Is AI-driven shopping curation a good thing?
Source: Amazon

Is AI-driven shopping curation a good thing?

MarketingCharts staff

Through a special arrangement, presented here for discussion is a summary of articles from MarketingCharts, which provides up-to-the-minute data and research to marketers.

A survey of tech-savvy consumers showed that more were confident that AI would narrow (56 percent) rather than expand their options (44 percent).

Interestingly, among those who felt that AI would narrow their options, a slight majority thought that it would limit them from seeing all of the options they would normally explore. Conversely, slightly fewer felt that AI would narrow their options, but in a good way.

The survey from The Integer Group of almost 3,700 U.S. consumers, of whom 3,615 described AI in positive terms, explored AI perceptions.

The findings bring to mind research from L2, which found that Alexa prioritizes Amazon’s Choice products over top-ranked items in conventional search. In presenting just a couple of options for an item, Alexa essentially makes a lot of decisions for the shopper, and the factors used in that curation may not rest simply on high-ranking products.

Nonetheless, there’s more to AI than just Alexa. Overall, 61 percent of respondents in the Integer Group survey felt that AI curation would benefit them either by expanding their exploration and helping them find new things (37 percent) or by narrowing their options in a good way (24 percent).

IntegerGroup Shopper Sentiment AI Curation Feb2018

Thinking ahead five years, the majority of respondents could see themselves letting AI find the best deals on things they regularly buy (some are already asking Alexa about deals), remind them when household items are low, and help make shopping lists.

But fewer than one-third would let an AI assistant automatically buy staples they typically buy, and only one in five would allow it to automatically buy perishable groceries. Likewise, just one-fifth would be comfortable letting AI proactively order items they haven’t purchased but might like.

There are also generational gaps here: Millennials (27 percent) were more than twice as likely as Boomers (12 percent) to let AI proactively order everyday household items for them.

Moreover, Millennials were considerably more likely than Boomers to see themselves letting AI stop them from ordering something because they had reached their monthly budget or because it did not align with their health or lifestyle goals. 

BrainTrust

"There is no doubt AI will influence shopping both today and into the future. How it does so remains to be seen, but I expect it to be a positive value"

Ricardo Belmar

Retail Transformation Thought Leader, Advisor, & Strategist


"There’s always a line with this sort of innovation where it goes from being helpful to being creepy."

Cate Trotter

Head of Trends, Insider Trends


"AI is the latest fad ... too many retailers are making big bets on it despite extensive evidence that it won’t make any large impact on sales volume."

Doug Garnett

President, Protonik


Discussion Questions

DISCUSSION QUESTIONS: Do you see AI-driven suggestions as transformative, somewhat beneficial or potentially detrimental to the online shopping experience? How open will consumers become to letting AI guide their purchasing decisions?

Poll

21 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Mark Ryski
Noble Member
6 years ago

AI-driven suggestions are already transforming the shopping experience for some shoppers and this will only increase over time. While there is no question that there is potential downside to placing all trust in shopping via AI device, as the survey results show, this does not to seem to be a significant issue for younger generations. As time marches on and AI impacts all aspects of life, there’s no doubt that it will play an increasingly significant role in shopping and purchase decisions.

Dr. Stephen Needel
Active Member
6 years ago

Those who are supplying AI suggestions need to consider how the user wants the engine to behave. Some people will want it to remember what they bought before, others will want it to bring up similar items, others will want it to bring up something different, and others will want it to bring up related items. There is no reason why AI can’t be configured for any of these and the user could opt into the mode s/he wants.

Max Goldberg
6 years ago

I see AI-driven suggestions as generally positive, in that the computer might surface items I had not previously considered, but am leery of letting AI into too much of my life. I want proof, through experience, that AI is always functioning in my best interest, is delivering the best prices, and is not subject to outside control — like putting Amazon Choice products first.

Ken Lonyai
Member
6 years ago

AI can be transformative, but there’s an underlying issue almost no one recognizes: AI algorithms are created by humans and machine learning algorithms originate with humans. Humans have bias, imperfect judgment, and make oversights. Humans also evaluate “the effectiveness” of AI/machine learning systems they create. So for the foreseeable future, AI will be imperfect.

That said, humans are responsible for non-AI systems as well. So given the two, I lean towards increased deployment of constantly iterated AI, but with the capability to consider intentional user input and to make personalized adaptations to user wants and needs (“never show me red shoes”) possible.

Art Suriano
Member
6 years ago

I think it’s up to the individual and how they use AI that will be the determining factor. Shopping is shopping … we all begin wanting or needing something. AI is another tool still being developed that will help us make our decisions. There will be times that we will allow it to purchase for us and times when we will not.

Millennials and younger generations will become as used to AI as boomers have become to sending an email. It’s just the next natural step. AI developers will continue to expand their technologies offering more services and conveniences. However, it’s too hard to predict today where it will all wind up. I think we have to sit back, go through a lot of trial and error and watch the evolution take place.

Chris Petersen, PhD.
Member
6 years ago

Retailers have been selectively curating products long before Alexa. Filters on a website enable a customer to curate thousands of products into “buckets that make sense.” Being able to view products by relevance, ratings and price are another type of curation. So brands are already “favored” or “disadvantaged” by how they appear on e-commerce websites. Without the filters product search would be tedious and overwhelming.

AI is automating filters with the power to recommend. AI has the power to take the pain out of the mundane. Most are willing to give up the range of all options and empower AI on purchases of consumables. Not so much for purchasing a dress or computer.

Adrian Weidmann
Member
6 years ago

Artificial Intelligence (AI) will drive (literally and figuratively!) and be woven into all aspects of our lives — whether we like it or not. The competitive, seismically changing retail landscape fueled by the insatiable expectations of the digitally-empowered shopper requires the use of AI into every thread in the fabric of our everyday life. Those brands and retailers that don’t embrace and begin to leverage this technology today will not only be marginalized but will quickly fade away to become footnotes in the global retailing story.

Nir Manor
6 years ago

Usage and adoption of AI-driven shopping curation is growing and the trend will continue the more confident consumers become with technology making simple and less-simple decisions for them. Indeed it is also a generational issue. It is interesting to see the different attitudes of different generations. While Boomers are happier to get help finding things they wouldn’t consider, and don’t look at narrowing the selection as a good thing, Millenials understand the value of narrower selection and are less afraid of “missing out.” One of the reasons for the success of DTC companies like Casper is that it gives fewer options to consumers and helps them decide. Consumers generally seek convenience and once they try and see that AI works well and delivers similar or better results than they would have gotten without it, they will join the trend.

Ricardo Belmar
Active Member
6 years ago

Much of what is shown by the survey results focus on how AI replaces human curation and discovery in some way rather than focusing on new ways in which AI will help shoppers. I would focus on one point — that viewed positively, AI is seen as helping shoppers surface new items they might miss. I see this as how shoppers currently view recommendation engines online but extended by voice assistants. How well these AI-driven assistants will help shoppers depends very much on how the shopper intends to use the information provided. The generational differences are mostly to be expected given other data we’ve seen in recent years about Millennials vs. Boomer shopping habits. There is no doubt AI will influence shopping both today and into the future. How it does so remains to be seen, but I expect it to be a positive value overall.

Doug Garnett
Active Member
6 years ago

AI is the latest fad. And far too many retailers are making big bets on it despite extensive evidence that it won’t make any large impact on sales volume or profit. What evidence? To keep repeating … As a loyal Amazon buyer for 20 years, why can’t Amazon recommend anything useful?

In fact, consumers in this survey may be projecting FROM their Amazon experience. If I were to rely on Amazon to point me toward products my world would be a far less interesting and valuable place.

So why would we expect that algorithms are going to improve that? Retailers should focus on what’s critical. In the words I heard from Tapestry this week at eTailWest: Product, store help, technology that ensures satisfaction….

Doug Garnett
Active Member
Reply to  Doug Garnett
6 years ago

Like, just today Amazon sent me an email recommending I buy … the book I wrote. Yeah … about that….

(And it’s not the first time. I think I’ve been presented with 5-10 emails of this type over the past couple of years.)

Lisa Goller
Trusted Member
6 years ago

Consumers (and retail companies) are drowning in data. AI makes sense of abundant information and helps shoppers discover their desired products faster.

AI is already improving customer service by predicting a consumer’s style with greater accuracy at the exact moment the individual is shopping. AI-driven visual search (used by retailers like ASOS, John Lewis, Shoes.com, Nordstrom, and Urban Outfitters) makes online product discovery more efficient.

If a shopper wants to find the perfect t-shirt, AI helps by tracking the individual’s history of online browsing and buying, and Instagram likes to find suitable options, which can boost sales, reduce returns and improve customer satisfaction.

Consumers and retail companies win because AI makes online shopping more personalized, relevant and efficient.

Nikki Baird
Active Member
6 years ago

I think the important thing to remember is to make sure that consumers understand the “why” behind the recommendation. Retailers — and the tech providers delivering AI recommendations — need to make sure they’re building in transparency, so that consumers can see things like, “Based on these three products that you’ve purchased, we think you’d like these too.” And then take that next step of giving them the opportunity to provide feedback on the logic. This recently happened to me with StitchFix, where I had pinned some fashion ideas for them on Pinterest, and they came back and said “Oh, we see you like lace.” And my first thought was, huh? Really? But sure enough, lace was a common theme among my pins, it just wasn’t why I had pinned them. So that dialogue between the logic and feedback on that logic is going to be important.

Cate Trotter
Member
6 years ago

There’s a wealth of options out there for pretty much anything that you could want to buy. As such, choice paralysis is a very real issue, as is a question of how much consumers are willing to sift through before their eyes glaze over. I think the opportunity for AI to really help customers find the best range of options for them is pretty big. But there’s always a line with this sort of innovation where it goes from being helpful to being creepy. Customers don’t want to feel that they have lost control so I think the uptake of using AI to proactively order things may be much slower. Again though this depends on the application — I can see people allowing AI to proactively reorder them coffee when running low a long way before they let it order them things they might like without their input.

Ralph Jacobson
Member
6 years ago

Shoppers are already letting AI guide their purchase decisions with several retailer and CPG Brand sites, like 1-800-Flowers.com with their “GWYN” program and The North Face with their XPS offering. The interesting thing is that many shoppers aren’t even aware that this is true machine learning in action.

As a retailer from way back, the aspect I appreciate the most is that AI is suggesting the best products and services for shoppers, regardless of how well those items are selling nor how much they may or may not be promoted. AI encourages shopping throughout the assortment/range without prejudice and helps build overall revenue growth via slow-moving item lift.

Jeff Miller
6 years ago

This might be one of the few places where consumers’ “openness” to a new technology may not really matter, since AI is coming and the main idea is that it will improve over time. So consumers may not like what AI decision making algorithms are providing now, but this will not stop the wave of innovation and use.

AI driven suggestions and how things rank will be transformational over the next decade. It will improve the shopping experience, but will have detrimental consequences for certain brands and products who have historically relied on other ways to reach consumers.

Steve Johnson
6 years ago

Curation is only as good as the algorithms. It will work better in some product categories than others. In fashion, we know (from our U.K. research) 30-60% of the average wardrobe is never worn and/or doesn’t look great on the purchaser. Think about this when you adopt an algorithm that recommends clothes based on previous purchases. Oh dear!

My view is AI offers big wins when adopted to improve supply chains. More so than at point of curation. I also believe that too many consultants are hoodwinking retailers that all the data is relevant. It really isn’t. This skill is knowing what matters.

“Data is the science. However the art has always been interpretation.”

Personally, I prefer to use the term “machine learning.” I predict it will be overused and disappoint many. Mainly because it will be cheap to action once the initial infrastructure cost is covered. Fashion retail already averages sending out an email every two days (180 per annum!) to signed up customers. Is that a good thing?

Mike Osorio
6 years ago

We are far more reliant on AI already than we appreciate, from Waze guidance to online machine learning based curation on most of the sites we use regularly. You can always find examples of a newer technology not working as intended (the Amazon examples), but that is a small subset of am otherwise increasingly sophisticated and effective utilization of AI.

I may be a boomer, but I’m right with the younger consumers who will readily allow AI to help narrow my choices to things that reflect my evolving tastes and desires, as well as automate purchases for consumer staples. My belief, based on dialogue with colleagues running businesses like StitchFix, is that the successful retailers and brands will be those that constantly and intentionally listen deeply to the consumer to continuously update their understanding (both machine and human) of the consumer.

Dave Bruno
Active Member
6 years ago

I would caution against casting too big of a net around AI curation based solely upon the Amazon Choice experience. Rather, we might look to the amazing things Sephora is doing with AI and their virtual assistants as another example of positive experiences built upon highly personalized and curated assortments. Through that lens, I am extremely optimistic about the role of AI as a curation tool that adds value to the shopping journey — in the near future.

Vahe Katros
Vahe Katros
6 years ago

This reminded me of that saying: “Products complicate, brands simplify.” Isn’t the assortment or the curation a brand statement?

I guess if your brand is an endless product catalog, cool. But if you are attempting to connect to an idea and an audience then how many factors are involved in coming up with what your show? This is where AI and computers fail relative to humans — the number of variants in making a decision is mathematically a “factorial” — that becomes a computing challenge real fast when the number of options are greater than 11? (not for quantum computers but that’s a sidebar).

AI should be used in the back-end to help merchants find the ballpark, but the last mile should happen in that computer called the prefrontal cortex and in that last mile lies your brand statement.

More watercooler thoughts from the left coast.

Shep Hyken
Active Member
6 years ago

AI suggestions work. “People who bought this, also bought that,” type of up-sell works. It’s time for AI to not only recommend, but also explain why. Once it has that capability, watch how sales will grow for the companies that implement that type of technology. Once consumer get that type of experience, they will not only enjoy it, they will want it, and in time, will expect it.