Autoscaling on “the cloud” is changing the game for online ventures

Entrepreneurs share a tense tangle of guesswork when it comes to the variables they weigh before entering a market. In the age of screen consumerism, one crucial variable affecting online businesses is computational capacity—how many computer servers might it take to handle a spike in transactions?

Enter autoscaling, the recent computational innovation that allows “the cloud” to automatically expand or narrow computational resources to exactly meet demand.

Research by Foster marketing professors Amin Sayedi and Jeff Shulman and doctoral student Amir Fazli predicts the effects of autoscaling on profits, pricing strategies, market entry decisions—and, ultimately, customer welfare.

The verdict is a mixed bag for new ventures.

Shulman, the Marion B. Ingersoll Professor of Marketing, says the model indicates that autoscaling could lead to a decrease in the number of Internet entrepreneurs entering a market. But this doesn’t necessarily guarantee greater profits. Autoscaling also could spark fierce price competition.

All of this—the promise of greater speed and lower prices afforded by autoscaling—is good news for cloud computing providers and, in many cases, consumers.

What is autoscaling?

Until recently, web- and mobile-based business owners have had to guess about customer demand and online traffic. They often overestimated and paid dearly for excess computational capacity—the number of servers it takes to cover all the transactions in a given day. Or, they underestimated and saw their servers crash from online overload—in effect shuttering their store just as the longest queues of customers stood ready to buy.

Autoscaling offers a compelling solution. Leading cloud-computing providers such as Microsoft, Amazon and Google now offer sophisticating scaling services that dynamically adjust to the size of the demand.

Problem solved, right? Not quite. Autoscaling, the Foster researchers discover, is also changing the rules of competition.

Shifting competitive landscape

Sayedi, an assistant professor of marketing and Michael G. Foster Endowed Fellow, explains that autoscaling is fundamentally altering firm decisions at the most strategic level. For instance, the decision to enter a market should be made based on whether a competitor is also entering the market—or is already there.

This is a big change. Before autoscaling, markets with low probability of success and high computational needs were not conducive to new firms. But now that autoscaling reduces the risk of failure, at least one firm will want to take a chance on even an iffy market. This is good for customers.

On the other hand, the anticipation of fierce price wars resulting from the presence of autoscaling can dissuade new firms from entering lucrative markets, potentially leading to monopolies which would not benefit customers.

Price wars

Sayedi elaborates on how, when multiple online retailers compete head-to-head, autoscaling can escalate to a price war.

Prior to autoscaling, he says, companies tended to pre-purchase limited server capacity allowing them to serve a share of their market. This allowed multiple competitors to avoid head-on competition and to succeed selling similar products. But with autoscaling, there are no capacity limits; the first to market can, in theory, serve every customer.

But what if competitors enter concurrently? In this case, autoscaling produces a classic “prisoner’s dilemma.”

“Let’s say two people want to fight,” Sayedi says. “You give them the option of using a knife. If one takes a knife then the other must also take a knife. But now that you’ve given them knives, both are in a worse situation than if knives had not been available.

“Autoscaling works the same way, allowing you to scale up really quickly, but also intensifying competition quickly.”

So even though autoscaling mitigates the risk of failure associated with computational capacity, it could ultimately lead to lower profits through increased price competition.

The forecast: much more cloud

Autoscaling certainly creates new opportunities for entrepreneurs. Markets that previously required heavy investments for entry are now available to all.

On the other hand, Sayedi, Shulman and Fazli demonstrate that autoscaling actually makes it more difficult for entrepreneurs with similar products to compete in the same market. With everyone deploying the same technology to ensure they can scale to customer demand, they can encounter price wars and “prisoner’s dilemmas” that might prevent multiple firms from entering.

So the authors advise, in the age of autoscaling, that entrepreneurs should do a more thorough competitive analysis before launching a new product or entering a new market.

“Our findings highlight how startups should use information on cost of entry, probability of success, cost of capacity and level of differentiation to evaluate their entry decisions, capacity commitments and pricing strategies in the presence of autoscaling,” the authors conclude.

“The Effects of Autoscaling in Cloud Computing on Entrepreneurship” is forthcoming in Management Science.