“Consumers who viewed this product, also bought that product…” Amazon’s been doing it for years, and for them and other ecommerce pros who are getting it right, product recommendations are a very effective way to beef up shopping cart sizes and pull in more revenue per customer.

This specific form of online targeting, however, hasn’t always been so easy for marketers to achieve. The good news is that with new automated solutions in play, they can now easily execute similar types of cross-sell and up-sell campaigns on their site.

As they venture into what may very well be new territory (despite being relatively commonplace), marketers should avoid making the following five common mistakes that plague product recommendation campaigns:

Mistake #1: You try to manually set all the recommendations rules

Until now, teal-time targeting and recommendations have been considered too difficult for marketers to implement because only rules-based models/technology existed. And let’s face it: we’re only human. It’s nearly impossible to predict all possible behavior combinations and rules necessary to target and tailor to shoppers needs and wants. Think about the number of products you hold. Now, multiply that number by the number sold per month (or even per day) and the amount of site traffic you have… The mathematical logic here is enough to make anyone’s head spin.

But now automated solutions exist that empower marketers to achieve this level of targeting without creating complex matrixes and rules, or bogging down their web developer and IT resources. When it comes to recommendations, your best bet is to find a software that automates the process for you, that employs a predictive model that learns and adapts about consumers behaviors over time, that can perform testing within your campaigns and of course, easily integrates with your POC system. Make it easy on yourself—the ROI rewards alone will be worth it.

Mistake #2: You don’t test your recommendations

Anyone who runs a complex ecommerce site has probably already done a fair amount of A/B testing and Multivariate Testing to test user journeys, shopping carts, search, buttons…and anything that impacts conversion rates and the user experience. You feel you’ve got content optimization down and are ready to move on to the next big thing: product recommendations.

But a common error most marketers make is forgetting to perform testing on the actual recommendation campaigns. The truth is, just like every thing else on the site, the “how” and “where” of the product recommendations’ presentation, including their content and design—is just as important as the recommendations themselves. Your site and consumers will always be evolving, and therefore testing and optimization is a continuous process for all important conversion elements.

Mistake #3: You recommend products that are out of stock

Next to a poor site experience, there’s nothing more frustrating than getting a customer interested and excited about a product only to find that it’s unavailable. Bottom line: Don’t kill your recommendations efforts by pushing a product that’s no longer available.

This is an easy one to avoid: your back-end POS system should carefully track stock quantities and this knowledge can be easily integrated with your product recommendations technology to ensure that no out-of-stock items show up in your campaigns.

Mistake #4: You don’t integrate product reviews

Next to price, product reviews have one of the biggest impacts on customer buying decisions. Whether you’re selling a hotel room or a cell phone, consumers want to be reassured about what they are buying—and will go looking for reviews wherever they can find them.

If you haven’t already, test out placement of product reviews in the buying phase and make it easy for the consumer to retrieve the information they want about products buy hosting product reviews from both a previous buyer and editorial perspective. Don’t risk them wandering away from your site and abandoning the check-out process.

One caveat here, though: leave the product reviews for the pre-shopping cart phase. Once your customer has already performed the “Add to Basket” or “Book Now” click, don’t distract them with information that isn’t imperative to entering credit card details and hitting a complete button.

Mistake #5: You offer a distracting or competitive product in the shopping cart

Your recommendations technology offers you insight that allows you to predict products your visitors might be interested in, and therefore, by default, products they either wouldn’t be interested in or those that will distract them. Distracting products are those that compete with known preferences.

For example, if your technology recognizes that a visitor has added a Blackberry to his or her shopping cart, it should not recommend a DROID. This type of recommendation not only interrupts their buying thought process and make them reconsider the purchase path they’re already on, it eliminates the opportunity to offer them complimentary items, such as a case or headset at the time of purchase. These recommendations increase cart sizes and overall revenue, rather than disrupt the flow of the purchase.

Another recommendation that might work here would be a “Buy one Blackberry, get a second for half-price” deal.

Bottom line: your recommendations campaigns are meant to enhance, not distract.