How Automated ‘Nudges’ Help Retailers Make Better Merchandising Decisions
There’s a growing trend today in big data and merchandising based on big data-driven “nudges,” which are critical for driving and improving merchandise decisions for retailers. These machine-learned nudges alert sellers about opportunities to avoid going out of stock, add a product that’s selling and sharpen their prices to be more competitive.
Nudges have also extended into the consumer electronics world, where products exist that remind people to do things on a daily basis (e.g., wake up, attend a conference call, pick up groceries, meet up with a friend, take 10,000 steps, etc.). Consumer products like smartphones or the more recently launched Apple Watch are constantly providing nudges with advice for managing one’s life and improving health.
Our lives are, to some extent, ruled by these nudges. In his 2015 letter to shareholders, Amazon.com's Jeff Bezos wrote that the retailer provides a steady stream of big data-driven nudges — more than 70 million in a typical week. These nudges translate to billions in increased sales for sellers on the marketplace.
Whether in the B-to-B or B-to-C world, we’re all seeing an emergence of big data-driven nudges based on sophisticated algorithms processing data and attempting to influence meaningful micro-decisions to create impact. Big data analytics are driving nudges across a whole of host of merchandising decisions retailers make on a daily basis. Let’s look at a few examples:
- Assortment: Merchandisers synthesize data outside and within retail information systems to provide very specific nudges like “keep product X,” “drop product Y” or “add product Z.”
- Regional: Retailers learn about what's trending locally by analyzing marketplace seller data (e.g., a surge in demand for snow shovels or air conditioners).
- Online marketing: Algorithms process and synthesize sentiments from consumer product reviews, and marketing teams are nudged not to spend too much on marketing for a particular product (i.e., if the reviews are poor).
- Pricing: Changing prices for products is based on a whole host of variables provided by nudges, such as competitors’ prices, stock quantities and projected demand.
- Product content: Product description elements can be modified to improve traffic and/or conversion (e.g., change meta tags or product descriptions, add more reviews and better images, change a title, have more backlinks on a page, etc.).
- Promotion: Changes on the timing of promotions are suggested via nudges, and merchandisers can use this to create personalized promotions.
- Seller: Retailers are able to make better decisions about which sellers to add to maximize sales.
Big data-driven nudges translate to billions in increased sales for sellers and increased earnings for Amazon. There's no doubt that these nudges are on their way to becoming mainstream, and some of them will even be automatically acted upon by machines.
In the retail world, however, there will still be many nudges that require the interpretation and judgment of a human category merchandising manager. The nudges should help them become more informed and efficient in their jobs.
In conclusion, the future is just around the corner. I will leave you with a nudge of my own: it’s time to start preparing now for this new paradigm of data-driven merchandising decisions.
Mihir Kittur is co-founder and chief commercial officer for Ugam, a leading next-generation data and analytics company.