Source: bestbuy.com
Consumers say online recommendations are the worst
Three out of four online global shoppers worldwide fail to regularly see product recommendations that interest them, according to a survey sponsored by SAP.
When asked how often suggested recommendations actually interest them while shopping online, only 25 percent answered either “almost all the time” or “half the time.” By region, only six percent of U.K. shoppers said that they regularly see relevant recommendations with 14 percent in Japan and 17 percent in the U.S. sharing that sentiment.
“There’s a striking gap in what consumers are seeing as they navigate online marketplace,” said Chris Hauca, head of strategy and GTM (go to market) for SAP Commerce Cloud in a statement. “… Having a 360-degree view of the customer will enable meaningful recommendations and unique shopping experiences — something organizations should have in mind as they enter the holiday shopping season.”
The challenges come despite a recent study from BRP Consulting finding that 60 percent of retailers were able to identify customers online before checkout versus 23 percent at physical stores.
A study earlier this year from Segment, the customer data infrastructure firm, further found that consumers have higher expectations on personalization online than offline – albeit disappointed at both.
Seventy-seven percent expect personalization from a “large online retailer” and only 23 percent felt they delivered. For a department stores, 51 percent expected personalization and 17 percent felt they delivered. Forty-seven percent expected personalization from a “big box retailer” and 12 percent felt they delivered.
An Adweek article from earlier this year found personalization expectations online extending beyond product recommendations to content. Brendan Witcher, principal analyst for ebusiness and channel strategy at Forrester Research, said, “All of these companies are trying to do this sexy personalization stuff, but the truth is they haven’t done the unsexy work of capturing, organizing and using customer data to create truly relevant experiences.”
- Online Retailers Are Missing the Mark on Customer Preferences, SAP Global Study Indicates – SAP
- Stores rarely ID customers before they check out – RetailWire
- 73% of Customers Want Order Tracking Across all Touchpoints but only 7% of Retailers Currently Offer “Start Anywhere, Finish Anywhere” Order Capabilities – BRP Consulting
- Study finds online retailers are ineffective at website personalization – Digital Commerce
- Personalization changes everything – Segment
- Personalization Is a Priority for Retailers, but Can Online Vendors Deliver? – Adweek
BrainTrust
Charles Dimov
Vice President of Marketing, OrderDynamics
Laura Davis
Founder, Branded Ground
Doug Garnett
President, Protonik
Discussion Questions
DISCUSSION QUESTIONS: What’s holding back a higher level of online personalization from retailers? Will consumer expectations around tailored recommendations and content likely be higher online versus physical stores for years to come?
Nothing is holding back retailers. The issue is that personalized recommendations are a low-percentage proposition. The most successful attempt would happen when the consumer is looking for exactly the type of recommendation that the retailer is touting at the same time. How likely is that to happen?
I think Huaca got it right: personalization is not just about identifying attributes or characteristics of shoppers. It’s about recognizing intent and context, which is far more subtle to detect. However, there is one surefire way to get around that issue: just ask.
If you think about it, what’s the question that most store associates ask when engaging with consumers? “How can I help you today?” Online, that question never really gets asked. And if it was asked, would the site be smart enough to do anything about the answer? AI gets smarter all the time, but it’s still not that smart. There’s a big difference between “I’m here looking to replace something that broke” and “I’m here looking for a gift idea for my niece.” Even personalization algorithms that are more heavily weighted on current browsing over past behavior aren’t going to be able to keep up with that kind of behavior shift from shopping visit to shopping visit.
Consumers know that online recommendations come from tracking their activities, and that sets their expectations high. I think it’s ironic that physical stores could probably do a better job – if only retailers staffed their stores to actually have someone there to ask the question in the first place.
I would actually like to see the nitty gritty of this survey. From my view as a retail consultant/practitioner, some retailers are excellent at it and others are not. I think the more interesting angle to this study would be to unveil those that are getting the high points for excellence, those that are not and the commonalities of both camps. How about it SAP? 🙂
This is all around the technology capabilities used by the retailer. There are true AI/machine learning solutions in the marketplace today that can provide very accurate real-time personalization that can eliminate irrelevant recommendations. The trouble will be to get the shopper back to looking at the recommendations now that they have been “trained” to ignore them to date.
Curious question given that systems like the Episerver’s e-commerce platform offer various personalization aspects, driven by their AI engine. Others like Salesforce Commerce Cloud, Shopify Plus, Big Commerce — all have varying degrees of personalization too. But it takes some will from the retailer’s side to get the systems interconnected; to make this a reality.
Online expectations will continue to escalate. But if you are in a physical store, there is a higher level of expectation of tailored suggestions. Aside from super busy days like Black Friday, stores have humans who should be able to help customers with their needs, making suggestions and recommendations to help them find what they need. There is no getting over the consumer’s need and want for human contact.
We continue to underestimate the complexity of human interest and overestimate the degree to which it can show up in data.
Let’s remember that personalization has been pursued aggressively for nearly 20 years. And Amazon, who appears to have the most thorough data and AI for making recommendations, still rarely puts anything in front of me that makes any sense.
Let me recommend a couple of things:
Retailers should back off expectations for what personalization will achieve. Amazon must find an increase in sales through their recommendations. But I’ll suggest the primary value they get is an advertising effect — they regularly put new things in front of shoppers. Their personalization likely has only a small return.
It seems to me a situation where simpler personalization will be of far higher profitability than putting in the extraordinary expense to create complex personalization systems because they aren’t likely to drive high sales volume.
To deliver personalized recommendations, retailers have to know a great deal about that customer and the objective of the current shopping occasion. I’m not sure how many customers want to share enough data to make that happen in the nanoseconds that recommendations are calculated and populated on the screen.
Relational database software with the latest and greatest 5th generation appendages are a long way from creating true interactivity outside from the laboratory. Both customer satisfaction and sales growth results demonstrate this quite clearly. In fact the consumer is now practiced to become repulsed by the idea of giving this technology a fair attempt at working for them.
With the root form of computing we have had since the development of present day digital processes, we may need a generation or two to get acceptable levels of willing consumer participation. In the world of information technology (IT) the evolution of categories like hardware and software have many advantages for today’s users. The single biggest disadvantage is when the consumer is made to believe in something like artificial intelligence, which has only proven to be a resident of an IT category called vaporware.
Consumers will always have higher expectations online vs. in-store since the company can leverage customer data and recommend from a broader range of products.
Retailers fail at product recommendations when they try to take advantage of customers by recommending products with higher margins or greater inventory rather than exactly what the data might suggest would be the optimal product for that customer. If retailers continue with that practice, they will end up driving away the very customers they hope to retain.
I agree with Brendan Witcher/Forrester — there are still way too many retailers and brands who have not gotten their data in order. You can’t implement advanced analytics for personalization at scale without proper data management.