This article is authored by MOI Global instructor Elliot Turner, managing director of RGA Investment Advisors, based in Stamford, Connecticut and Great Neck, New York.
Elliot is a featured instructor at Wide-Moat Investing Summit 2019.
Of the many powerful lessons learned from reading “The Snowball: Warren Buffett and the Business of Life” by Alice Schroeder in 2016, one subtle takeaway continues to emerge in my semi-regular musings. Warren Buffett’s grandfather owned and operated Buffett & Son, a grocery store where a young Charlie Munger had worked as a clerk. Schroeder wrote about it within the following context:
Buffett & Son was a direct descendant of the oldest grocery store in Omaha and Ernest’s demanding ways were all in pursuit of an ideal vision of service to his customers. He felt certain the discount national chain stores that were encroaching on the neighborhoods were a fad that would disappear because they could never provide a comparable level of service. Sometime during this period, he wrote confidently to one of his relatives: “The day of the chain store is over.”
Today’s supermarkets all build around a layout which was introduced around this time, where you enter the store in the produce section, travel through the packaged goods aisles and finish collecting items in the dairy and egg shelves. In the pre-WWII America, a grocery store was completely different. As a fascinating blog post about “Grocery Shopping in the 1920s” explains, instead of walking through the aisles and selecting produce for oneself, “In the early 1920s, the customer approached the counter, made a request, and the clerk selected the merchandise for the shopper. Much of the merchandise was kept behind the counter.”
In fact, the customer rarely, if ever handled any product until they were in the comfort of their own home.
Ever since the rise of the Internet, people have mused about the eventuality for Internet-driven delivery, but interestingly the grocery delivery model is an old idea that worked incredibly well within a certain model. The key technology facilitator for delivery in groceries was the telephone. When competition emerged from the early incarnation of the modern supermarket, the town grocer, as Ernest Buffett lays bare, relied on their relationship and customer service as key points of differentiation. As we now know today, lower prices ultimately won out over personal touch and convenience.
Although the town grocer lost out to the scaled supermarket, embedded in the old model is an interesting concept worth exploring. The grocery, as a customer-service centric model conferred a high degree of trust on the grocer himself (back then they were mostly men). The customer service principle included the job of curator. It was the grocer’s responsibility to select the products that checked off the everyday needs of his customer. When a customer called in an order the ask would be for “bread, cheese, soda, a can of beans and toilet paper” in reference to the generic.
The transition to a grocery shopping experience that empowered customers to walk the aisles and select their own products would not have been as simple without the concurrent emergence of brands like Coca-Cola, Kellogg’s and Kraft in the supermarket. The rise of the supermarket aisle saw bottled Coca-Cola overtake fountain sales fueled by the six-bottle carton, the introduction of ready-to-eat cereals from Kellogg and the launch of a powdered beverage named Kool-Aid from Kraft. Given Mr. Buffett’s proximity and familiarity with the fall of the local grocer and rise of the mass supermarket, we perhaps have an explanation for his infatuation in branded CPG like Coca-Cola and Kraft, two names which populate Mr. Buffett’s portfolio today. Brands were the most important shortcut in helping customers navigate the aisles of the supermarket. In other words, there would have been no empowerment absent the power of brand. To that end, brands subsidized the very existence of the supermarket and the resulting relationship was thus symbiotic.
Many have observed that often what’s old becomes new again, and new old. While the Internet has yet to unleash scaled grocery delivery, we can muse about the effects that the online world already hath wrought on brands and one of the biggest consequences is the reemergence of a layer of curation which dilutes the value of brand. There are many great writings on the impact the Internet has had on brands, but we find the history in thinking about this story interesting for analyzing who will win and lose today. To that end, there is one kind of company we are increasingly enamored by for how it fills a vital role in today’s ecommerce landscape: the app that owns the customer. An abundance of really high-quality companies are emerging to help consumers navigate a fragmented ecosystem, full of friction with little differentiation.
A few verticals where we have specific investments include payments (PayPal), content (Roku, and Disney going DTC), home services (IAC, specifically the HomeAdvisor vertical) and food delivery (Grubhub) which we will cover for the remainder of this letter. Each of these referenced companies owns the customer in some way, where customer’s decision-making process starts effectively in-app and leads to a kind of transaction. It is interesting to contemplate the economic reality of a food delivery order placed by searching “Pizza hut delivery” with explicit intent versus an experience that starts with a hungry diner opening up the Grubhub app, given some of the limited screen real estate anyone has on their phone, and browsing through a list of restaurants organized by a combination of proximity, paid targeting (the restaurant giving up a higher take rate) and a diner’s historical preferences. Explicit intent brings little value to the restaurant with no likelihood of incrementality, while owning the customer in app and sending them to a restaurant premised on curated taste is uniquely valuable. Keep this in mind as a point of context while we discuss our investment in Grubhub.
Mr. Market Delivers Time and Again
One of the best positions in our firm’s history, which inadvertently has neither been written up nor presented, though has been discussed in depth with many of you is Grubhub. While our portfolio had a great start to 2019, Grubhub also happens to have been our worst stock in the first quarter. From the first day we bought shares in the company, it has been one of our most controversial positions. We bought our first shares in January of 2016 and from there the returns look stellar; however, over the past six months, shares have shed over 50% of their value.
Grubhub and its SeamlessWeb division (hereinafter we will simply refer to Grubhub) are collectively the patriarchs of the Online Delivery Provider business model. The two brands started as pure online marketplaces, built to connect diners with restaurants for a fee (the take rate). Diners place their orders online at Grubhub, after which restaurants receive a fax detailing the order, manually enter their acknowledgment of receipt of said order, then deliver the food to the diner’s door. As a pure marketplace, Grubhub was just a middleman, collecting its generous take rate in exchange for connecting restaurants to incremental sources of demand and simplifying the process for consumers.
Own the customer, build your TAM:
In building its business, Grubhub removed two crucial pain-points in the food delivery process for consumers. First, consumers often didn’t know all the delivery options in their area and what each had to offer. Grubhub gave a platform through which restaurants could share this information and strategically purchased Menupages in order to improve their own access and awareness to this crucial piece of the puzzle in an effort to drive its salesforce in the right direction. Second, calling in an order was prone to error and in high-volume restaurants, subject to repeated busy signals. For a diner to painlessly click in their order online and know reliably that it will arrive without error was a big change. Grubhub’s success on these two fronts conferred important power on the company as a source of demand generation for restaurants. The site became a hub through which hungry diners commenced their search for food delivery and consequently, while most expected Grubhub’s take rate to compress over time, it actually rose as the company added in the capability for restaurants to offer more take rate in exchange for a higher listing in a diner’s search.
So successful was Grubhub in removing key delivery frictions that in the process, the diner appetite for delivery grew alongside the company itself. When Grubhub started, its opportunity was converting phone orders to online and the total addressable market was clearly defined in that sense; however, as Grubhub proved tremendously successful, the opportunity grew to include shifting a bigger portion of restaurant consumption generally to the online ordering world.
This is all amidst the backdrop of a huge shift from “food at home” to “food away from home,” best demonstrated by this chart shared with us by a sharp analyst:
Source: BLS CPI data, Bloomberg
Making a S***** Business Yummy:
When we first commenced this position, Grubhub shares were weighed down by the emergence of the turnkey delivery business and well-capitalized competitors with no imminent profit mandate like Uber and Amazon entering the delivery business. Pure marketplaces are phenomenal businesses, with high profits and little capital necessary to support the business. With well-capitalized competitors offering to do delivery for restaurants in exchange for a higher take rate, Grubhub was hit with a tough dilemma: do you stick with the marketplace and subsist on your lush economics or do you build your own turnkey operation and expand the platform’s offering?
Grubhub, led by CEO/founder Matt Maloney and CFO Adam DeWitt chose the harder, but more ambitious path of launching their own delivery business. At the time, Maloney called delivery a “s***** business” and lamented that “I’m running my delivery-based business with the explicit goal to break even. That’s not fun for me, and normally I’d say that’s the dumbest business you could ever be in. Why run a break-even business? That’s a pain in the butt.” Grubhub was all-in on this plan in 2015 with the acquisition of three regional delivery services (RDS), commencing its own investment plan and the hire of Stan Chia from Amazon to run operations and logistics for delivery. While Maloney did indeed call delivery a “s***** business” what he meant is that on delivery itself, there would not be much room to generate a margin, even with Grubhub’s scale advantage over both venture-backed and well-capitalized competitors.
It just might be that delivery when added to a marketplace is a better business. Delivery is effectively a three-sided marketplace, which connects diners (first side) to restaurants (second side) via an independent contractor driver (third side). Maloney’s vision was that if delivery could run at break-even, then Grubhub could continue to earn its marketplace margin on the “demand generation” piece. The take rate for a restaurant on the pure marketplace side is around 15%, while the take rate for one who uses Grubhub’s fleet of driver’s is 30% plus a$2 delivery free. Think of this 30% as two separate pieces—15% for the demand generation piece, 15% for delivery. It thus became Grubhub’s imperative to squeeze the cost structure of its delivery service into 15% of the average order. In doing so, Grubhub could reach what they call “economic parity” between marketplace and delivery orders as measured by EBITDA per order, leaving the platform agnostic, the diner indifferent and the restaurant empowered to choose the model that best fit its own needs.
The introduction of EBITDA per order parity was important in setting expectations, because as delivery increases as a percent of gross food sales (GFS), revenue grows faster than before (due to the 2x take rate), while margin is suppressed. EBITDA per order thus gave us analysts a way to track the company’s progress on monetizing orders, in contrast to modeling out revenue and margin assumptions that would now be dis-anchored from past trends. The economics break down as follows:
We note that Grubhub’s delivery take rate of over 30% is far superior to UberEats’ comparable take rate of under 20%, due to UberEats over reliance on McDonald’s as a source of demand at extremely low take rates. Per Uber’s S-1: “the large chain relationship we charge a lower service fee to certain of our largest chain restaurant partners on our Uber Eats offering to grow the number of Uber Eats consumers, which may at times result in a negative take rate with respect to those transactions after considering amounts collected from consumers and paid to Drivers.
In Grubhub’s Q4 2017 earnings report in February 2018, management declared its achievement of “economic parity” between marketplace and delivery alongside its most important strategic partnership to date: Yum! Brands, owner of KFC, Pizza Hut and Taco Bell invested $200m in exchange for a ~3% equity stake with the two set to rollout a nationwide partnership with Grubhub providing delivery services to Yum’s subsidiaries and its franchisees. The relationship with Yum was about more than just bringing Yum on platform. With proven economic parity in their existing model, Grubhub was able to use their lessons learned along the way, leveraged with the Yum platform in order to expand their delivery business from what had been exclusively first tier cities to second and third tier cities. As Maloney explained, “by leveraging our chain partnerships, where they [Yum] have a lot of footprint and a lot of excitement where they can see your [sic] more growth, we can kind of create those anchor tenants and then we fill out the mom-and-pops as we continue to expand our market coverage in these nascent markets.”
For any new delivery market launch, Grubhub had to foot certain overhead expenses (like driver subsidies) in order to ensure enough scale is in place for where delivery volume would go, even before the demand is evident in the market. Yum’s national footprint of brands helps bring the demand necessary in order to justify the cost of starting in these less dense markets. This leaves open a key question: if Grubhub has been a “demand generation” platform historically, what would the economics with Yum look like? CFO Adam Dewitt explained as follows:
“It’s impossible to think about just the economics of the Yum! deal without thinking about the value of the advertising that we’re getting, the value of the cobranding that’s not explicit, that’s more implicit, the new diners that we’re getting, the value of the — the value of just having the brands on the platform and driving new diners. And so you need to take all that into consideration, and we still think that — we still feel really good that over time, even in the kind of short to medium term, you’re not going to see a detrimental impact to EBITA per order related to the Yum! deal or other deals.”
Yum, with its big brands and marketing prowess accomplish the following for Grubhub:
1. Help with customer acquisition by introducing new diners to the platform
2. Support brand marketing for Grubhub alongside the Yum brand names with paid marketing on TV and a branded presence in each of the restaurants.
3. Bring enough demand in order to justify the cost of operating a delivery network in cities that don’t have the same scale and density of New York and Chicago.
Yum effectively introduced a third key pillar of the Grubhub business model beyond marketplace and delivery—a white labeled, outsourced delivery and technology partner. Whereas Grubhub gets compensated for “demand generation” in marketplace or delivery, with a large chain partner Grubhub gets a take rate but also builds the website, the app and handles delivery in addition to the benefits of scale and awareness that come with a large chain’s marketing prowess. As Dewitt explains, “in aggregate, that we get to an EBITDA per order that is not significantly accretive or dilutive, and instead, just adds a lot of growth.” If Grubhub could support Yum “at cost” with its take rate (a secret between the two parties), then the company’s insistence that the deal would be “EBITDA per order neutral impact” would require the scale benefits provided by chains to drive down its “operations and support” costs per order and the halo of the Yum brand portfolio to lower customer acquisition costs and boost lifetime values for Grubhub diners.
Despite the vague economics, investors were initially enthusiastic. Enthusiasm gave way to doubt in the fourth quarter of 2018 when the rollout of the new delivery markets was slower than hoped and with a series of large funding transactions for Doordash on the heels of its reemergence from near demise. With a huge influx of cash, a growth mandate blessed by Softbank with profitability an afterthought and a series of aggressive tactics, like selling non-partner restaurant food, skimming tips from drivers and positioning smartly in Google’s new “delivery” console Doordash was able to quickly take share.  This raised concern about Grubhub’s first mover and scale advantages and the captivity of customers to any of the delivery platforms. By some accounts, Doordash overtook Grubhub in aggregate US market share during the month of March in 2019. While this may be the case, there are four important reasons why aggregate market share is misleading:
1. The market is growing rapidly and even in seemingly losing share, Grubhub is growing alongside the delivery market itself.
2. A lot of the growth is coming from Tier 2 and 3 cities, where Doordash launched first, meanwhile core existing markets remain far more stable than aggregate market share data would have one think.
3. Doordash services “non-partner” restaurants whereby the company earns no take rate and marks up menu prices at the diner’s expense in order to cover the cost of the delivery. Many of these restaurants don’t know they are listed on Doordash nor do some of them want to be.
4. An unknown but meaningful portion of Doordash’s “market share” comes from non-restaurant deliveries, most notably their Walmart This is a good “last mile logistics” business, but is not comparable to GFS at Grubhub.
Competition puts opportunity on our plate once again:
We must confess that the rise of Doordash caught us by surprise; however, we think the fears surrounding Grubhub are as much about the swinging of Mr. Market’s pendulum from euphoria to despair as they are about Doordash or Grubhub itself. The market took the Yum partnership and subsequent acquisition of LevelUp (which we could spill much ink on, but will spare you for now) to such an optimistic extreme that the stock became setup to fail. The launch in tier two and three cities happened alongside Doordash’s ascent and analysts were challenged to figure out whether the EBITDA per order decline across the business was the result of operating nascent markets sub-scale or the influence of competition driving down core economics. In the Q4 2018 earnings call, management shaped expectations as follows:
“We expect adjusted EBITDA and EBITDA per order to increase throughout the year as we gain efficiencies in the newer delivery markets, start to get increasing leverage from the recent marketing and technology investments and achieve general operating leverage as we continue to grow.
As a result, we expect to exit the year generating EBITDA per order much closer to our third quarter 2018 rate of $1.57 than our fourth quarter ’18 rate of $0.98. If you take our 2019 order volume and apply the third quarter of 2018 level of EBITDA per order, it implies that the business would generate $40 million to $50 million in EBITDA above our full year guidance for 2019. So when considering 2020 and beyond, we believe the base EBITDA from where we’ll continue to grow should be closer to that figure than our full year 2019 guidance.”
With sentiment now squarely reset, the risk reward in the stock is as good as it has been since November 2016. There have been several key catalysts since then that have enhanced the company’s advantaged position and some setbacks. They key variables to follow on this thesis are the rate of net new diner additions and operations and support as a percent of delivery revenues. Net new diner adds continue at a robust rate and while competition is forcing customer acquisition costs upwards, the lifetime value to customer acquisition costs remain robust for the company. Further, Yum’s co-branded marketing efforts with Grubhub only kicked in during the first quarter. Consequently, we have so far only seen the impact of the costs associated with supporting the Yum partnership (sub-scale utilization in nascent markets) without seeing any of the benefits. We think management’s perspective here is smart. Maloney summarized it as follows:
“If we wanted to grow at a slower pace, if we wanted to grow similar to what we are doing in the first quarter or the second quarter last year, we could add $20 million to $30 million EBITDA right now, but this is the right choice for the business long term, right. We have — as I mentioned in the call, we’ve accelerated our DAG growth every quarter for the last 5 quarters with a couple hundred basis points consecutively for the third quarter and fourth quarter.”
With this stock trading at 4.40x consensus 2019 sales estimates, the market is implying that growth will slow and future margins will be well below the “economic parity” margins the company generated just one year ago. In fact, this is the lowest price to estimated sales in the company’s history since going public. The Q4 2015 P/S dip below 5x coincided with skepticism surrounding the initial launch of the delivery business.
As Maloney and team executed on achieving profitability goals, the multiple expanded. We see many parallels to late 2015 when looking at the following chart:
We find market cap to forward GFS an especially insightful and important valuation metric:
Source: Sentieo, RGAIA
EBITDA to gross food sales has hovered around 4.5% historically and as such, a consistent market capto forward GFS is an expression that the economic value Grubhub extracts per dollar of GFS will remain consistent over the long-term. At today’s market cap of around $6b and expected GFS of around $6.4 billion in 2019, the stock is trading below a 1:1 ratio. At around a one ratio, an investor could reasonably expect to earn a return commensurate with the growth of GFS over a holding period, assuming economics stay consistent. In our fairly conservative estimate for GFS, which notably excludes any economic benefit from LevelUp, we expect a 15% compound annual growth rate (CAGR) over the next three years and around 14% over the next five years, with exit growth beyond that period expected to maintain growth rates into the double digits. We think there is the reasonable opportunity for EBITDA/GFS to improve over this time period, which would be accretive to returns, though we are not underwriting the position under the presumption this happens. The biggest opportunities in improving EBITDA/GFS stem from leveraging the Sales & Marketing. Technology the Operations and Support expense lines over time.
In many respects, we view this situation as analogous to buying Booking Holdings (Priceline Corp) when it was at a similar revenue run-rate in 2006, only better (for added color, consensus expects 36% top line growth for Grubhub this year, vs 17% growth for Booking in 2006). Grubhub is similar to Booking back then, in how there are questions about competition from well capitalized players and industry incumbents and general frustration with the high fees the platform was able to generate their respective merchant bases. The similarities are deeper in how Grubhub and Booking each earn the majority of their economics off of an incredibly fragmented customer base (in Booking it’s the largely independently owned European boutique hotels, while with Grubhub it’s the independent restaurant landscape). Grubhub is superior is how the company owns the customer with its presence at the top of funnel where search and discovery commence, while operating in an area where the frequency of interaction with the company is far more regular. Booking unfortunately must reacquire customers each time someone plans travel, which for many customers happens only once or twice a year. Booking merely offers both parties the removal of friction, Grubhub offers demand generation as an added layer and thus should earn superior economics in the very long run. Notably, Booking was able to grow its top line at a near 24% CAGR since 2006—a span of twelve years. If Grubhub’s high growth time period could meet Bookings, even if at a much lower rate, the returns to long-term holders would be incredibly attractive. In our DCF, we think the mid-$80s per share is justified today. Each year we increase our forecast period beyond year five adds upwards of $30 in expected value. This goes to demonstrate just how important the company’s ability to extend its high growth period before reaching terminal value will be in driving long-term returns.
 Schroeder, A. (2008, 2009). The Snowball: Warren Buffett and the Business of Life. New York, NY: Bantam Books. Page 69.
 Grubhub Q1 2018 Earnings Call Transcript.
 Grubhub Q2 2018 Earnings Call Transcript.
 Grubhub Q1 2018 Earnings Call Transcript.
 Grubhub Q1 2018 Earnings Call Transcript.
 Grubhub Q4 2018 Earnings Conference Call
Past performance is not necessarily indicative of future results. The views expressed above are those of RGA Investment Advisors LLC (RGA). These views are subject to change at any time based on market and other conditions, and RGA disclaims any responsibility to update such views. Past performance is no guarantee of future results. No forecasts can be guaranteed. These views may not be relied upon as investment advice. The investment process may change over time. The characteristics set forth above are intended as a general illustration of some of the criteria the team considers in selecting securities for the portfolio. Not all investments meet such criteria. In the event that a recommendation for the purchase or sale of any security is presented herein, RGA shall furnish to any person upon request a tabular presentation of: (i) The total number of shares or other units of the security held by RGA or its investment adviser representatives for its own account or for the account of officers, directors, trustees, partners or affiliates of RGA or for discretionary accounts of RGA or its investment adviser representatives, as maintained for clients. (ii) The price or price range at which the securities listed.