P&G’s eStore was prompted by a mediocre year in online sales for the consumer goods manufacturer. According to a recent article by Andy McMains of AdWeek, less than one percent, or $500 million, of P&G’s $79 billion in global revenue last year stemmed from online sales via sites such as walmart.com and amazon.com. Through the eStore, P&G hopes to learn from consumers how to better market their products and improve online sales across all e-commerce channels. One of the main goals of this initiative is for P&G to learn about consumer buying patterns and share this customer intelligence with its retailer partners to boost sales.

eCommerce is unique in that Websites allow retailers to quickly identify consumer shopping patterns, and automatically adapt their merchandising programs based on this insight. This is simply not possible in brick-and-mortar stores. With Baynote’s technology, for example, retailers such as Urbanoutfitters.com and Bluefly.com are able to more quickly recognize even the most unexpected trends, such as consumers purchasing two seemingly unrelated products, and then personalize product recommendations in real-time. This improves the next shopper’s experience, builds loyalty and boosts sales.

Since most products are currently sold via retailers rather than directly by manufacturers, manufacturers are often left in the dark on the valuable data that their retail partners can obtain about consumer behavior. While brands have been able to utilize the Internet well to engage with consumers via email marketing, social networks, online product reviews, and a slew of other activities that require their explicit feedback, they have not had a direct link into the less obvious but often more telling data that ecommerce affords. This includes leading, often termed “weak” signals about consumer intent that can only be gleaned from the online shopping experience. Think clicks and hovers, not hard transactions.

This new online environment should enable P&G to monitor and respond to shifting consumer trends in real-time, and we are looking forward to seeing how their customer engagement initiative will change the overall online retail landscape. We predict that the insight gained from eStore will improve both indirect and direct sales, as P&G will be able to better collaborate with their distributers on why certain products are selling well, and how to improve upon the merchandising and online sales strategies. If managed well, the P&G eStore should also allow the manufacturer to bring innovative new products to market significantly faster based on real-time visibility into the online browsing behaviors of their customers.

While many marketers, website owners, and other business owners already know this, you’d be surprised how many of them use tools that rely on this overused and inaccurate metric. Many analytics, and even recommendation vendors, are relying on clicks as a KPI to either display to their customers or in the case of recommendations, to power their algorithms. The quality of recommendations or any kind of targeting based on user observation is no stronger than its ability to understand when and how a user has succeeded. Bounce rate, as Avinash points out, is just one of those metrics. At Baynote we track dozens of them, and clicks are by far the least utilized when validating success or failure.

Bottom line: When you are talking to analytics vendors, ad networks, or recommendations vendors ask them what KPIs their algorithms are relying on. If they say hits or clicks, they are idiots by Avinash’s definition, and he’s a smart guy.