WPC Smart Recommendations for WooCommerce (Premium)

Recommendations feed on your history, not on thin air

Everybody wants the "customers who bought this also bought that" line the big shops have. What almost nobody mentions is where that line comes from: millions of past orders. Recommendation is not magic, it is statistics about what already happened.

That has an uncomfortable consequence for a new shop. If you have forty orders, there is no pattern to find. The system will do the only thing it can, which is show the best sellers in the same category, and you can achieve that with a hand-made list.

The practical conclusion is not that it is useless. It is that its value grows with your history, and it pays to know where you stand before expecting results. With volume, it works on its own and finds combinations you would never have put together.

How it decides what to show

• Bought together: it analyses completed orders to find which products appear in the same purchase.
• Viewed together: it records which products get looked at in the same session, an earlier signal than buying.
• Category affinity: when there is not enough data, it falls back to what performs best within the category.
• Wishlists and compare: it uses data from WPC Smart Wishlist and WPC Smart Compare if you have them.
• Interest scoring: a background process, hourly, aggregates all those signals into a score.
• Priority rules: by age, price, stock, on sale, featured, category, tag or type, to boost, penalise or exclude products.

Manual rules are what make it useful from day one

Here is the part that makes up for the missing history. While the algorithms fill up with data, priority rules already let you apply what you do know about your business.

Excluding out of stock items avoids the worst possible moment: recommending something you cannot sell. Penalising what has been published for a long time refreshes what gets seen. And boosting what carries more margin turns recommendation into something that pays, not just something that moves traffic.

The sensible mix is rules first, algorithm later, not the other way round.

Where to place them and where not to

They can appear on the product page, in the cart, at checkout, on the thank you page, in popups, in emails, through shortcodes and widgets, as a grid, a carousel or a list.

Being able to does not mean you should. The thank you page is among the best, because the customer has already bought and is receptive without risking the sale. The cart works well with inexpensive add-ons. Checkout is the delicate spot: every distraction there competes with finishing the order, and what you gain in extra sales you can lose in orders that never close.

The popup deserves a separate warning. On mobile, notices covering the content are a practice Google flags as a poor experience and can affect your ranking, so if you use one, make it follow a visitor action rather than firing on arrival.

Requirements and what it does not do

It needs WooCommerce and works with high performance order storage. The score is recalculated hourly in the background, which depends on scheduled tasks working.

It does not do machine learning or genuine artificial intelligence personalisation, however much the word gets used in this industry: it aggregates signals with known rules and algorithms. And recently viewed items are stored in the visitor's browser, not on your server.

In shops with little order history, its effectiveness drops. The maker acknowledges this itself.

Alternatives in the catalogue

If your aim is targeted cross selling rather than discovery, look at UpsellWP or UpStroke, which work with funnels you define.

When the catalogue is large and the customer does not know what they want, asking them sometimes works better: the recommendation quiz does exactly that. And if you sell downloads, there is the Easy Digital Downloads module.

Frequently asked questions

Does it work in a brand new shop?
Partly. The algorithms need orders; the manual rules do work from day one.

Does it use artificial intelligence?
Not in the machine learning sense. It combines purchase and browsing data with rules.

Where should they go?
The thank you page and the cart are safe ground. At checkout, carefully: it competes with closing the sale.

Can I exclude out of stock products?
Yes, with the priority rules, and it is one of the first things to configure.

Does it need the other WPC plugins?
No. If you have their wishlist or compare, it uses that data, but it works without them.

Are recently viewed products stored on my server?
No, they live in the visitor's browser.