Marketers Bet on Recommendations While Shoppers Want Saved Carts

Most marketing teams have a clear picture of what personalization should look like: a feed of relevant products, a message timed to the moment and an AI layer that learns from every click. Shoppers, asked the same question in a matched survey, describe something much plainer. They want a fair price, a cart that remembers what they put in it and a checkout that does not make them start again.

That gap matters now because the holiday quarter is when brands lean hardest on recommendation engines and new AI features. The research released this week suggests the investment is aimed at the part of the experience customers value least, while the friction they notice most goes unfixed.

Two Surveys, Two Definitions of Value

The Wunderkind and Cordial study, released on October 13, put the same questions to 1,000 US consumers and 150 US-based senior marketing professionals. On price, the two groups barely overlap: 56% of consumers cite the lowest price as a leading consideration when deciding where to buy online, against about a quarter of marketers. Half of consumers say discounts are among the strongest factors in whether they notice a brand message at all, while 26% of marketers agree.

Personalization shows the same split. Some 61% of marketers say relevant product recommendations have a major impact on conversion, but only 32% of consumers say recommendations make shopping easier. What consumers credit instead is the plumbing. According to Cordial's full report, which was fielded with MX8 Labs and Dynata and completed in August 2026, 48% of consumers value saved carts, yet only 15% of marketers believe saved carts have a major impact on conversion. Faster checkout scored 48% with consumers as well.

Channels are misread too. Consumers say they prefer email (82%) and SMS (64%), and marketers underestimate both, putting them at 73% and 47%. Social media runs the other way: 74% of marketers emphasise it, compared with 39% of consumers who prefer it. AI is where expectations diverge most. Some 59% of marketers expect AI-powered personalization to have the biggest impact on customer marketing over the next three years, while 16% of consumers say AI currently improves their shopping experience and 27% say they do not trust AI to personalize at all.

Brands Grade Themselves on a Different Curve

The gap is not unique to one survey. Deloitte Digital's personalization research, published on September 29 and based on blind surveys of 1,000 US consumers and 480 director-level and above leaders at large B2C brands, found 87% of brand leaders rating their personalization good or excellent. Only 29% of consumers agree. Brands estimate they personalize 62% of customer experiences, while consumers say 44% actually feel personalized.

The demand is real when it lands. In the same research, 52% of consumers say they are more likely to buy from brands that personalize, and they report spending an average of 30% more with them. But the forms of personalization consumers value most are practical: experiences that proactively resolve an issue (67%), save money (61%) or deliver loyalty benefits (58%). That lines up closely with what Cordial's respondents asked for. Personalization, as customers define it, is mostly about removing effort.

The operating model is part of the reason brands miss. Deloitte Digital found brands take an average of 4.7 business days to activate personalization based on new customer data, and 44% name integration as the top constraint to scaling it. On the Cordial side, 42% of marketers say more accurate customer data would most improve customer marketing performance, and limited customer data and fragmented systems tie as the top personalization barriers at 19% each. A recommendation engine built on a stale, disconnected profile will show a shopper the item they bought last week.

Trust Is Built on the Data Customers Hand Over

The most useful finding for brand teams is about what earns permission. In the Cordial research, 46% of consumers say clear explanations of how their data is used would increase their trust in personalization, while 18% say relevant recommendations alone would. Some 41% value control over their preferences, and 41% want brands to use only information they have shared directly. Deloitte Digital found 58% of consumers value clear and respectful use of their data and 38% want more direct control over how companies personalize for them.

A separate survey from Mews of 3,250 travelers across eight countries, released on October 15, draws the same line in another category. Willingness to let AI draw on data falls steadily as the data gets further from the customer: 61% for stated preferences, 56% for past stay history, 43% for online behavior and 34% for external and social data. Some 57% say they prefer to stay in control rather than let AI act on their behalf, even as 63% would use it to monitor price drops.

“Closing the customer marketing gap requires connecting customer identity, behavioral insights and reliable data,” said Rob Garf, chief strategy officer at Cordial. The surveys suggest the order matters. Customers will accept a more personal experience when it is built on what they chose to share, explained plainly, and aimed at saving them time or money. Inferred interest served back as a carousel is the version they trust least.

Closing the Customer Marketing Gap

• Fix the friction before the feed. Saved carts and faster checkout each matter to 48% of consumers. Audit cart persistence across devices and sessions before adding another recommendation slot.
• Lead with price and value where shoppers look for it. With 56% of consumers weighing lowest price first, make the deal visible in the message itself rather than burying it behind a personalised module.
• Rebalance channel plans toward email and SMS. Marketers underestimate consumer preference for both by 9 and 17 points. Check whether holiday budgets reflect where customers say they want to hear.
• Explain the data exchange in plain language. Clear explanations of data use beat relevant recommendations as a trust driver by 46% to 18%. Put that explanation at the point of sign-up and inside preference centres.
• Build on stated preferences first. Comfort is highest for data customers provide directly and lowest for external or social data. Use preference capture and purchase history before inferred signals.
• Shorten the time from signal to action. An average of 4.7 business days to act on new data means personalization arrives late. Measure activation speed alongside conversion, and fix integration first.

Neither survey argues against personalization. Consumers in both say they reward brands that get it right. The difference is in what counts as right: a remembered cart, a fair price, a clear account of what the brand knows and why. Brands that grade themselves on those terms are likely to find the gap between their scorecard and their customers' a lot narrower.