June 7, 2012
In my post yesterday, Put Product Recommendations in the Hands of Merchandisers, Not Machines, I talked about the importance of making sure a human element exists when implementing product recommendations on your website.
Let's now look at three specific pain points you can expect when launching product recommendations, if you don't take advantage of some of the new tools available. If done correctly, you will be able to minimize risk, make smarter decisions, provide your customers with the very best shopping experience, improve key metrics, and help to maximize ROI.
1. Speculative Impact. Unless you can test your recommendations against a control group, it might be hard to judge the success of automated recommendations. In a perfect world, you are currently merchandising your cross-sells manually and have ample data against which you can judge automated recommendations.
2. Lack of Information and Input. Product recommendation engines should not live in a black box. The merchandiser or marketer needs to be able to set rules, judge results, and make the appropriate level of changes to the recommendation engine.
3. Limited and Inflexible Page Real Estate. This pain point might be one of the biggest to overcome. But if you are able to, it will change your world. Ideally, you want the installation of your product recommendation to be the only time someone from IT touches the project. Beyond that, the best-in-breed recommendation engines will allow for the merchandiser or marketer to place recommendations virtually anywhere on your website.
Implementing product recommendations should not be a painful process, especially if you consider these four features that I think are critical to any product recommendation engine:
• Transparent Algorithms with Controls. New tools remove the black box and give control back to the merchandiser/marketer.
• Intelligent Segmentation. Allows the merchandiser to set rules regarding targeting; for instance, recommend new products to returning visitors and all products to new visitors.
• Easy Testing with Decisive Reporting. Stop leaning so hard on your opinion or that of the HiPPO (Highest Paid Person’s Opinion), and let a test against a control show you the way.
• Rapid Code-Free Deployment. Remember when I mentioned not needing IT after the initial installation? It’s not a dream, it is a reality with the latest generation of recommendation tools.
Robert Gilbreath manages the Ecommerce Marketing and Analytics Department for Calendars.com and DogBreedstore.com. Robert is responsible for the online/traditional marketing channels, site merchandising, social media, and reporting/analytics. He has over 10 years of ecommerce and online marketing experience. Before joining Calendars.com, he was GM/Brand Manager at AcademicSuperstore.com/JourneyEd.com.