This is the third post in a series about deconstructing Amazon’s Feedback Loop, an attempt to understand how its components work both as individual units and together as a collective system. See the previous posts, Convenience (and Diaper Stench): Amazon v. Walmart and All Convenience Infrastructures Not Created Equal.
The Feedback Loop is about pressing those levers (price, selection, convenience) for the purpose of earning the patronage of millions of shoppers that comprise the broad middle. It’s about driving growth.
Here’s the big difference between being a web retailer as opposed to a traditional one; an Amazon versus a Walmart:
To become more convenient and attract more shoppers, Walmart must build ever more stores. It must go where the customer is. Proximity between seller and buyer is a function of how far a shopper must drive to reach a supercenter.
To become more convenient and attract more shoppers, Amazon benefits from more consumers being connected to the internet with each passing day. Proximity between seller and buyer is a function of how many steps a shopper must take between his seat on the couch and the nearest web-enabled device.
The same forces that bring more people onto the internet every day bring those people to Amazon. It must simply be prepared for their business.
Scalability and the Check-Out Scenario
Those are the starkest of contrasts when considering the factors that drive convenience in the physical world versus the digital realm. And while differences in cost are big, the implications for scaling – increasing sales faster than infrastructure – are staggering.
Consider the programming logic that turns the gears behind Amazon’s 1-Click check-out process. While that code base was enormously expensive to develop, requiring high-salaried programmers to build it, test it, and improve it over months of iterated effort, it now sits on an Amazon server. It may require occasional maintenance tweaks, but by and large that digital instruction manual can facilitate millions of check-outs each day about as easily as it can conduct just one.
More importantly, that code can facilitate thousands upon thousands of check-outs simultaneously without slowing down the flow of commerce, the speed of transactions, or the convenience of quick turnarounds for shoppers.
That is ultimate scalability. The kind you get when you depend on a convenience infrastructure built on technology.
To keep picking on Walmart, let’s consider the contrast of store-based retail scalability. Its cashiers are cheap on an individual basis. Each is paid a piddling hourly wage, but each can only facilitate the check-out of one customer at a time.
If it wants to be able to check-out 100 shoppers at the same store simultaneously, it must have 100 different check-out aisles with 100 different cashiers working 100 different cash registers at the same time. Since that’s impractical, it permits long lines to form at each register during peak hours, thereby reducing its convenience to customers.
Scaling its convenience infrastructure is – for Walmart and its ilk – always constrained by physical world limitations. In the physical world it’s difficult to scale many of your convenience factors too far beyond that one-to-one ratio (like one cashier to one shopper). In the digital realm, scaling seems nearer to one-to-infinity (theoretically at least) than that one-to-one ratio.
The 1:1 Ratio Rears Its Head Again
Consider this other constraint of store-based retailing…
In 2010 Walmart added about 1,400 stores to its existing world-wide base of 10,000 (give or take a few). All that additional real estate – those stores that brought Walmart closer to more shoppers – funded a nearly ten percent increase in revenue, taking the business to $447 billion in sales. If we assume each new store cost, on average, somewhere around $10 million (to build, equip and stock each), it means Walmart invested $14 billion in the main component of its convenience infrastructure.
So, to increase revenue ten percent, it invested $14 billion and grew its number of physical locations by about 14 percent. This suggests something pretty close to a one-to-one relationship between opening new stores and growing revenue.* And while that ratio is not a precise formula for Walmart’s growth, it does highlight the natural constraints that exist for growing your business under the rules of a physical world: you have to invest significant cash to build more stores to access more customers and (finally) to grow your revenue.
This puts to a cap on how quickly Walmart can grow because everything is governed both by how much cash it has to plow back into its convenience infrastructure and how many new stores it can possibly open in a 365-day span of time.
It’s almost like a gravitational pull that keeps its ability to scale in check, making it difficult for traditional retailers to get much beyond that one-to-one ratio of having to increase its base of stores by (for example) ten percent in order to increase its revenue by about the same amount.
Walmart’s Rate of Growth Anchor
Those physical world limitations create an anchor on rate of growth as the base of legacy stores (those that have been open for more than a year) gets bigger and bigger.
Let’s consider that 1,400 new store openings push the upper limit of what Walmart (or any retailer for that matter) could do in a given year. That’s a lot of construction, requiring a lot of resources in the way of cash, management attention, and use of supply chain bandwidth. They could probably do more (in fact, I believe they have done more in previous years), but I doubt they could do considerably more on a sustained basis.
When calculating rate of growth (a percentage), new stores are the numerator and the base of legacy stores is the denominator. If 1,400 is the max new stores in a given year, the numerator is pretty much a fixed number. But the denominator grows larger with each passing year. Those 10,000 or so from 2010 become 11,400 after 2011, 12,800 after 2012, 14,200 after 2013 and so on. So, for each of those years, 1,400 divided by the growing legacy base gives us a smaller percentage (14 percent drops to 12.3 percent then 11 percent then 9.9 percent, etc.) as time passes.
The rate of growth slows. The ability to get to ever more customers is bounded by the constraint of only being able to open so many new stores in a given year.
Now this is mostly theorizing. I’ll grant that the numbers are likely a decent-sized understatement of reality. Walmart can probably open more than 1,400 stores a year if it wanted to. But not a dramatically higher number. So the general rule stands true: the base of existing stores creates an anchor on rate of growth. It will slow as the legacy base gets bigger. Trees can’t grow to the sky.
Conclusion and Segue to Convenience Barriers
The point is that Walmart operates in a world of limits, and while we can quibble about the exact numbers, the facts remain that the limits approximate (at least roughly) 1.) that one-to-one ratio for new stores to revenue, and 2.) the rate of growth slowing with a fixed numerator being anchored down by an expanding denominator.
Web retailers don’t have that same challenge. Investing in technology – those pieces of code whose logic churn across millions of server processors to transact millions of transactions – is much less expensive and much more scalable.
Over the long haul, this scale difference has a compounding effect. Much like the difference of a few (seemingly) small interest rate points for savings accounts may not amount to much over a few years’ time, the difference is amplified to dramatic proportions as time marches on and the effects of compounding take hold.
Amazon’s advantage of lower convenience infrastructure expense and better scalability means it can take those savings and invest them in making up for any deficiencies it might have in the competitive struggle with traditional retailers.
And Amazon has been very disciplined in making these investments both in the other growth levers (lower prices and wider selection) and also in attacking the bulwark of the best convenience defense store-based retailers retain against web competitors; namely, the ability to satisfy customers’ desire for immediate gratification, that ability to walk out of a store with purchase in hand.
That’s the convenience barrier, and Amazon’s has been whittling away at it over the years.
* Note that this is the roughest of calculations on several fronts, the most important of which is that Walmart’s revenue growth does not only come from new store sales. In most years, the lion’s share of growth comes from selling more through its existing base of stores (a better scaling proposition because they don’t have to invest much more in the existing stores to drive more sales volume through them). However, given how anemic same store sales growth was in fiscal years 2010 and 2011, it’s fair to conclude that new stores were responsible for most of its added revenue.
The bigger point is, however, that there exists some sort of gravitational pull – governed by the natural constraints of a physical world – that pulls store-based retailers back toward that 1:1 limitation on scale. Even if they do better than 1:1 for some time (say a ten percent increase in number of stores increases revenue 50 percent), gravity will pull down that ratio over the long haul.