If there is one thing that is constant for retailers, online or offline – it’s change. Changing products, consumer tastes, and new trends.
With relentless pressure on the bottom line, change seems like the enemy. For example, when merchants can’t foresee change in consumer tastes they experience stock outs or are forced to sell off extra inventory at a deep discount. Retailers that can spot trends as they emerge, versus months or even weeks after the early signals are able to make the changes in inventory in time to make a difference to their business.
So, why can’t retailers see these trends clearly today? One big problem is that many are looking at the wrong data. When thinking about what’s going to be popular, online retailers look at clicks or what’s selling.
Unfortunately, clicks don’t really indicate interest and with single digit conversion rates, purchase data is a very small part of the picture. Looking at what people are searching for, engaging with and spending time on can actually be a much better way to predict potential interest than purchase data. Even better, engagement data like this doesn’t require any personally identifiable information (PII), eliminating any privacy concerns.
Using engagement data, retailers can see trends in sales before they happen. Consumers will tend to spend time on items of interest before they purchase. In fact, Baynote’s recent study about the holiday shopping experience showed the overwhelming majority of online shoppers conduct research before making a purchase.
As a result, search traffic to ecommerce sites will pick up in areas or around products that are becoming hot. Likewise, a decrease in engagement and search traffic looking at particular products can signal a downward trend long before sales really start to fall. And let’s not forget, that the goal is to consummate a sale. That means offering the right product at the right time, when that customer is finally ready to buy.
Tracking engagement data allowed one US-based online apparel retailer to experience dramatic increases in sales volume and average order value. Since this retailer sells limited quantities of highly priced items and wants to move inventory frequently, time is money. By looking at how its shoppers were engaging with new SKUs to immediately identify their interest, and creating affinities between new products and other items, those hot new products could be dynamically recommended across the site.
This greatly reduced the lag time between new product introduction and the time the item sold out: a net positive for inventory turns. New inventory frequently offers a retailer their best shot at maximizing profit and revenue. By highlighting new products in cross sell and upsell recommendations, this retailer was able to move new items faster which pumped up order volume and order size.
Other non-apparel retailers don’t have it so easy. In another case, a home goods retailer sells unique one of a kind items. As an online only retailer, initially they felt their only choice was to manually merchandise their catalog due to its eclectic nature.
So how can you possibly recommend products and create affinities when every product is so different? By studying how their shoppers were engaging with different products they were able to develop contextual connections between products in surprising ways.
While the company worked to fine tune their automated recommendations, they realized that their manual merchandising system was also creating additional load on their database servers. By moving to an automated system that could track users, products and context in real time, they saw page load times decrease at the same time they offered their customers a more engaging experience on their website. So even in tricky situations like this, the ability to track engagement across the whole catalog can yield terrific business benefits.
Lesson learned: with increasing revenue goals and tight margins, engagement data is one of the keys to effective inventory management. If retailers can understand what consumer tastes and trends are before they impact purchases, then they can use that data to get head of the curve.