It’s easier to sell to a customer that is already on your site rather than one who isn’t. And it’s easier to sell to an existing customer than to a brand new prospect who is less familiar and comfortable with your business. But just how effective can cross-selling and upselling be? And how can you do so more effectively?
UPSELLING AND CROSS-SELLING BY THE NUMBERS
- Existing customers have a 60-70% probability of making a purchase whereas new prospects have only a 5-20% probability of making a purchase (Marketing Metrics).
- Upselling drives more than 4% of revenue (Econsultancy).
- 80% of a company’s future profits will come from just 20% of their existing customers (Gartner Group).
These stats clearly show that 1) it’s easier to sell to existing customers and 2) upselling/cross-selling to customers can increase your revenue. Therefore, it only makes sense to upsell and cross-sell to your customers, but giving them bad item suggestions–whether these be physical products, videos, music, etc.–can be counterproductive. Indeed, Bain & Company found that a customer is 400% more likely to go to one of your competitors if they experience a price or product-related problem with you.
HOW TO UPSELL AND CROSS-SELL EFFECTIVELY
There is a wealth of insightful and useful information on the web about how to upsell and cross-sell effectively. Central to this advice are roughly three crucial elements:
- Knowing your overall audience.
- Knowing the preferences of specific customers.
- Knowing your items and their features and how these relate to customer preferences.
These elements allow you to make good product suggestions. However, you won’t, of course, be able to personally recommend items to everyone that visits your site. This is where tools that enable your website to make suggestions come in handy. These tools are called recommendations solutions, recommender engines or recommender systems.
Not all recommender systems are the same. Different recommender systems use different methods to make recommendations to varying degrees of effectiveness. Many e-commerce websites or content management systems include some form of recommender system. For example, they will suggest similar blog posts or products to a user that has some relation to the product or blog post they are already viewing. However, these are usually rudimentary and arbitrary and do not scale well. Often, basic recommender systems make recommendations based on some set of features of the item in question (i.e. blog post, product, song, video, etc.) that is specified by the administrator.
While this sort of system may be effective for a small site with relatively few items and users, it becomes clunky and ineffective as the number of items and users increase. For example, a small e-commerce site with a small number of products in its database doesn’t require recommendation. But as the number of products increase and the customer base expands, the differences grow and often become more subtle and therefore difficult to parse for a rudimentary system. In this case, product suggestions start to seem random, or worse pushy and schlocky, which could alienate potential customers–recall the above stats to understand how damaging this can be to a company’s profitability. In order to make precise recommendations in an environment with a large number of users and items, more sophisticated technology is required.
Without getting too technical, this technology uses advanced techniques like data mining and artificial intelligence to make highly accurate and personalized item suggestions to users. In the hyper-competitive world online business, effective recommender systems aren’t simply nice to have, but increasingly essential. Bad item recommendations can alienate would-be valuable long term customers. In contrast, good item recommendations can dramatically improve user engagement, increase customer retention and increase revenue per customer.