Will Predictive Tech and Same-Day Delivery Revolutionize Retail?
In the often-referenced sci-fi movie "Minority Report," Tom Cruise walks into a Gap, is instantly recognized and receives tailored purchase recommendations. But what if predictive technologies accurately anticipated his needs before he ever got to the store and delivered those items to his home?
A Wired article by Marcus Wohlsen explores how the rollout of same-day delivery by Google, eBay, Walmart, Amazon and others coupled with algorithms that support personal digital assistant apps, such as Google Now, could lead to such a utopian commerce scenario in the future.
Speaking to Wired, John Sheldon, global head of strategy, marketing solutions, eBay Enterprise, termed the concept "ambient commerce," describing its basics as "consumers turning over their trust to the machine."
He offered a few hypothetical examples, such as the Nike+ mobile app ordering a new pair of shoes after the runner has completed 300 miles. A sensor could determine when a bike helmet is damaged enough to require a replacement. In a more advanced scenario, a shirt with a sensor detects rain, prompting a courier to deliver an umbrella or rain jacket at the user’s exact location.
"The ability to both sense a need and fulfill on that need in that compressed a time frame is very, very powerful," Mr. Sheldon told Wired.
In some ways, ambient commerce appears to rely heavily on automation. Such technologies are exemplified by the popular IFTTT (If This Then That) app, which automates small tasks between internet-connected services. IFTTT recently expanded into home automation with SmartThings, which, among other things, could set an air conditioner to turn on once a front door is opened.
But ambient commerce also relies on predictive technologies. That’s exemplified by Google Play, which gathers information from a smartphone user’s e-mail, calendar, location and social accounts to send alerts and recommendations before the user even knows that they needed them.
While Mr. Wohlsen pointed to the complexities of accurately depicting needs, not to mention conducting wide-scale same-day delivery, he still thinks the concept may have some merit.
"Between what these companies know about our interests, our friends, our whereabouts, our purchases, and anything else we’re willing to feed them, whether by e-mail, Twitter, Facebook, GPS, or credit card, they probably should have a very good idea of what we want and when and where we want it," wrote Mr. Wohlsen in Wired.
Discussion Questions
How could the combination of predictive technologies and same-day delivery influence retail sales performance in the future? What scenarios make sense for how predictive technologies may work their way into retail in the near and long terms?
I’m picturing spending a good part of my day declining items that an algorithm has delivered. We are nowhere near the level of accuracy required of such a system. Nor am I likely to give such a system access to my bank accounts and credit cards to do this. I’m not seeing it.
This is where the future is going. Customers should be able to subscribe to a variety of items – diapers, ink cartridges, motor oil, and a host of other predictable purchases. Allowing them to opt into regular replacements of consumables like shoes also makes a lot of sense.
What’s undetermined yet is how technology will help us do this. But it’s clear that brands will want to lock in future purchases and demonstrate an appreciation of customers’ loyalty and lack of free time.
Interestingly, I am involved in the formation of an AI ambient computing project, with one of the principles being no commerce/marketing/advertising allowed.
John Sheldon makes some intriguing hypotheses that I don’t think will quite come to fruition, as much as some brands would salivate for it to be so. Most consumers do not need to be told when it’s time to buy replacement running shoes and likely would react negatively to the invasiveness that implies.
Sensors are “the thing” that the press latched onto after this year’s CES, but they have existed (in modern terms) at least 100 years. So, yes they will be incorporated with more advanced algorithms, and real-time data to anticipate behavior and try and sell more stuff to more people. However, a lot of these prognostications have more to do with willful thinking than true research into shopper habits through the lens of user experience. Either people will draw a line regarding how far they want their wallets invaded, or we will be witnessing the true dumbing down of civilization.
If perfected, it would be a true game changer.
Assuming it worked perfectly, for example, there really would be no more need for stores as we conventionally think of them. If Gap, for example, knows what you want essentially before you do, why wouldn’t they just ship it to you?
But … predictive technologies don’t work all that well because purchasing is – in many respects – an irrational activity. Would Amazon really like to institute a program where it automatically shipped a book a week to you based on its algorithms? No, because the return cost and process would be staggering.
Making these technologies really effective would require being able to access an individual and a household’s total purchase histories, tracked in real time.
And, nobody has access to that kind of data … well, O.K. except maybe for the N.S.A.
Without that real-time access to metadata, predictive technologies come down to slightly more interesting, high-tech forecasting tools – and we all know how well those work.
Using algorithms to suggest items in which I may be interested is one thing – even a welcome service to point me in the direction of new items. However, I can review the items or not as I am interested, purchase or not. Having those items delivered presumes I want to buy them.
What if the sensor determines that my favorite shirt needs replacing and the new item is not the same? What if the recommended item is based upon the purchases I made for a gift for someone else? The recommendations are welcome. The presumption that the consumer wants to buy the item by automatically delivering it, eliminates one level of choice and may not be well received. Making the recommendation known to the consumer along with the option of one day delivery with a one-click checkout may be very attractive.
It’s all about personalization. In it’s simplest form, Walgreens has an automated system that calls me when my prescription is about to run out. Add deliver to that and you have a winning combination.
Tracking the habits of a customer can help you deliver a personalized/customized experience. This is a higher level relationship that depends on the customer being loyal to you. Loyalty goes both ways. The customer commits to you, and you commit to delivering what the customer needs (or wants), when they need it.
The combination of these technologies is attractive on the theoretical side, but on the practical side, I’d choose to forego the convenience, and to stay in control of my own shopping experience.
Predictive technology is predicated on (1) Technology being able to accurately predict the needs of humans (2) willingness of humans to trust that technology and its prediction, to spend their money without an explicit approval.
Given that we humans and our activities are inherently unpredictable, the prediction is going to be a hit-or-miss (miss mostly) and hence the second step isn’t happening just yet. For now predictive technology will primarily function as “Alerting” technology, alerting users that they potentially need something and hoping the user will agree and “initiate” the 2nd step.
I completely agree with Dr. Needel. We are nowhere close to having the accuracy required to deliver a quality experience.
Today, even retailers with the best systems and supply chains are having issues with predicting demand and keeping core items in stock on the shelf, let alone same-day delivery.
Accurate predictive technology at a shopper level requires that the consumer disclose a lot of personal identity information. With the recent NSA disclosures, there is already a consumer backlash on what they are willing to disclose.
While a lot of these scenarios seem far-fetched or too difficult to imagine being effective in everyday life, the technologies enabling them are moving fast. Looking back more than six years (before the introduction of the iPhone and the start of the smartphone revolution), who imagined all the things we’re now doing on our phones?
That said, it hardly ever goes as predicted. (I’m sure there’s a good Yogi Berra quote for this situation.) Technology changes are usually accompanied by alterations in the business model. For example, one of the reasons apps became possible is that cellular companies moved from the walled garden approach.
With almost everything being instrumented – through the internet of things as well as the many personal uses of the smartphone – there will be a lot of data along with ways to test and check the effectiveness of predictive technologies. Google Now is a great example, and that is still in the very early days. Combine that with incredible logistics capabilities and same day delivery is a real possibility for many more customers.
The effect on traditional store retail will be huge, especially for consumables where the purchase patterns will be readily available. As already mentioned, the consumer can still be in control through a reminder and a one-click purchase (or decline) process. The key is relevancy, timeliness and perhaps some sort of simple, convenient user-in-the-loop feedback system.
Imagine a web browser predicting what website you want to see every time you start up. How exciting does that sound?
Now imagine a search button or a mobile voice search button that allow you to input what you are looking for or interested in and return results.
Minority Report is a nice movie, but the future of retailing will resemble the real world of a customer providing input to an interface (via voice or data share) and retrieve relevant results.
The more things change, the more they stay the same.
The world of predictive technology is old but young in retail, as it is applied or leveraged with shopper information. Being “young” in retail tech, it is at a stage where it will be abused. Say what? In retail, when we have a new, retail-applicable technology, all kinds of new ways and stupid ways to use the technology appear. Some ideas settle and become a reality and others are just dinner talk at retail tradeshows.
The secret behind predictive tech is the shopper or customer. Using predictive in ways that they see helpful and letting them control it will be the winning application of the science. Let the shopper ask “if I buy these pants, what are the best tops to match based on new spring fashions?” Call it the Fashion Advisor. Let the shopper leverage the predictive tools and predictive will fly beyond the really old movie’s application – like “Minority Report.”
The social networks will make sure that the right predictive based apps fly fast….
Tom…the Majority Answer.
Part of this I liken to subscription services, which is very up and coming in many retail categories. That is a service that will prove itself to be a real winner and already has, in some cases.
As for using truly and purely predictive technologies, we are not there yet. A true understanding of the human mind and thought process would have to be captured for this to work in mass.
Don’t get me wrong, there are some scenarios that would work. The Nike shoe, for example. Say you are an athlete, yes, if the tread of your shoe could be monitored and an order placed, this would be super cool. Same day delivery is not, however, a must here.
For likely higher-end categories where this type of cool factor might WOW, I can see some retailers scoring points and perhaps loyalty. I say higher end because I am assuming there will be some higher costs associated with making this happen. IMHO
Traditionally, the retail and CPG industries have not enjoyed the luxury that other industries, such as airlines and hospitality have in that their customers make reservations for their purchases to give them advance notice of incoming product and services demand. Retail and CPG don’t typically have that luxury, however, newer technologies are changing that game. These capabilities will have virtually limitless opportunities for application in both retail and CPG.
Same day delivery will become the expectation in the near future, so merchants need to leverage that bandwagon ASAP.
I like shopping for the things I want. It’s enjoyable. For decades in the supermarket business we rhapsodized about helping shoppers to automatically replace the pantry items they use the most. We would know their preferences and their purchase cycles for basic items and then offer them discounts on these items at just the right time.
It is and will remain a pipe dream. There are too many choices and reasons to change purchase behavior, and no algorithm can keep up with shoppers in this context. I like to try new things, and most shoppers do, also. Imagine allowing a clothing store to dress you, like the example of Tom Cruise in a futuristic Gap store. How fast can you say “no?” We are not Betsy McCalls (man, you young’uns are really going to have to go to Google for that one – here and here.