In the industrial economy, “scale” was a noun. In the information economy, “scale” is a verb. The difference is important.
Richard Waters, in today’s Financial Times, discusses Google’s great expectations. The company, he writes, “is after everything” in the advertising business: “if the adverts you see are not brought to you directly by Google, it will at least have a big hand in the process.” Google’s “immodest ambitions” place the firm “among a growing band of technology companies [like eBay and Dell] that believe they can own it all”:
What all these companies have in common is a belief in method: a confidence that what sets them apart is not the services or products for which they are currently known but a way of doing things. If Dell makes cheaper PCs by collecting customers over the telephone and internet and building machines to order, then why not apply that principle to any type of electronic product? If Google has used technology to bring a higher level of targeting to advertising, why not use that brain to squeeze the inefficiency out of other corners of the advertising industry.
To put it another way, these companies have a faith in the “scalability” of their “business model.” It used to be you’d beat your competitors by achieving greater scale in your operations, enabling you to spread your costs over more products and thus push down the cost of producing each product. Scale was tangible, a manifestation of plant and equipment and other real assets. Today, you strive to beat your competitors by creating an idea or a model that can scale without constraint, expanding easily and flexibly to handle ever more business. Scalability is intangible.
Given that the concept of business scalability emerged in the tech business, particularly among dot-coms, it’s not surprising that it has its origins in software engineering. A scalable software system is, in brief, one that can easily expand to handle more transactions or other sorts of throughput. Ideally, there are no limits to that expansion. In the view of engineering-led companies like Google, building a good business is not all that different from building a good data-processing system. Create a good model (or method), and then keep running more transactions through it. The closer your business model is to a piece of software, the better the idea of business scalability works – as Google’s enormous success with its ad auction shows. But scalability, unlike scale, is a metaphor, and metaphors have the nasty habit of breaking down as you try to extend them.
As Waters notes, “grand plans such as these run into a couple of problems. One is that a process perfected in one market often does not transfer quite so neatly into another [witness Dell’s struggles in consumer electronics] … The second problem is that, as competitors catch on, they start to erode the technology or cost advantages of the pioneer – or they are quicker to apply the lessons to new markets.” A software engineer does not have to account for competition; a business manager does.
Scale and scalability both have strengths and weaknesses as business strategies. We know the strengths and weaknesses of scale pretty well by now. We’re only beginning to understand those of scalability.
Interesting metaphor comparing software scaling to business model scaling. I agree with your points.
One additional note: In the world of software, we consider software to be “scaleable” as long as the per-transaction costs (in terms of CPU, memory or other stuff) stay reasonably constant.
However, in the software world, many people realize that systems often scale linearly — but only to a certain point. Often, there is some “breakpoint” after which all hell breaks loose and the performance of the system changes dramatically. This is why enterprise software engineers do performance testing — to simulate what might occur under heavy loads.
I would argue that business models may manifest a similar phenomenon. They may be scalable — but only to a point. After that point, all hell may break lose. One big difference is that its hard to test business models and “simulate” what might happen.
So, in addition to the challenges of scaling a business model or process into another dimension/market, there’s also a challenge of scaling it even within the market a company is in now. It may simply be a case that they haven’t hit their breakpoint — yet.
Success does attract competition and prices do tend to erode.
Yet there are a few still optimistic points that tend to get eroded too in this knee-jerk piling on Google that’s popular now that the stock price backed off.
Google revenues are a bellwether of international web traffic — which will triple-to-quintuple in 10 years. Prices may erode while traffic growth mitigates that effect on profits.
No one knows how these trends will play relative to eachother. But why would ad prices be peaking when 5.6 Billion people have not yet chosen their operating system? Mind, a Google OS might double the “venues” on the desktop where ads could be placed and clicked upon. That too would ssupport pricing.
Google’s algorithms are pretty naive compared to where they will evolve. That may support prices or increase revenue-yield from same unitary traffic, or both, or either, with potential profit rate increases.
Google’s hidden competitive moves on the desktop might support ad price increases if it gums up the general works at other competitors.
I enjoyed Richard Waters piece today, and agreed with his conclusion that we don’t know what is going to happen. All the Google reporting except Richard’s has been reactionary, simple-minded, unimaginative & full of Schadenfeude.