Virtual Reality and Theory of Modularity

If you're anything like me and don't know much about mobile gaming, you haven't paid a ton of attention to virtual reality. It seems like a cool technology for sure, but useful primarily for playing computer games. That's what I thought, at least, until I heard Marc Andreesen on the A16Z podcast in August of 2015. 

After hearing Andreesen, I realized that VR can change lives for many people around the world. Most people's, especially those in war torn or developing countries, "real reality" is nothing to envy. VR can give people an experience that would be nearly impossible in actual reality. Imagine a student from anywhere in the world being able to sit in a Stanford classroom and interact with students and professors as if he were actually there. VR simulations are already being deployed outside of gaming; they already have the ability to create 3D models of patients' anatomy, can make history and science classes come to life, and allow auto manufacturers to test drive a car that doesn't yet exist. 

Despite all the promises of VR, the technology has mostly remained inaccessible to the masses, and Google intends to do something about that. The month of May brought us Google I/O and its most talked about announcement of the mobile virtual reality platform called Daydream. Daydream follows Google's initial foray into VR, which was in 2014 through a cheap, disposable headset called Cardboard. But Cardboard came with a latency problem which could make users sick, a problem  that was solved by higher-end VR headsets such as the Oculus Rift, Oculus+Samsung's Gear VR and HTC's Vive. 

Daydream's introduction fomented a debate about who is likely to win the VR headset race. Daydream, according to Gizmag, is more reactionary than innovative. Gizmag argues that the quality of Daydream will likely continue to lag behind the higher-end headsets, which have multiple controls and are already working on things like positional tracking. "Daydream shouldn't pose much of a threat anytime soon," claims the article. 

What makes Daydream more interesting is Google's announcement that the VR platform will be based on the next Android version called Android N. Consequently, many phones that run Android will come optimized to run VR experiences and be Daydream ready. These phones will be certified by Google and will be required to have various VR friendly components such as "high-quality sensors for head tracking or screens that can reduce blurring by showing images in extremely short bursts," according to The Verge. 

In addition, Google will curate the Play Store to have content optimized for Daydream. In fact, "Google VR head Clay Bavor specifically mentioned Hulu, Netflix and Lionsgate as some of the companies bringing media content to Daydream," according to Variety. Moreover, Bavor mentioned that the Company has already built YouTube from the ground up for VR. The accessibility of Daydream is expected to shift relevance of VR from niche PC gamers to any mobile user who wants to experience various apps in a different way. 

Google's playbook for Daydream is straight out of the renowned Clayton Christensen's Innovator's Solution. Christensen begins by defining interdependence and modularity. He says that "an architecture is interdependent at an interface if one part cannot be created independently of the other part- if the way one is designed and made depends on the way the other is designed and made." He goes on to say that "a modular architecture specifies the fit and function of all elements so completely that it doesn't matter who makes the components or subsystems, as long as they meet the specifications. Modular components can be developed in independent work groups or by different companies working at arm's length." 

Each product can have some components that are interdependent and some that are modular. An iPhone and iOS are interdependent, but the apps on the iPhone are modular.

What kind of architecture is best for VR today? Christensen argues that when a product's functionality is not yet good enough to address customer needs, firms that build their products around proprietary, interdependent architectures enjoy a competitive advantage because standardization in modularity takes too many degrees of design freedom away from engineers and performance cannot be optimized. He goes on to say that, "one reason why entrant companies rarely succeed in commercializing a radically new technology is that breakthrough sustaining technologies are rarely plug-compatible with existing systems of use." So if we think today's VR technology isn't good enough for mobile, then an interdependent structure, or the one Google hopes to create with Daydream, could win out because it will be easily fit and function within all the Android phones.

Modularity, however, becomes the dominant design when products become good enough, and there is a performance surplus from the product. Once the requirements for functionality and reliability have been met, products begin competing on speed of upgrades or responsiveness to customers. With modular architectures, companies can introduce new products faster because they don't have to redesign everything. "Whereas in the interdependent world, you had to make all of the key elements of the system in order to make any of them, in a modular world you can prosper by outsourcing or by supplying just one element." If we think that mobile VR technology is good enough, then those companies that innovate faster and are more responsive to consumer needs, such as Oculus and HTC, would become the dominant players.

The trajectory of product architecture as defined by Christensen, is depicted in the figure below:

The Oculus and HTC VR solutions are almost textbook examples of modular designs. Baldwin and Clark in "Managing in an Age of Modularity" note that modular designers "rapidly move in and out of joint ventures, technology alliances, subcontracts, employment agreements and financial arrangements as they compete in a relentless race to innovate." Baldwin and Clark note that since designers achieve modularity by partitioning information into visible design rules and hidden design parameters, modularity is only benefitial if the partition is "precise, unambiguous, and complete."

When it comes to mobile VR, it will be important for the VR architecture to be designed well to fit the phone's architecture. Here, Google's Daydream presents a truly mobile experience, not one that we initially for PCs and later fitted to smart phones. In that context, it wins by providing accessibility of VR to a broad range of users. As John Nagle of Gyoza Games said, "With the launch of Daydream, Google is again further democratising VR, making it accessible to a vastly broader audience than was ever before possible."

As VR becomes more accessible, Google's platform will be able to provide developers with standards and design rules, allowing for modularity in applications that could be used in a VR setting. Nagle goes on to say, "From a development perspective, including a controller and providing a ‘Daydream Ready’ hardware spec is a great advantage, because it means that we can focus on building great content instead of spending time and money developing and testing on so many disparate hardware platforms."

Therefore, Google wins at first with its interdependent design, even if it doesn't have the best VR solution, just because it is able to make the technology accessible for both users and developers. The VR technology just isn't seamless enough yet with mobile phones to be useful to the majority of smart phone owners. Google's reference device that will allow manufacturers to bring their own headsets will create a much needed standardization in the market, which will allow developers to focus on content rather than on hardware.

As Daydream as a VR platform becomes more prevalent, however, the industry will move towards modularity. That is, phone makers can create their own headsets as long as they meet Google's specs. Developers can create content that will align seamlessly with Android. And the market will move towards a performance surplus as VR vendors innovate quickly to offer performance and the rich content to meet user demand. 

Oil Prices, Tech Market, and the Economy, Part II

The Tech Bubble

The New Yorker

Asset bubbles, or the appearance of them, concern almost everyone due to their omnipresence in the media. Last year, Mark Cuban wrote a paroxysm of how the 2015 tech bubble is worse than the 2000 bubble because retail investors can't participate (Cuban's a smart guy, but this makes no sense. Isn't that a good thing if we're in a bubble and retail investors aren't participating? It would mean fewer people hurt by the popping of a bubble if companies are private instead of public). Everyone, from TechCrunch to Vanity Fair is talking about the tech bubble. I'm just waiting for Drake and Meek Mill to have diss tracks arguing about the existence of a tech bubble. 

I think it's smart to begin by defining an asset bubble, which, surprisingly, isn't easy, followed by some empirical evidence and economic theory.

Definitions: 
Some economists define an asset bubble as an upward price movement in an asset over an extended range of time that then suddenly implodes. For the most part, this definition is too ambiguous because it doesn't convey how high the prices should move and why that movement is not justified. 

A more precise interpretation would define a bubble as a "situation where an asset's price exceeds the fundamental value of the asset," according to Gady Barlevy of the Federal Reserve Bank of Chicago. He goes on to note that an asset's value ought to be the present value of its future cash flows. If the price of an asset increases significantly, perhaps in a short period of time, without any expected change of future cash flows, then there may be an asset bubble. Again, there ought to be a difference between prices when expected cash flows grow by 10% versus 1000%, Thus, this definition is also bereft of quantifying how large the movements have to be to constitute a bubble, but the idea of prices being disassociated with fundamentals is key.  

I also think it's imperative to note what a bubble is NOT: 

1. There is not necessarily a bubble just because prices in a certain asset class are higher than they used to be. Pricing must be divorced from fundamentals in order to constitute a bubble. 
2. In the same vein, higher fundamental valuations don't always point to a bubble. Many reporters cite the growing number of unicorns as a sign of a bubble. If the present value of that company's future cash flows is over $1 bil, then the valuation could be justified. That's not to say high valuations are always justified, and indeed, in many cases they are not. But, in order to proclaim a bubble, one should dig several levels deeper to figure out if the entire asset class ought to be generalized as overvalued beyond fundamentals. 
3. Lower stock prices do not necessarily mean that there was an asset bubble that is in the process of bursting. Day to day volatility in the stock market is a lot higher than day to day volatility in valuations of companies. If stock price compression is due to political reasons, over-reaction to market data, or other exogenous factors, it probably isn't a sign of a bubble popping. A correction from speculative levels of valuations to levels that more adequately reflect future cash flows, however, could be a sign of a bubble deflating. 
4. Failure of companies does not mean there was a bubble that is imploding. Ben Thompson of Stratechery made a great point about the winner-take-all market: there will be failures in industries where there's only room for one major player due to network effects. Advertising is a zero-sum game and some apps/social media sites will lose out to others. That doesn't mean there was a bubble; that just means expected cash flows from one company were transferred to another.

So what could have caused the recent downturn in tech stocks recently? We can't ignore the impact of the Chinese stock market and recent tech earnings. Ben Thompson, however, makes a great argument:

"I think the recent chill in valuations and fundraising is about coming to terms with the fact that a lot of those unicorns are in the same boat as Facebook and Google’s advertising competitors: they have already missed out to the dominant player in their field (or, that their field was never viable to begin with). In some respects it is tech’s own inequality story: the average and median company and startup will increasingly bifurcate. It’s not a bubble, it’s a rebalancing, and the winners are poised to be bigger and richer than anything we have seen before." 

Empirical Evidence:  
How can we ascertain the existence of a bubble? When valuations reach unrealistic levels, we see more and more financial capital chasing companies in a particular industry. So, first, we can look at how much money has gone into the venture market now and back in 2000. In 2000, over $100 billion had been invested in VC, compared with $59 billion in 2015, according to the PWC MoneyTree Report. So there has been a lot of money spent in VC, just not as much as there was in 2000.

Moreover, during bubble times, investors put money into companies at the "idea"or very early stage. Speculators and non-venture investors enter the market at seed or angel rounds in hopes to land the next unicorn. If investors are chasing newfangled investments in hopes for another gold rush the way they were in 2000, one would expect to see much more capital into the seed and early stages of investments.

According to the PWC MoneyTree data (graph below), there was $21 billion of capital into the seed and early stage VCs in 2015 compared with $29 billion in 2000. Approximately $8MM per venture went in at the seed or early stage  in 2000 compared with $8.6MM into the same rounds in 2015.

It's interesting to note that the average dollar size per investment during seed and early stages is higher this time around, even though total capital deployed is lower, implying that there could be more of a winner-take-all strategy that Ben Thompson alluded to earlier.

So are companies this time of higher quality than before, justifying the higher level of investments? We can look at the quality of tech companies that have raised public financing this time compared with in 2000 to help determine that. While this isn't an apples to apples comparison, it gives us an idea of how mature the companies are when they go public, and ultimately, how far the investments could fall if their valuations aren't justified.

I looked at 2002 revenues of Nasdaq companies that IPO'ed between 1998 and 2002 and compared that with last twelve month (2015/2014) revenues of Nasdaq companies that IPO'ed between 2012 and 2015. The idea was to see how much more traction today's companies have before the public, non-VC investors jump in. Intuitively, we know that since companies have been waiting longer before going public, they should have greater revenue traction, which was corroborated by the data below.

The data above confirms the suspicion that companies are in later stages and have more control over their expenses (which the revenue/employee is supposed to indicate) now than they did in the early 2000s. Which means that if there is a bubble, the speculative nature of it isn't nearly as bad as it was in the early 2000s. Perhaps, instead of acting with irrational exuberance, investors are merely unreasonably quixotic.

So it appears that there's not much of a bubble in the public markets or later stage VCs since, outside of a few companies, valuations are more or less close to fundamentals this time around with higher revenue traction and lower expenses. As in, we're seeing "real companies" at the later stage VCs and public companies than we had in 2000, when we saw more investments due to speculation.

There could be a bubble in the early stage or angel investing staged companies since there is almost as much capital invested at those stages now as in 2000. The perilous impact from the bursting of that bubble would be limited to investments in just those stages.

What would cause the implosion of high valuations in tech companies? Katie Benner of New York Times and Jason Calacanis proffered in an engaging TWIST round-table that what we might see this time around may be similar to what we saw during the 2009 crisis. That is, we may see a negative impact on the tech industry if other industries or consumers began buying less technology. As mentioned in my previous blog post, if oil companies, for example, began spending less or had massive layoffs so consumers couldn't spend on technology, we'd see lower tech revenues. Jason calls this a contagion, or a downturn caused by exogenous factors rather than the inherent overpricing and then correction of tech valuations themselves. The magnitude of the contagion affect this time would be a balance between how much more investment there is in technology now versus in 2009 and how much of an impact a downturn will have on tech revenues now than in 2009.

Economic Theory: 
For economic theory about asset bubbles, there's no better source than Carlota Perez's Technological Revolutions and Financial Capital. In it, she describes four phases of technological revolutions: Irruption, Frenzy, Synergy and Maturity.

The Irruption Phase is when "new revolutionary entrepreneurs outstrip profit making potential of all established production sectors, and there is a rush of financial capital towards them, readily deploying new appropriate instruments when necessary." In this period, there is idle money in search for profitable use, and it leans towards investing in these new entrepreneurs to obtain high yields.

The Frenzy Phase is when there is a decoupling of financial capital and production of new innovation. Financial capital becomes arrogant from the highly profitable "bets" made by investors. Financial capital becomes a powerful magnet to attract investment into new areas, which become the "new economy". The entrepreneurs are forced to do whatever is necessary to attract the investors, in this case, the  VCs.  This is also when uncontrollable inflation sets in, debt mounts at a reckless rhythm, and a vast disproportion between paper wealth and real wealth becomes apparent.

After the Frenzy Phase, there is a turning point, which brings with it a collapse and a recession. Bubbles begin at the end of the Frenzy stage and burst during the turning point.

"There are three structural tensions that make it impossible to keep the frenzy profit going for an indefinite time. There are tensions between real and paper wealth, between the profile of existing demand and that of potential supply in the core products of the revolution, and between the socially excluded and those reaping the benefits of the bubble."

The Synergy Phase is where there is a re-coupling of financial capital and production. Innovation and growth can take place across the whole productive spectrum in this phase.

In Maturity, some disappointment comes from highly profitable sectors reaching their limits in both productivity and markets. Profits begin dwindling, and we begin to see "idle money" in the financial markets again.

So which phase are we in? According to Ms. Perez, we're likely in the midst of the turning point.

While it is possible that there will be several small tech valuation bubbles followed by corrections during the Turning Point, it seems from Perez's work that the big tech crash and related recession have already occurred. And it appears, from the empirical evidence, that we're facing less of a bursting of a tech bubble and more of a contagion effect on the tech industry at the moment. So, are we in a tech bubble? Probably not.

Oil Prices, Tech Market and the Economy, Part I

IS-LM Model

Dow Jonesy enough for you?
The New Yorker

One of my colleagues and I were recently discussing the dizzying number of stimuli trying to play tug-of-war with the US economy in general and the financial markets (public and private) specifically. There's so much going on! Collapsing oil prices, prodding at the "tech bubble", the presidential elections, burgeoning threat of terrorist activities, Kim Jong-un inching towards insanity... What does all of this mean for our economy and our financial markets. 

As an economist, I like to (over)simplify the world by thinking in terms of frameworks, and the one framework that spoke to me in my macro courses was the IS-LM model, which illustrates the Investment/Savings - Liquidity Preference/Money Supply equilibrium. Behind this almost nonsensical jargon is a simple concept that interest rates and GDP are a function of how much money is sloshing around and whether people would rather invest or save that money.  

For those of you who want the details, the IS and LM curved are derived from the aggregate demand equilibrium where output (Y) = C (consumption) + I (investment) + G (government spending) + X(net exports)

The LM curve is a little harder to understand, but an easy way to think about it is that i(interest rates) is the price of holding on to money. That is, we all would rather have money in our checking accounts, readily accessible (holding money), but if i (price of money) is high enough, we'll let someone else (mutual funds, banks, etc) hold our money and pay us for it. 

The IS curve is pretty relevant today given the collapsing oil prices. Exogenous variables (not just the price) have resulted in a glut of oil supply. Conventional wisdom would tell us that lower oil prices would boost production of items where oil is an ingredient since input prices would go down. As The Economist mentions, however, that doesn't seem to be the case at the moment. 

"Cheaper fuel should stimulate global economic growth. Industries that use oil as an input are more profitable. The benefits to consuming nations typically outweigh the costs to producing ones. But so far in 2016 a 28% lurch downwards in oil prices has coincided with turmoil in global stock markets. It is as if the markets are challenging long-held assumptions about the economic benefits of low energy prices, or asserting that global economic growth is so anemic that an oil glut will do little to help."

Let's go back to our IS-LM framework: oil prices are low, so investment in oil should decline, shifting the IS part of the curve to the left (from IS to ISd). If investment in oil is lower, producers will be reluctant to produce oil (and many may not be able to produce given their much lower income). 

As The Economist says above, it could have been possible that although the I part of Y=C+I+G+X would decline, increase in production of goods could have boosted the C and X portions of the equation, hence leaving the IS curve unchanged or even higher (to the right, to ISi). That doesn't seem to be the case this time, however. It appears that C and X, for whatever reasons, need more to boost them than lower input prices. As an example, you'd like to buy a Lucite table that costs $500 last year and now costs $400 (Consumption), but you still don't think it's a good enough deal to purchase (perhaps wages haven't risen enough, prices of other goods have risen more than you had expected, your taxes have increased, etc). 

The impact of a lower IS curve is that GDP on the X axis (output, income, yield, or whichever other measure you'd like to use) is lower and the economy is producing less. What can be done to counteract that? 

Well, usually, the Fed could just shift the LM curve to the right by spurring money supply. That is, the Fed could increase the money supply by buying bonds (less bonds, more money in the market), which is the mechanism used to achieve what everyone refers to as "reducing interest rates". With lower i (interest rates, the price of money), instead of letting someone else hold your money (put it in a savings account, a mutual fund, etc), you hold it yourself as just a store of value. 

As we know, however, interest rates are pretty close to the lowest they can be. Although rates can be reduced a little more and into negative territory with quantitative easing, there's likely a limit to how low they can go before the economy becomes topsy turvey.

Paul Krugman at the New York Times says that at the rate of 0, the LM curve should be flat, like this:

When rates are at 0, people have no incentive to buy bonds or put money in a savings account; they'd rather hold cash. Changes in the money supply have no impact, which is known as a liquidity trap.

"And IS-LM makes some predictions about what happens in the liquidity trap. Budget deficits shift IS to the right; in the liquidity trap that has no effect on the interest rate. Increases in the money supply do nothing at all," Paul Krugman.

So if the LM curve can't be changed to counteract the lower IS curve and if lower oil prices aren't doing much to spur consumer demand, then we're left with the lower IS curve, meaning lower output and rates.

What does that mean for the rest of the economy? A lower IS curve is further away from full-employment, so we'll likely see higher unemployment, primarily from the energy sector. But rates will also remain low.

Low interest rates lead many investors to continue searching for higher yields elsewhere. That is, investors will take on riskier endeavors because risk-free investments (treasuries) or low risk investments (investment grade bonds) won't yield much. This impacts the technology market and silicon valley enterprises too. The lower the yields, the more money investors will be willing to put into startups and other technology ventures that have the possibility of an out-sized return. And we've seen this phenomenon for a while now since rates have been close to zero.

The demand side of the equation for technology and venture startups may be impacted negatively if products were meant to be sold to firms related to oil production. At this point, however, it doesn't seem likely that we'll see a major revenue slowdown for tech companies just because of lower oil prices. It also doesn't seem like lower oil prices are having a recessionary impact on the rest of the economy because it hasn't hit industrial production, the real GDP, real income or wholesale retail sales.

What about the "tech bubble" then? And do the weakness in the public markets portend an impending disaster? Should we expect a deleterious impact from a repeat of the 2000's tech crash? Or is this time different? I will address these issues in the context of a technological revolution in Part II of this post. 

500 Startups and Conglomerate Theory

In early October, I had the good fortune of seeing Dave McClure, founding partner of 500 Startups, at the Geekwire Summit in Seattle, WA. McClure, with his humor and charisma, had the crowd roaring with laughter as he opined on the venture capital industry. 

McClure's philosophy of venture capital is unique: he wants to be big, but not just for the sake of being big, but also for the sake of diversification. He hates the idea of having a few portfolio companies, knowing that only about 10% will be wildly successful. If the rate of success is so low, why not just have more portfolio companies? "I would say spray not pray" says McClure of his investment thesis. The diversification in his portfolio mimics asset allocation techniques of typical Wall Street managers.  

Fund of funds, however, are circumspect about investing in 500 Startups for various reasons. Michael Kim of Cendana Capital likes investing in VCs that take big bets because the chance of big returns is higher. 

Jason Lemkin of SaaStr notes that while IRRs have been higher than most VCs at 500, given its early stage of investing, cash-on-cash returns may take longer to transpire at 500 than at other VCs. 

When I heard McClure talk about valuing diversification so highly, I instantly thought of conglomerate theory in corporate finance. Although there are a lot of differences between a traditional conglomerate and a large VC, there are certainly some similarities. Early stage VCs such as 500 do have operational roles in their portfolio companies, the way a Berkshire Hatahway or GE do. VCs are heavily invested in their portfolio companies, and, especially in the case of 500, the synergies are more financial than operational. 

What are some of the reasons investors eschew conglomerates? For one, there is an inherent bias against growing for the sake of growth, and it is believed that conglomerates are in a precarious position for doing just that. "Many conglomerates remain obsessed with empire building, sacrifice value for growth, overpay for acquisitions, hang on to businesses that will never prosper (or would perform better in other hands) and fail to develop structures, impose disciplines and create cultures that sustain value growth." according to Kaye and Yuwono of Marakon Associates. So the question related to 500 Startups would be: Do we think McClure is growing the VC just for the sake of being big? At some point, does adding lots of startups mean that investment criterion have become less stringent? 

Another reason money managers prefer non-conglomerates or core businesses is because they view diversification as a job better done by portfolio managers, not management teams. An investor may want to invest in the healthcare sector of GE but not the oil and gas sector. So if she does buy GE stock, she'd to pay a discount for having to invest in an industry that does not interest her. In terms of 500 Startups, do the fund of funds believe that they can do a better job than McClure of diversifying the venture portfolio? What if investors want to support startups in Asia but not Europe and would rather choose geographically focused VC? 

To be fair, there is a conglomerate premium for conglomerates outside of the US, as noted in this HBR study, for reasons that may also apply to 500 Startups. For example, a conglomerate in an emerging market may be able to operate each segment of the business better together because it may be able to navigate the regulatory and political environment better than anyone else. 

McClure argues that the reason for diversification of 500 Startups is so that the companies can access a wider network of experts, thus providing an operational edge to each of those companies. He says that VCs would like to believe that they are brilliant, but investing in 30 companies is no guarantee that one will hit the $1bil mark. But doesn't buying up a myriad of companies in all different sectors and geographies smell like an index strategy? 

Axiomatically, I don't like the idea of playing the numbers game in VC. To me, it sounds like accumulating the most start-ups is akin to investing in the market or chasing beta. By saying VCs aren't "brilliant", he's implying that there's no alpha, so he's employing a benchmark strategy.

But there IS alpha in venture capital! There has to be, right? VC is one of the least efficient markets out there: transaction costs are high, information is asymmetrical, access to investments is limited in many cases, VCs are not rational, and pricing is turbid. All of these inefficiencies, at least theoretically, allow for opportunities to outperform (or underperform) the market consistently (not based on luck, as efficient market hypothesis would suggest). Isn't this why Sequoia and Benchmark outperform, because they are able to generate alpha? 

If 500 Startups is outperforming, I think it's because McClure is more "brilliant" at picking and operating good companies than he is giving himself credit for, and not just because of having lots of companies in his portfolio. 

For more on pluses and minuses of conglomerates, go here. 

Food Delivery Services: Lessons From Struggling Grocery Stores

This post was originally published on Distressed and Turnaround Blog in July 2015. 

With increased competition, sky high operating expenses, and little room to pass on any increased food costs, American grocery stores have long faced a tumultuous battle to survive. And yet, online food and delivery services are winning VC capital with no end in sight. It's difficult to imagine that online comestibles shopping will stultify grocery stores the way Amazon has for brick and mortar retailers. It is however, prudent for each food-related sector to learn survival tricks from the other. 

In regards to grocery stores, barring the recent Albertsons' IPO (which has its own turnaround story from the SuperValu days), news has been pretty dismal. There are reports that A&P is close to filing for bankruptcy protection for the second time in five years due to high debt payments, competition from Whole Foods & Trader Joe's, and high pension costs. Haggen, which bought some Safeway and Albertsons stores recently, announced  that it will be cutting some labor expenses. 

And last November, Dahl's Foods, an 83-year old Iowa based filed for Chapter 11 and agreed to be acquired by Associated Wholesale Grocers. The company claimed that it was "slow to recognize the competitive threat and to make the operational changes necessary to remain viable. During this period, the debtors over-leveraged their assets as their revenues declined and became under-capitalized."

In contrast, web-based food and grocery delivery services has been one of the hottest VC sectors, with more than $1 billion invested since 2014, almost a four-fold increase year-on-year, according to TechCrunch. 

Actually, TechCrunch's piece on Food Delivery Wars is a good primer on various offerings backed by Silicon valley today.

So what lessons can these hot delivery services learn from supermarkets and their struggles? Some of the obstacles that led to the demise of grocery stores won't apply to online services, such  as high pension or rent expenses. Other challenges, however, may appear to delivery services in disguise. Here are a few tips: 

Customers are fickle and it's best to have the option to adapt. Consumers crave variety over time, and you want to avoid A&P's fate of not being able to make operational changes to remain viable. For example, Sprouts Farmers Market increased its productivity by 15% by shifting its focus from being a specialty store to an everyday healthy grocery store, according to SuperMarket news. If a delivery system isn't able to get adequate revenue or customer growth, it might be time to think of adjacent products to offer, both to retain existing customers and entice new ones. This sounds intuitive, but it's surprising how many delivery services have a limited scope, whether it's a small number of restaurants contracted or services that only deliver cold pressed juices. And it that vein... 

Data are your friend. Since consumers are fickle, the only way to stay abreast of trends is by collecting and analyzing data. Albertsons', for example, uses extensive data to systematically monitor emerging trends in food and source new and innovative products, according to its S-1.  

Munchery is a good case study of an online service using data to optimize its sales. The food delivery system couldn't figure out why some of its chef-prepared gourmet meals weren't selling. Contrad Chu, the entrepreneur behind the concept, used Desk.com from Salesforce to track customers’ order histories and food preferences, as well as company mentions on Twitter and other social media feeds. This helped Munchery spot and react to macro taste trends such as kale mania, the Paleo diet or ethnic foods. 

Data are especially important for delivery services that touch products, like Munchery, Plated or Blue Apron. If you're putting in the effort to obtain ingredients, prepare (or pre-prepare) the meals, and package them, you want to make sure that they'll be a hit. Social media is a great tool to glean these trends. Fast food restaurants such as Taco Bell use data from social media to get an early indication of which products are working and why. For earlier stage startups trying to understand their addressable market, there are a number of analytics firms that excel at collecting and interpreting these data, which could be helpful before product launches.

Keep expenses tight. Unlike traditional grocery stores, delivery systems don't have the problems of yesteryear such as pensions. Having lean operations from the get-go, however, will help online companies weather difficult times. And while many services, such as Fresh Direct, have an excellent supply chain that affords it to minimize shrink and waste, start-ups without that kind of scale may have not have that negotiating power. 

Instacart, for example, has a powerful software that that keeps a tally of how fast orders can move through the system at any given time. “We have an algorithm that runs every minute of the day that evaluates what orders we have, what supply we have, and whether or not we can take a one-hour order and place it on time. Before you’ve even placed an order, we’ve done all the math," said Instacart co-founder Max Mullen.  What that means is Instacart doesn't need to have an extra worker, driver or even packaging material with its ability to forecast demand. With the information collected from its app, Instacart is able to predict when customer orders will come in and for what items, so that it is able to manage its supply chain and operations efficiently. 

Obviously, operating expenses are less important for SaaS companies that don't actually touch the products, such as GrubHub. In such cases, something in addition to the software may help sell the product, such as a network effect, also offered by GrubHub. 

Expect a lot of competition. Everyone eats, so the total addressable market for food should grow more or less with population. Sure, maybe there's some room to expand the pie, but for the most part, when one company gains market share, another loses. Grocery stores have learned that the hard way, with increased competition from superstores such as Walmart and Target, discount retailers such as Aldi or even the dollar stores, and lastly, the advent of online food delivery services. And many grocery stores aren't taking it lying down; they've realized they need to get into the home delivery and online shopping on their own. 

Although the Internet is rife with advice for entrepreneurs about gaining a competitive advantage, one way to gain scale is to collaborate with providers for adjacent products, because that is essentially what the big box retailers do. A group of best-of-breed online food services can gain scale quickly and control costs by having increased negotiating power with suppliers or vendors. They'll have more data to assess consumer behavior and sell what works. Sure, it requires some creativity and innovation, but there's no shortage of that in the food start-up land. 

Fashion Risk

This post was originally published in Distressed and Turnaround Blog in July, 2014. 

Fashion is risky business. Who knows what works? In the investment circles, buying stocks or bonds of a company that sells modish products is referred to as taking on "fashion risk". Rightfully so, because many of them fail, with the most recent example being Coldwater Creek. In this post, I will use Coldwater Creek as a case study for some of the possible pitfalls for an apparel retailer and discuss strategies for convalescence.

Coldwater Creek, a specialty clothing retailer, filed for bankruptcy on April 11. According to the recent liquidation plan, all term loan, priority and secured claims will be paid in full. Unsecured claims would recover an estimated 4.43%. TheDeal.com

The retailer was once beloved by middle-class, professional women for its effortlessly chic styles. It started as a catalog company back in the 80s and soon began opening retail stores in the 90's.

"The shopping centers, located in upper-middle-class neighborhoods, cater to the company's core audience: women who earn an average of $70,000 a year and who are drawn to such Coldwater Creek staples as $79 burnished silk jackets and $65 reversible suede belts... 

(Dennis) Pence is rushing to capitalize on an emerging demographic group that retail experts have dubbed the zoomers. They are baby boomers with a zest for living" Bloomberg BusinessWeek

Fast forward to April 2014 and the Company filed for bankruptcy after attempts by the debtors to refinance the debt, recapitalize the balance sheet and even sell the enterprise outright failed. What happened?

       

2007 happened, and everything went downhill from there. Net sales declined approximately 32% from its peak in 2006 to $742 million in 2013. The Company has cited everything from a slow-down in traffic to merchandising issues for the steep decline in same store sales. Following the reduced revenues, the buildup in inventory and its subsequent markdowns resulted in a deterioration of margins.

James A. Bell, EVP, COO and CFO of Coldwater, stated in the Declaration in Support of First Day Motion:

"From 2011 to 2013, the Debtors attempted a targeted turnaround process, which focused on the following:  a) incorporating cross-channel discipline into product and creative functions b) establishing the foundation of product assortment architecture c) acquiring retail-centric talent d) developing and implementing a real-estate optimization program e) positioning the brand strategy to ensure focus on target customer and f) re-engineering design and product development functions." Pacer

In order to gauge what the problems could have been, I did a little digging online for reviews from customers and employees. Here are some quotes that sum up many of the issues.

"Coldwater Creek USED to be a place women could go to find elegant, classic fashions. This year (2013) in particular, they have seemed to slide down that same slippery road as JC Penneys and KMart, which has resulted in imported, tacky, frumpy-looking clothing." Customer, May 2013

"I hear CWC is trying to grow back its customer base, but I just took a look at their website. Styles are dowdy and colors are drab. With my past experience with CWC poor quality and fit, there is nothing that tempts me to try again." Customer, October 2013

"Return policy is too liberal. End the 'anything, anytime, for any reason' deal." Employee, November 2011

"People bring stuff back from 5 years ago and get money or a gift card." Employee, December 2011

"While the employee discounts are good, the quality of the products have dramatically declined over the past 3-5 years." Employee, August 2012

"They are trying to change their entire customer base and it makes it difficult for the store emplyees to help the customers that made the company so popular to begin with. Not every customer wants short sleeve tops, slim leg jeans, and shapeless tees. Bring petites back to the stores we are losing customers everyday because we are only carrying pants in the store. Same with 3X sizes for customers." Employee, August 2012

A few observations:

1.  Returns Policy- I wouldn't recommend any turnaround strategies without a through quantitative analysis of costs and benefits. That said, anecdotal information could point to where corrective actions may be necessary. Depending on the negative contribution margin due to returns, the policy may be significantly hurting profitability OR it may just be a perk that doesn't make a significant difference in the bottom line. It would be helpful to do a margin impact (since nonperforming inventory is liquidated to offset some of the cost) curve based on the difference between time of sale and time of return to see if the returns are a significant hit to profitability. 

2.  Fast Fashion- In the mid 2000's with increased globalization and textile outsourcing, the fashion game changed. Zara is often noted as the pioneer of "fast fashion", an idea that fashion should be as responsive as it is innovative. Unlike traditional fashion labels that produced two main collections a year, fast fashion concepts such as Zara, H&M and Topshop design, manufacture and deliver many collections over the year to drive traffic. The two main determinants of fast fashion are short production and lead times and highly fashionable product design. Here's more from a Wharton paper:

"Short lead times are enabled through a combination of localized production, sophisticated information systems that facilitate frequent inventory monitoring and replenishment and expedited distribution methods. 

The second component (trendy product design, Enhanced Design) is made possible by carefully monitoring consumer and industry tastes for unexpected fads and reducing design leadtimes. Benetton, for example, employs a network of "trend spotters" and designers throughout Europe and Asia, and also pays close attention to seasonal fashion shows in Europe." Cachon and Swinney, "The Value of Fast Fashion".

The benefits of quick response strategies influence consumer behavior by reducing the frequency and severity of season-ending clearances. Enhanced design capabilities result in products that are of greater value to the customer in the present time and exploit this greater willingness-to-pay by charging higher prices on trendier pieces than on basics.

In contrast to the fast fashion front-runners, Coldwater Creek explains its merchandising process in the 'Risks' section of its 10-K:

"On average, we begin the design process for apparel nine to ten months before merchandise is available to consumers, and we typically begin to make purchase commitments four to eight months in advance. These lead times make it difficult for us to respond quickly to changes in demand for our products...

Our inventory levels and merchandise assortments fluctuate seasonally, and at certain times of the year, such as during the holiday season, we maintain higher inventory levels and are particularly susceptible to risks related to demand for our merchandise. If the demand for our merchandise were to be lower than expected, causing us to hold excess inventory, we could be forced to further discount merchandise, which reduces our gross margins and negatively impacts results of operations and operating cash flows." 

Coldwater Creek 10-K. 

Essentially, the Company is making bets about what will work in fashion 4-8 months in advance. For equity holders, that is like investing in a company that buys forward contracts based on fashion trends!

I am not suggesting that an apparel designer and retailer aimed at women in their 30's to 50's ought to carry pieces trendy enough for teenagers, but it does help to be more nimble when it comes to responding to consumer demand. And in order to be more reactive to consumer demand, faster lead-times are essential.

Of course, there is a trade-off: faster lead-times and enhanced design require giving up a lot of design control and following trends in an efficient way. For example, Zara's designers work to imitate fashion, rather than innovate fashion. Only the fabrics are ordered before the season starts due to long lead times, but even those are ordered uncolored so there is flexibility on changing them right before the order. Suppliers have more autonomy in design, so they become more of strategic partners rather than just an operational necessity. Zhelyazkov, "Agile Supply Chain". 

3. Technology- As a retail grows into servicing customers through multiple channels, it's IT system becomes incredibly important. Macy's merchandising initiatives called My Macy's and Omni-Channel were initiated in 2009 as a way to delight customers at a local level and provide a a seamless shopping experience regardless of the channel. Of course, this required significant IT capacity. 

"A single platform or visual merchandising software enables retailers to also extend the reach and relationship with their product suppliers, giving them insight into the merchandising process and ensuring they are able to act more quickly and sharpen their merchandising execution."  Retail Customer Experience

"Most retailers cannot match Macy's spectrum of Omni-Channel initiatives because they have not created the multi-year master plan needed to achieve it. Many retailers have not allotted the investment required to create accurate, real-time views of inventory, order management, supply chain." Seeking Alpha

Migrating everything over to a single ERP platform from various databases can be very costly and operationally awful; just ask Levi Strauss about its SAP disaster! In addition to an ROI analysis, it would be useful to examine scenarios where such a migration could go wrong and assess the possible impact of the worst case scenario before deciding on a solution. 

Lastly, an important public service announcement: Many firms lose touch with employees on the field and their customers as they grow. It should go without saying that employees in the front lines should be empowered to voice their concerns, those opinions should be an important part decision making process, and the reasoning behind those decisions should be freely shared with the employees. "Store employees can in turn provide faster feedback when a campaign has been executed so corporate has a clear understanding of compliance, the impact an accurately executed campaign has on sales and the customer feedback from that geographic market," Retail Customer Experience. Open communication within a company goes a long way and could make the difference between success and failure.