The Grinch Bot Before Christmas: A Security Story for the Holidays

It was Christmas 2017 and I needed one thing: a Fingerlings. There was just one problem: Not a single store had these tiny stuffed animals in stock.

That's when I found Bill from Chicago, who was offering 30 Fingerlings on eBay. The price was 20% higher than it would have been at a toy store, but like so many other consumers I was stuck. I made the purchase, with a slight sense of aggravation at overpaying but also a sense of relief. It was done.

It wasn't until months later that I realized I'd been had — by a Grinch Bot!

This might sound like a page out of the retail Nightmare Before Christmas, but it's a real story about the difficulties of buying a product online and an example of the effect that bots have on the ability of consumers to buy products — and of manufacturers to control their market access, pricing power, and supply chain.

Machine learning-powered bots and algorithmic purchasing are becoming a gray-market economy. Aided by hacker cartels and fronted by online middlemen, the trend is affecting consumers and business buyers as well as manufacturers and brick-and-mortar retailers.

The Sandman Clothing and Sports Company
Another real-life example, this one drawn from a customer engagement, describes the challenge a large US sportswear manufacturer and online retailer faces daily.

The company — we'll call it the Sandman Clothing and Sports Company — manufactures a range of sportswear and associated gear, distributing the products through its own retail outlets and also through agreements with national and local retail channel partners, brick-and-mortar sporting goods stores, and online sales through a network of e-commerce partners. The enterprise has long led its sector with big-splash releases of clothing and other sports items endorsed prior to release by high-profile social media influencers. By the time an item is released, so much buzz has been created that the item is bound to sell out.

Enter the Grinch Bot.

Over the period of a year, the company realized each new product release followed a pattern: The second a new item was made available, it sold out, before sufficient physical stock could be shipped by the company. In an instant, inventory had virtually disappeared from the primary market. Hours later, the item would show up on the secondary market — for example, Craigslist — at significant markup. Retail channel partners were furious — their loyal customers were swarming the stores and venting frustration. Online channel partners watched helplessly as the newest desirable product popped up on other e-commerce sites, offered by sellers with little track record or feedback. The company was losing money not only in its own retail properties but also from channel partners; its supply chain was unable to adapt quickly enough to meet the unprecedented demand.

It quickly became apparent that the company's analytic data, which should have provided a control on orders, was not acting as an early-warning system for suspicious purchasing behavior.

How It Works: Buy, Hoard, Sell
Where once you purchased only from a certified retailer, new markets are now ruled by hoarders who use bots to amass goods. How do retailers fight back? By deploying machine learning and advanced security analytics to quickly detect bots and protect inventory so the downstream brand is not affected.

But wait — not all bots are Grinch Bots.

While it's easy to think all bots are malicious, nothing could be further from the truth. According to the "Automated Threats to Web Applications" project report from the Open Web Application Security Project, more than 73% of bot traffic on the Internet involves, directly or indirectly, humans directing "good" bots to manage processes: search engine crawlers, API access, and system health checks. The remaining 27% represents the activity of "bad" bots and web application risks.

It's unlikely the average consumer would recognize most of the activity generated by bad bots, which puts pressure on businesses to detect, deflect, and defend against the predations of bad bots.

Protecting Your Reputation, Taylor Swift-Style
It's possible to fight bots with a good strategy. Taylor Swift's recent Reputation tour is an example of managing a brand — and the value of a ticket — to maximize value and reduce risk of bot-powered scalping and other threats.

Swift, who lost money to bots and scalpers on her 2015 tour, worked with Ticketmaster to offer her Reputation tickets only through the verified fan program, an AI-driven platform that requires buyers to register for ticket access in advance, then doles out access to tickets after examining a fan's social profile and other data, using breadcrumbs from the fan's web activity. "Verified fans" are sent a code that gives them limited-time access to acquire tickets. Swift sold up to 50% of her tickets via the platform, then released the remainder at market price. According to reporting in the Financial Times, this approach limited the availability of tickets on the secondary scalping market to 5% for the Reputation tour, as opposed to 30% for Swift's earlier tour. Swift came out of the tour with "about $1.4m a show that had been lost to resellers on her previous tour" as measured by a Financial Times analysis — recouping more than $50 million.

Beating bad bots, not only at Christmas but all year long, requires vigilance, real-time access to analytic results, and a laser focus on security. For businesses like the so-called Sandman Clothing and Sports Company, the benefits are clear: more control of products as they move through the retail chain, renewed pricing power, fewer supply chain surprises, and better relationships with retail channel partners and customers. The key to this transformation is data — data the company had in its systems all the time but is hidden by slow analysis that reduces visibility into operations.

Related Content: