Product Recommendations: Helping Shoppers Find More to Love

Picture a shopper standing in a wonderful physical shop. They came in for a single pair of running shoes, but a friendly assistant notices the shoes and says, "Those pair brilliantly with these cushioned socks, and a lot of runners grab this water bottle too." No pressure, no hovering, just a helpful nudge from someone who clearly knows the stock. The shopper leaves happy, with three things instead of one, feeling looked after rather than upsold.

That quiet, generous moment is exactly what good product recommendations try to recreate online, where there is no assistant standing nearby. In this guide you will learn what recommendations actually are, the psychology that makes them work, the different types worth using, where to place them, how to keep them feeling personal rather than creepy, the mistakes to dodge, and how to measure whether they are pulling their weight. By the end you should be able to picture a recommendation strategy for your own store that helps people find more to love.

What product recommendations really are

A product recommendation is simply a suggestion: a small set of items shown to a shopper because they are likely to be relevant to that person, right now. You have met them a thousand times. "Customers who bought this also bought…", "You might also like…", "Complete the look", "Recommended for you". They appear on home pages, product pages, in the cart, in search results, and in email.

Underneath, recommendations range from gloriously simple to genuinely clever. The simplest is a hand-picked list an owner curates manually, like a staff favourite shelf. The cleverer ones use patterns in shopping behaviour — what people view together, buy together, or return to — to surface suggestions automatically. The word you will hear for this is personalization, which just means showing different things to different people based on what is likely to suit them, rather than showing everyone the identical wall of products.

Recommendations are a cousin of upselling and cross-selling, but the mindset is gentler. Upselling tries to move someone to a better version; cross-selling adds complementary items. Recommendations sit underneath both as the discovery engine: they decide what to suggest in the first place, and they can just as easily help a browsing shopper find the right product as they can grow a basket.

Why a good suggestion feels like a gift

Recommendations work because choosing is tiring. A store with hundreds of products can quietly overwhelm a shopper, and an overwhelmed shopper often does the easiest thing of all — nothing. Psychologists call this choice overload: when faced with too many options, people freeze, postpone, or abandon. A thoughtful recommendation cuts through that fog by saying, in effect, "here are a few good options for you", and that reduction of effort feels like a genuine kindness rather than a sales push.

There is a second, gentler force at play. When a suggestion lands well — when it surfaces exactly the accessory a shopper half-knew they needed — it creates a small spark of delight and the sense of being understood. That feeling builds warmth toward the store, and warmth is what turns a one-time buyer into someone who comes back. The goal, then, is never to trick anyone into spending more. It is to be the knowledgeable assistant who makes the whole experience easier and more enjoyable, so the shopper leaves thinking, "that was a great shop," not, "I got upsold."

The main types of recommendation

It helps to know the handful of approaches stores actually use, because each answers a slightly different question for the shopper. You do not need all of them, and the best stores mix two or three thoughtfully rather than bolting on every widget available.

Behaviour-based suggestions

These look at what large numbers of shoppers do and find patterns. "People who viewed this also viewed…" and "frequently bought together" both fall here. The charm is that you do not have to guess the connections; the crowd reveals them. A camera and a memory card, a dress and a matching belt, a coffee machine and descaling tablets — the data quietly notices what humans pair up.

Personalized-for-you suggestions

These draw on an individual shopper's own history: what they have browsed, bought, or saved. A returning customer who keeps looking at lightweight jackets should see more lightweight jackets, not heavy parkas. This is where wishlists and saved carts earn their keep, because a saved item is a strong, explicit signal of intent you can build on.

Curated and editorial picks

Not everything should be automated. A "staff picks", "new this season", or "complete the look" set chosen by a human carries taste and personality that algorithms struggle to fake. Curated picks are especially powerful for newer stores that do not yet have enough traffic for behaviour data to be reliable.

Where recommendations belong on your store

Placement matters as much as the suggestions themselves. The same recommendation that feels helpful in one spot can feel intrusive in another. The trick is to match the type of suggestion to what the shopper is trying to do at that exact point in their journey.

A quick word on the cart and checkout. This is the most delicate spot of all. A tiny, relevant add-on — the socks with the shoes — can lift order value gracefully. But anything large, loud, or off-topic risks distracting a shopper who has finally decided to pay, which is precisely how you create avoidable drop-off. If you want to protect that fragile final stretch, lean light, and read more on the broader topic in our guide to recovering lost sales after checkout.

Keeping recommendations helpful, not creepy

There is a fine line between "this store gets me" and "this store is watching me". Cross it and you erode the very trust that makes a shopper buy. The good news is that staying on the right side is mostly common sense applied with discipline.

First, suggest things that are obviously relevant to the page, not a random grab-bag. Second, avoid eerie over-specificity — referencing something a shopper looked at once, weeks ago, in an unrelated category, tends to unsettle rather than delight. Third, always leave room for genuine discovery; a recommendation engine that only ever shows variations of what someone already bought becomes a small, boring echo chamber. A dash of "you might not have thought of this" keeps browsing fun.

Trust underpins all of it. Shoppers are far more relaxed about personalization on a store that already feels credible and safe, which is why your trust signals and your honest reviews do quiet double-duty here. When people believe you have their interests at heart, a smart suggestion reads as service, not surveillance.

Common recommendation mistakes to avoid

Even well-intentioned stores undermine their own suggestions in predictable ways. The most common is showing too many. A wall of twenty recommendations does not feel generous; it feels like another overwhelming catalogue, and it dilutes the few suggestions that genuinely matter. A short, confident handful almost always outperforms a sprawling grid.

A close second is recommending things the shopper has already bought. Nothing says "we are not really paying attention" like suggesting the exact item sitting in someone's order history. Equally jarring is recommending out-of-stock products, which turns a helpful nudge into a small frustration. Keep your suggestions tied to what is actually available, an effort that depends on the same accurate inventory discipline behind any good store.

The final trap is letting recommendations grow stale. A "trending now" carousel that has shown the same five products for a year is neither trending nor convincing. Suggestions should breathe with your catalogue and your seasons. Treat them as something you tend, not something you install once and forget, and they will keep earning their place.

Recommendations for stores that are just starting out

If you are early and short on traffic, do not despair that you lack the data for clever algorithms. Curation is your superpower. Hand-pick "goes well with" sets on your top products, build a small "our favourites" collection, and use clear category links. These manual touches often outperform thin automated suggestions, and they double as great content for your email marketing. As you grow and the behaviour data thickens, you can gradually hand more of the work to automation while keeping a human eye on the results.

It also pays to feed your recommendations good raw material. Crisp product titles, accurate categories, and tidy tags help any system — human or automated — make sensible connections. Recommendations are only as good as the catalogue underneath them.

Measuring whether they actually work

A recommendation widget is not automatically a good thing; it is a claim that needs testing. The headline numbers to watch are simple. Are shoppers clicking the suggestions? Are those clicks turning into purchases? Is average order value rising for sessions that engage with recommendations versus those that do not? And crucially, is overall conversion holding steady or improving — not quietly dropping because a busy carousel is distracting people from the buy button.

The honest way to answer these questions is to test rather than assume. Run a controlled comparison: some shoppers see the recommendations, some do not, and you watch which group buys more in total. Our guide to A/B testing your store the sensible way walks through how to do this without fooling yourself. If a placement does not earn its space, move it, change what it shows, or remove it. Real estate on a page is precious.

Connecting recommendations to the bigger picture

Recommendations rarely live alone. They work best alongside thoughtful product bundling, which packages natural pairings into a single, better-value offer, and alongside loyalty programs that reward the repeat behaviour good suggestions encourage. You can even let recommendations follow the shopper into your inbox flows; pairing them with automated email communication turns a one-off purchase into an ongoing, helpful conversation. The story your store tells about its products matters too, which is why strong brand storytelling makes every suggestion feel more like a recommendation from a trusted friend.

Treat recommendations as a living part of the store, not a set-and-forget feature. Revisit them seasonally, retire suggestions that have gone stale, and keep asking the only question that matters: is this genuinely helping a shopper find more to love? If you would like help designing a recommendation strategy that fits your catalogue, our team is happy to talk it through with you.

Frequently asked questions

References

1. McKinsey & Company. "The value of getting personalization right." mckinsey.com.
2. Baymard Institute. "E-Commerce UX Research." baymard.com.
3. Nielsen Norman Group. "Recommendations and Personalization in E-Commerce." nngroup.com.