Monday, November 23, 2015

Machine Learning on Azure

As part of my role in the Tauber Institute, I'm excited to be teaching the application of Machine Learning techniques in operations. Although there are many cloud-based providers of this out there, I have to give special kudos to Microsoft for their AzureML platform. It makes ML remarkably easy and accessible (which of course has its dangers as well), and I think they've created a system with very high potential. I'm waiting for their documentation to catch up so that information about the details is easier to find. I'm looking forward to bringing this to my students tomorrow!

Thursday, November 12, 2015

Competition for locker space

Amazon's locker program is now in its fourth or fifth year, but it turns out that Walmart is also trying this out. From today's WSJ:
7-Eleven Inc. is making space for more lockers at a number of its North American stores, in a bet that growing e-commerce volumes will help drive Slurpee sales.
The company has added lockers where customers can pick up packages from FedExCorp. and United Parcel Service Inc. over the last year, and announced in October that it would install Wal-Mart lockers in six locations in Toronto as well. The additions mark a significant expansion in scope of a program first piloted with Inc. in 2011. Now, any retailer that ships via UPS or FedEx has the lockers as a delivery option.
I have to say, I didn't expect Walmart to move towards lockers. Walmart has approximately 4000 stores in the US, compared to 7-Eleven's approximately 10000, but one would think that Walmart's existing network is extensive enough. I suppose the difference is that Walmart's stores are not in the dense urban areas where ecommerce volumes are growing fast, while 7-Eleven's most definitely are.

I'm also surprised that Amazon does not have an exclusivity agreement that prevents 7-Eleven from partnering with other companies. This raises many interesting questions. Are we going to start seeing competition for locker space within 7-Eleven? Are other companies with networks of many urban locations going to start hosting lockers? Are we going to have locker banks, where many different retailers can deliver stuff to individual lockers?  

Friday, November 6, 2015

What if drones delivered all ecommerce orders?

We're seeing so much hype about drones, so my research assistants and I decide to do a little hypothetical calculation. Suppose all ecommerce purchases in the US was delivered by drones. Of course this is a very rough calculation, but roughly what numbers could we be looking at? Here's some output from our research; email me if you want the details.

Roughly 1 trillion pounds of "stuff" is sold to consumers every year in the US. This includes food, apparel, electronics, health and beauty stuff, household goods, etc. Roughly 6% of retail sales in the US is ecommerce; in terms of weight, let's be even more conservative and say 5%. That's about 50 billion pounds of stuff per year, or 136 million pounds per day. Say a single drone can deliver one 5 lb. order in 30 minutes, and is delivering for about 12 hours a day (with the remaining 12 hours for charging, maintenance, loading, etc.). That equates to a little over 1 million drones doing deliveries at any point of time. For comparison, at any point of time in the US, there are between 1000 and 5000 airplanes in the air, and about 250 million cars in the US.

[--- With research assistance from Megan Graham and Sujay Shetty.]

Tuesday, November 3, 2015

How Do You Want Your Product? An article on the different ecommerce supply chain approaches that companies use to get more of their products to their customers.

Ecommerce is an evolving medium of trade, and producers across various industries are competing with each other to find the best solutions to best meet the desires of the consumers while minimizing costs to themselves- this magic combination hopefully resulting in higher profits.  With this goal in mind, companies have taken different approaches to delivering their products to their consumers.  Each of these approaches has its own various virtues and disadvantages, from the perspective of both the company and the customer.  As a result, companies have had to effectively strategize when venturing upon the largely unmapped terrain of ecommerce.  The purpose of this article is to analyze these approaches to ecommerce supply chains, and see the degree of success they have had within the context of consumer items such as electronics, houseware items, and/or clothing.  Below you will find a chart with ratings of “H”, “M”, or “L” - meaning “High”, “Medium”, or “Low”- from the perspective of both the company and the consumer along with selected justifications providing our rationale behind these ratings.     

Selected Justifications:
Company Value of Item Pickup and Delivery by Contractor:  
We classified the company benefit from this arrangement as “high”.  The reason for this is that services in which pickup of items and delivery of these items is handled by a third party at no big costs to the company is a great deal for the company.  It gives the company an additional e commerce avenue for getting their products to customers- providing an additional source of revenue at little cost. Contractors, like Instacart, offering these services may treat the product purchasers as their clients, so fees are collected from the store customers.  According to Juggernaut, Instacart revenue comes from charging customers a delivery fee or fees from the Instacart Express program, along with markups of 15% or more on store prices.   This means that the company does not have to pay for the expensive “last mile”.  Furthermore, store employees will not have to be diverted from their store duties to pick-up and prepare items for delivery to the customer.  This kind of setup provides a convenient option for customers who may have otherwise have visited the store.  Furthermore, “customers” are still coming into the store. The hired “shoppers” that contractors (like Instacart) use, are potential customers themselves to the store.  Furthermore, according to Fortune Magazine, it is possible that Instacart is also receiving revenue from selling consumer data from customers who opted-in with interested companies.  This means that at a premium, companies have the ability to still have knowledge about their customers and make informed marketing and sales decisions despite not being the direct seller to the customer.    

Customer Value from Pickup at Locker:
            We classified the customer benefit from this arrangement as “medium”. The lockers are placed in convenient locations, so the trip to go retrieve the product is not normally an extra one. Another benefit for the customer is the fact that they can pick up the package on their own time. Where this differs from regular package deliver is the freedom of time allocation that the buyer receives. They are now able to go about their day and simply pick up their package from a location convenient for them at a time that is also convenient for them. The downsides for the customer in this situation is that they may still be waiting for relatively long periods of time for the arrival of their product, and they have to leave their residence to receive the product.

Company Value from Customer In-Store Pickup:

            We classified the company benefit from this arrangement as “medium”. For the company, the last mile of delivery is the most expensive one. The reason that having the customer pick up their items at the store is so beneficial is the fact that it does not cost them anything for the last mile. An employee must simply go and pick up the package and get it ready, which does cost time. The last part that makes it overall beneficial for the company is the impulse buying aspect. When the customer picks up the item curbside or buys offline, it ruins the possibility of impulse buying. However, when the customer has to come inside and get their item, they are more likely to purchase something that beforehand they had no intentions of getting.  However, the company still has to divert employees or specifically assign employees in filling ecommerce orders, which can get expensive in a store not optimized for filling ecommerce orders.

Company Value of Delivering Items Themselves:

We classified the benefit of company delivering items through their own delivery systems as “low”. There is much extra cost and time involved in this vector. The company must purchase a fleet of delivery vehicles, hire workers, and plan an affect delivery system. In the extra time this costs on the company's side, whether it is analysis on the process, working with employees, or simply delivering the products- the time compounds quickly. Since other companies, such as UPS, specialize in delivery and have convenient options for large companies, that is normally a better option due to the ease and convenience of service.

Customer Value of Delivery by Contractor:

We classified the benefit to the customer of delivery by contractor as “medium”, with the caveat that this depends largely on the relationship and delivery systems contractors have with the companies whose products they are delivering. This system of delivery is faster than delivery by the company or by a delivery agency, as a contractor is able to be more flexible with it’s delivery times and can implement more customized delivery schedules.  With the possible increase in delivery speed, however, prices may be more unpredictable with possible events such as “surge pricing” occurring during high volume periods.

--- Megan Graham and Sujay Shetty

[Megan and Sujay are freshmen at UM, and are working with me on ecommerce research via the Undergraduate Research Opportunity Program. As always, everything published on this blog is my responsibility, so please direct questions to me. ---Amitabh]

Monday, November 2, 2015

Robots and drones

While Walmart and Google are also now testing drones, here's news from Slate/Business Insider about a land-based autonomous delivery vehicle.
The co-founders of Skype, Ahti Heinla and Janus Friis, are building an autonomous, self-driving robot that will deliver up to 20 pounds of groceries for $1.50 in under half an hour, starting in 2016. 
Twenty pounds for $1.50 in half an hour would be extremely economical! As the article says, these robots don't have to be dispatched all the way from the store; a bunch of them can be dispatched from a mobile van that arrives with multiple orders in a neighborhood. Of course, this technology has its own challenges in terms of what kinds of locations it could operate in, etc. However, in terms of economics, I think it will beat drones pretty handily, simply because of the carrying capacity.

Wednesday, October 28, 2015

Free shipping is still preferred

A long-running trend continues to hold: Another survey by Deloitte also finds that customers prefer free shipping to getting the product quickly. This is something we've commented on before as well, along with our mystification on the rush by retailers to implement costly quick delivery systems (here and here, for example). There's an interesting quote on InternetRetailer's coverage of the Deloitte article: 
One way for retailers who have both an online and physical retail presence to grab a slice of that pie is to offer a buy online, pickup in store option this holiday season for online purchases. 43% of shoppers surveyed say they’re likely to use the perk, with 67% citing the ability to save on shipping charges as a key reason for doing so and 35% saying they’ll shop for other items when they go to pick up their online purchase in store.
What I find interesting is that the article calls Buy-Online-Pickup-In-Store a "perk".  I think of it more as a cost-saving and competitive-edge-gaining move by the retailers: the expensive last mile of delivery is shifted to the customer, while the ability to have the product available quickly offers a competitive feature against's fast delivery methods. It's also noteworthy that such a large fraction of customers expressly state their interest in buying more stuff when they go inside to pick up items.

Wednesday, October 21, 2015

Continued growth of curbside pickup in groceries

Readers of this blog know my perspective: the future of ecommerce in the US is in drive-through pickup, as supported by our research. In the grocery market in particular, there is tremendous room to grow in this direction. Well, it turns out that the major retailers are not staying idle in this direction. Wal-Mart is expanding their pilot service to five more cities. Kroger piloted in Cincinnati, and is expanding as well. So is Giant Eagle. Safeway is testing. As the following quote in the WSJ article about Wal-Mart says,
“The U.S. is a car-based country. Convenience for many Americans isn’t waiting at home for something to show up,” said Bryan Gildenberg, chief knowledge officer at Kantar Retail, a research and consulting firm.
I couldn't agree more. Amazon, are you listening?