Company and Business Model Overview
Groupon is a fascinating case study of one of the fastest growing companies in the world that is now trading at 85% off its IPO price. Armed with the bold vision of bringing the $3T+ local commerce market to the Internet, Groupon’s business model entails connecting local businesses and local consumers. Its value proposition is simple: for merchants, Groupon can get butts in seats like no other form of advertising (online or offline) can; customers can discover great businesses on Groupon at unbeatable prices. Founded in Chicago, Groupon started by offering just one deal per day per city and drove massive user adoption as users scrambled to invite new users to the platform to secure that day’s deal. However, as Groupon scaled, its operating and business models fell out of alignment.
Operating Model Misalignment
As Groupon outgrew its daily deal roots and aspired to become a marketplace for local commerce, it ran into misalignment on three core dimensions: product, process, and culture.
When Groupon was pushing one deal per day, there was no marketplace element to the product. There was no selection for users to browse (past deals of the day were no longer purchasable) and no focus on personalization since all users received the same deal. In shifting to a marketplace model, Groupon sourced more merchants to the site, personalized a feed of deals for a given user to purchase, and built a search experience. It overindexed on the “Buy” node of the Local Commerce Process flow highlighted in Exhibit 1. Yelp was the leader in “Discover” as it had ample merchant reviews to jumpstart a user’s local commerce journey. Square and Paypal were leaders in the “Redeem” step as they had point-of-sale systems that tapped into merchants’ transactions data. Not thinking through the user flow holistically left Groupon in a difficult position to serve as a one-stop-shop for users.
Another contentious point is the practice of discounting. The very name “Groupon” implies users will receive a discount, and indeed Groupon operates very efficiently in sourcing 40-80% off discounts. However, discounting limits the scale of the marketplace. Good merchants who are doing well don’t need to discount, so Groupon appeals to early-stage or struggling merchants. This offering in turn, attracts price-sensitive customers, setting off a vicious feedback loop (see Exhibit 2 below). Further, discounting has the potential to tarnish the image of high-end brands. Breadth and depth of selection are obviously critical pillars to success in marketplace design, and Groupon as constructed today, largely features 3-star-Yelp-equivalent merchants.
Groupon’s go-to-market strategy involves pushing curated deals to users via email (and now mobile). When Groupon started with just one deal per day per city, push distribution via a single daily email was great as it drove purchases without inconveniencing users too much. As the company moved to its marketplace business model, its old operating process of push distribution was not a great fit. To become a marketplace, Groupon had to supplement its push distribution with pull distribution. It had to source a broad selection of deals with density within a particular geo-locale, build a robust search experience, and train users to consume Groupon in an entirely new fashion. While Groupon recognized this paradigm shift was needed, it proved difficult. As mentioned in the Product section, lack of discovery and limited selection muddled Groupon’s ability to communicate to its users its new battle cry of “Check Groupon First.”
Exhibit 3: “Check Groupon First” ad campaign (1)
Headquartered in Chicago, Groupon initially took on a sales and marketing culture. As the company scaled, it built up a strong engineering, product, and design presence, with many of these tech hires joining the new Silicon Valley office. Chicago was proud of its “feet on the street” strategy that brought Groupon much of its early success. It felt building relationships with local merchants who perhaps aren’t the most tech-savvy required a heavy dose of human touch. However, the Palo Alto folks came at problems from different perspectives and looked to automate many processes. For example, in the old days, a function within Sales Operations known as City Planning would hand pick the deals to send out to customers. Technology folks wanted to automate this function and built sophisticated algorithms that leveraged internal data sets and machine learning to get better at learning users’ preferences. Overall, tension between sales and engineering needs to lead to healthy compromise, as overindexing on one or the other leads to misalignment between the business model and operating model.