turns out that this year, demand for labor is increasing while supply is decreasing. Demand is increasing simply because more and more retail volume is moving to ecommerce. Supply is decreasing because the overall unemployment rate in the US is falling. Basic economics dictates that in such a situation, the price of labor should go up. And that's what we are beginning to see. From the Wall Street Journal:
Employment agencies for retailers and logistics companies say they are having trouble finding warehouse workers to stock early holiday inventory and employees to train for work in fulfillment centers, where holiday orders will be packed and shipped.
Few could have predicted the nation’s unemployment rate would fall to 5.1%, as it did last month, amid such red-hot growth in e-commerce. As a result, retailers and delivery companies expect to have to raise starting pay in some places.
On a related note, this brings to mind something I keep observing in my discussions with industry executives: the role of automation. Already in warehousing there is plenty of automation, and as that increases, the demand for labor will decrease. In delivery, driverless cars and drones may play a similar role. So although right now it looks like there is a labor crunch, it may not last very long as more automation comes in.

Sunday, September 13, 2015

Future of US Retail

There's a fascinating new research paper out from a couple of economists from the University of Chicago. It's titled "The ongoing evolution of US retail: A format tug-of-war", NBER Working Paper 21464, by Ali Hortascu and Chad Syverson. The paper contains a detailed study on the volume of different formats of retail, going back many years, and including features such as average size, employment, etc.

The main point of the paper is the following. Two retail formats have been expanding: warehouse clubs, and ecommerce. This expansion has been accompanied by a decline in traditional format retail stores. However, warehouse clubs have had a much bigger impact than ecommerce. Their main findings are best summed up by the following sentences at the end of the article, which I quote:
While warehouse stores have had much more influence on the sector to this point, ecommerce has had its own effects and may be growing in relative importance. Perhaps this concurrent expansion and strength of ecommerce and a physical format portends a retail future not dominates by either, but rather with a substantial role for a "bricks-and-clicks" hybrid. The formats may end up being as much complements as substitutes, with online technologies specializing in product search and discovery and physical locations facilitating consumers' testing, purchase, and returns of products (A.T. Kearney, 2014).
I completely agree with that part of their conclusion, that a bricks-and-clicks hybrid is the way of the future. In fact, many retailers have stopped separating ecommerce vs. physical sales from their annual reports, as a recognition of this convergence.

The paper focuses on empirical data about the retail sector. It would be fascinating to connect these findings with an understanding of consumer preferences: what is it that customers care about, that has led to retail evolving in this way? Figuring that out, along with understanding macroeconomic trends as in this paper, would provide a much better idea about the future of retail.

Wednesday, September 9, 2015

Shameless self-promotion: big data class

Leaving the ecommerce focus of this blog once again, here's an article published on our B-school news site about a big data class I launched a few months ago. It's sort of like a coding camp in disguise, and is a lot of fun to teach. Here's a brief quote that gives a sense of what the class covers:
The project in the first class was to identify what makes tweets go viral. Students wrote code to download tweets, analyze the text using an algorithm they created, use SQL to find patterns, and display their findings using Tableau.
“This is the kind of stuff done in social media marketing analysis, and the students were very proud they were able to do this with six weeks of learning, many of them starting from scratch,” Sinha says.
Meanwhile, here's an article yesterday on The Economist. Money quote:
As Marc Andreessen, the co-creator of the Netscape web browser, likes to say, in future there will be two kinds of jobs: those that involve telling computers what to do, and those that involve being told what to do by computers. 
And lastly, I cannot possibly not link to that big giant thing that Bloomberg published a short while ago, also titled "What is Code?"

Thursday, September 3, 2015

America's driving habits and ecommerce

Readers of this blog know that I'm bullish on the so-called "click-and-collect" models of ecommerce delivery: customer buys stuff online, but then picks it up in a drive-through facility (see here and here for two recent examples). So far, I haven't used any data to support that stand.

Well, here's some analysis of America's driving habits, courtesy Advisor Perspectives via RedState. The analysis shows that the population-adjusted miles driven grew at a fairly steady rate from 1971 until 2005, after which the recession led to a dip. It's been rising again this past year, as also evidenced by the return of traffic jams and long commute times. On the flip side, the first article also mentions that older people drive less and Millenials are more attracted to a public-transit based urban lifestyle. Additionally, there appears to be no significant impact of gas prices on miles driven, possibly because fuel economies are also increasing dramatically.

So where does this leave us with respect to ecommerce models? I think if the trend of the past year continues and we see more driving, especially with lower gas prices, the case for click-and-collect ecommerce models only gets stronger. Even otherwise, I think there are a substantial number of people who would utilize such drive-through services, much as they do for banks, pharmacies, fast food, etc. A more rigorous analysis is still awaited, however.

Thursday, August 27, 2015

Tauber Institute Video

Here's a nice video describing the organization I now lead. As always, feel free to reach out if you want to connect about this.

[With apologies for going off-topic.]

Monday, August 24, 2015

Wal-Mart omnichannel strategy: ship from store or ship from FC?

We are hearing more and more about omnichannel retailers implementing store-fulfillment models: customers buy online but the products are shipped from retail stores. The purported advantages of this are many: the store network has lower average distance to the customer than the fulfillment center network, so customers can get it faster; ship-from-store allows for inventory pooling between the store and ecommerce channels; ship-from-store enables easier inventory rebalancing when some stores are overstocked and others are short, etc.

However, there are disadvantages as well. The operational cost of fulfilling ecommerce orders is far cheaper in fulfillment centers, because of labor and process efficiencies. Fulfilling from FCs also allows for inventory pooling across different regions that generate ecommerce orders.

So which is better? The answer, of course, depends on a number of retailer-specific factors such as demand characteristics, cost structures, etc. At Ross we're conducting research on this very question, although that is still premature. However, here's some evidence from the country's largest omnichannel retailer about how these costs may play out. From an article from WSJ:
Greg Foran, chief executive of Wal-Mart U.S., said on the company’s earnings call Tuesday that the company has been holding more of its inventory at distribution centers rather than in the backrooms of its stores, a strategic choice that gives the retailer more flexibility in meeting demand.
Wal-Mart reported that inventory grew 2.2% from a year earlier, slower than its sales, which grew 4.8%, or by $3.4 billion, in line with its goal of shedding excess inventory.
Supply chain experts say this is a necessary move as Wal-Mart increasingly embraces e-commerce and what the retail industry calls an “omnichannel” strategy, in which companies make the inventory across all their warehouses and stores available to all customers, both online and offline.
By keeping smaller amounts of a wider variety of goods at distribution centers, Wal-Mart can make products available to customers without having to stock the items in every store, said Kevin O’Marah, head of research for supply chain talent development firm SCM World. And if customers want an item in a distribution center, it can be shipped quickly and more efficiently than if a store employee were to ship it.