Product Recommendation Statistics That Actually Matter for Ecommerce ROI (2026)

TL;DR

Only 7% of shoppers click a product recommendation. Those shoppers generate 26% of ecommerce revenue and 24% of orders (Salesforce Shopping Index).

71% of consumers expect personalized interactions. 76% get frustrated when they don't get them (McKinsey).

Personalization leaders lift revenue 10 to 30% on average and capture another 1 to 2 points of margin on top (McKinsey; BCG).

Real Clerk.io customer results: Carlsberg lifted AOV 17% and basket size 24%. Eva Solo grew AOV 125%. Hairlust generated €21k in 30 days plus a 35% lift in repeat-customer revenue. See all three on our customer stories page.

Why product recommendations are no longer optional

Ten years ago, “customers also bought” strips were a nice-to-have. Today they're table stakes. The brands still treating them as a plug-in widget are losing ground to the brands running recommendations as a revenue channel in their own right.

The numbers below come from primary research (Salesforce, McKinsey, BCG, Epsilon, Accenture) and from Clerk.io's own customer base.

The revenue case

1. Recommendations punch far above their weight. About 7% of shoppers click a product recommendation. Those shoppers generate 26% of revenue and 24% of orders (Salesforce). That's a greater-than-3.5x revenue-per-visitor multiplier on the minority who engage with recommendation widgets.

2. Personalization leaders lift revenue 10 to 30% on average. McKinsey's personalization research puts the typical revenue lift in the 10 to 30% range, with top performers landing well above that (McKinsey, The Value of Getting Personalization Right).

3. The margin story matters too. BCG found that retailers using advanced personalization capture roughly 1 to 2 points of additional margin on top of the revenue lift (BCG, Profiting from Personalization).

4. The cost of not personalizing is $2 trillion. McKinsey estimates that companies who get personalization wrong leave $2 trillion in collective revenue on the table globally across retail, CPG, and services (McKinsey).

What shoppers actually want

5. 71% of consumers expect personalization. 76% get frustrated without it. McKinsey's personalization research has reaffirmed both numbers across multiple editions (McKinsey).

6. 80% of shoppers are more likely to buy from brands offering personalized experiences (Epsilon, The Power of Me).

7. 91% of consumers say they're more likely to shop with brands that recognize, remember, and serve them relevant offers (Accenture Personalization Pulse Check).

8. 49% of shoppers have bought something they didn't plan to buy after seeing a personalized recommendation (Salesforce, Connected Shoppers Report).

The Amazon benchmark

9. Product recommendations are estimated to drive about 35% of Amazon's sales. This is the most-cited data point in the category. It traces back to a McKinsey analysis from 2013 and has been recycled ever since. Read it as “material, and well above any competitor,” not as a fresh 2026 measurement (McKinsey, How Retailers Can Keep Up With Consumers).

What this looks like in practice: three Clerk.io customer results

Industry averages tell you what's possible. The next question retailers usually ask: what does it look like when it actually works? Three examples from our customer base.

Carlsberg: 17% AOV lift, 24% bigger baskets, 1.6x conversion

The global beverage brand integrated Clerk.io across its B2B webshop. Behavioral recommendations went on the homepage, category pages, product pages, exit-intent popups, and checkout. Shoppers who clicked a Clerk.io recommendation were 1.6x more likely to convert. Average order value rose 17%. Basket size grew 24%. Carlsberg's global ecommerce manager describes the overall impact as “double-digit growth from baseline, in most cases.” Read the full Carlsberg customer story.

Eva Solo: 125% higher AOV

The Danish home and design retailer paired Clerk.io's recommendation engine with personalized email content. Average order value rose 125%. The win came from connecting on-site behavior into lifecycle email. Recommendations shoppers saw during a browsing session showed up again in the next campaign they received. More results on our customer stories hub.

Hairlust: €21k in 30 days, +35% repeat-customer revenue

The hair-care D2C brand uses Clerk.io across 12 domains. Search and recommendations alone generated €21,000 in attributable revenue inside 30 days. After automating lifecycle emails, repeat-customer revenue grew 35%. Read the full Hairlust customer story.

These are standout results, not guaranteed outcomes. They show what's reachable when a catalog, a search experience, and a recommendation engine get tuned together rather than treated as separate widgets.

What to do with this data

1. Measure your recommendation revenue separately. If you can't tell what share of revenue flows through recommendations, you can't optimize it. The Salesforce 7% / 26% engagement-to-revenue split is the benchmark to compare yourself against.

2. Move beyond static “related products” rules. Rules-based widgets can't adapt to in-session behavior. Behavioral engines can.

3. Test cart-page recommendations. Lowest risk, fastest AOV signal. For more on this, see our guide to increasing basket size.

4. Close the loop with email. The behavioral signal that powers on-site recommendations compounds in lifecycle email. Hairlust's 35% repeat-revenue lift came from exactly that loop.

To see what Clerk.io would forecast for your specific catalog and traffic, the ROI calculator takes about two minutes.

FAQ

What percentage of ecommerce revenue comes from product recommendations?

Salesforce's Shopping Index reports that the roughly 7% of shoppers who click a recommendation generate 26% of revenue and 24% of orders. It's the most-replicated benchmark in the category.

Do product recommendations actually increase conversion rates?

Yes. McKinsey puts the average revenue lift from good personalization at 10 to 30%. BCG finds 1 to 2 points of additional margin on top. Clerk.io customer Carlsberg saw shoppers who clicked recommendations convert 1.6x more often than shoppers who didn't.

How much uplift can a retailer realistically expect from product recommendations?

Industry averages land in the 10 to 30% revenue lift range (McKinsey). Clerk.io customers include Carlsberg's 17% AOV lift, Eva Solo's 125% AOV growth, and Hairlust's 35% repeat-customer revenue increase. Those represent strong outcomes, not typical ones.

Why is Amazon always cited in product recommendation stats?

McKinsey estimated in 2013 that around 35% of Amazon's sales come from recommendations. The figure has been repeated ever since. Treat it as directionally useful, not as a fresh benchmark.

What's the biggest mistake retailers make with recommendations?

Treating them as install-and-forget widgets. The lift compounds when placements get segmented, messaging gets tested, and the on-site signal connects to lifecycle email.