Calculating the length of an internet year
We know technology moves fast these days. But how fast? And which technology?
Most folks know a “dog year” equates to about seven human years. Although this is not a particularly accurate actuarial device (little dogs live longer than big dogs), it does give us a rough idea as to when Bowser is going to be chasing that chariot in the sky.
We have other measurements, too. For example, Moore’s Law
suggested back in 1965 that the number of transistors in a chip would double about every two years. Again, a generalized barometer that has proven to be more accurate and useful than one would have thought.
So where are we at with the internet? With the web? How can we measure the maturity of our apps? Or predict business cycles in the online world?
I think I've found a useful answer. Here’s how I worked it out.
Establishing a baseline
My first task was to find a suitable industrial-era baseline. I needed to find a well-established, highly industrial segment that had demonstrated:
- a period of invention, followed by
- a period of adoption, followed by
- a definitive end to an era of pioneering, followed by
- a long period of slow, incremental innovation
- a long-term, on-going global presence
Finding such a segment is easier said than done. The obvious-in-hindsight choice was the automobile industry – a sudden inspirational flash in the middle of a presentation on innovation I was giving in Italy.
Autos and the web have a lot in common
The first commercial vehicle was a Daimler, produced in the United Kingdom in 1896. Interestingly, the first commercial websites started to appear about 100 years later. In my calculations, I will use 1993, which marked the introduction of the first true graphic browser, Mosaic
. Most experts agree that the introduction of Mosaic represented a turning point in the history of the Web – much as the original Daimler represented a turning point for those experimenting with horseless carriages.
Certainly, no one would contest that websites in 2009 cannot represent the final phase of online evolution. If we compare ourselves to automobiles anno 1909, it would seem we haven't come very far at all. If nothing else it strongly suggests that a "calendar year" is significantly longer than an "internet year".
End of the pioneering period
The era of pioneers, where most of us working in the online arena were pretty much making things up as we went along have long since passed. Today, we have pretty good sets of best practices. But when did the age of pioneering actually end? We need a date for our calculations. Although the current economic crisis has caused unimployment in our industry, it's doing that in all industries. We are not seeing the great "weeding out" of questionable practices that we saw back in the early years of this decade. From a development point of view, we need to go back to the burst of the dot-com bubble
Even given the rise of social media and other innovations during the past decade, the market reaction to events of 1998-2001 equate, from a business point of view, to many other technology inspired booms, including autos in the 1920s (other similar technology booms include railroads in the 1840s, radios in the 1920s, and transistor electronics in the 1950s).
So which year represents the end of pioneering for the automobile industry? The introduction of mass production by Ford in 1908? The U.S. entry into World War I in 1917? I pondered this for over a year before I accidentally came across a footnote in a book on antique cars that stated “The stock market crash of 1929 marked the end of the pioneering period for car manufacturers.” Conveniently, a crisis again seems to have marked the end of an era.
The similarities observed in the aftermath of both 1929 and 2001 have erie commonalities. For example, a study of the car industry suggests that there were more makes and generally better cars in the 1920s than in the 1930s. After the market crash of 1929, bad ideas (and poorly built cars) became more prevalent. And I would argue that in many semi-developed online markets (Scandinavia for instance), websites did
deteriorate in quality during most of the decade following the dot-bomb; while pretty, these applications did not successfully build the shared frames of reference needed to establish credibility, trust, and a willingness on the part of site visitors to deal with these business entities. Indeed, there is still far too much "brochure-ware" polluting the ether.
But back to 1929. If the Wall Street Crash marked the end of the pioneering era in automobiles, it should be possible to work some numbers.
Doing the math
If 1896 and 1929 mark the age of pioneering for the auto industry, we have a period of about 33 years. And if we accept that 1993 and 2001 represent watershed years for the Web, that works out to seven years (late 1993 to early 2001). The months of introduction are critical when calculating the length of the pioneering period for the Web as the difference between seven and eight years has far too significant an effect on the calculations.
Setting up a simple ratio, we find:
And that means X = 4.7 years.
Proof of concept
So is 4.7 years a viable figure? Just like Moore's Law, Reiss' Law will be proven or disproven by history - it is impossible to provide hard proof. But the anecdotal evidence is already compelling. For example, economists put the average business cycle (as defined by Burns and Mitchell
) somewhere between 3.5 - 7 years. This appears pretty much the same for cycles triggered/ended by exo- and endogenous causes. The average time needed for a traditional business to establish a business model, gain goodwill, and prove its worth is about eight years – longer than a single business cycle, but usually shorter than two.
So if my number works, one calendar year should roughly represent slightly less than one complete business cycle in the online environment.
Curiously, if you look at the online ventures that have succeeded this past decade, you’ll find that an incredible number of them have been sold or expanded their ownership base within the first two years of operation – which fits surprisingly well with the timing of the offline experience, using my 4.7-to-1 adjustment; if we look beyond the get-rich-quick IPO mania of the late 1990s, many successful offline ventures typically seek alternative financing early into a second business cycle.
If you look at the online ventures that have failed this past decade, you’ll find that the same cyclic pattern repeats – ventures have roughly two calendar years to make it or break it. We, of course, knew this from our emprical observations over the years. But using the numbers I’ve laid out here, it’s easier to see why this is so from a business-economic standpoint.
Of course, I haven't yet identified the triggers that mark the beginning or end of an online business cycle. But I'm working on that, too.
So where are we now?
If we compare, for example, websites to cars, we’re 15 x 4.7 years into our development. With a calendar starting point in 1896, that puts current web development on par with the car model year 1960.
If we continue to use cars as our barometer, we can see that a number of things have been invented and standardized – the number of wheels, shift patterns, basic controls (pedals, steering wheel), the placement of heating and ventilation controls, etc.
In web terms, perhaps this suggests that many of the basic navigational devices we use today will be around for some time. But it also leaves us wide open for innovation. For example, web servers account for an incredibly high proportion of CO2 emissions – almost as much as the aviation industry according to the UK’s Health Protection Agency task force
Should we be using gray backgrounds rather than white to reduce electrical consumption? Maybe AJAX isn’t a good idea seen from a sustainability point of view.
If we look at the development of the automobile these past 50 years, two issues really stand out: safety and fuel economy.
So my question to you is, what are OUR safety and fuel economy issues? And how long do we have to make these improvements? Can we use my magic number to predict the future of our industry by examining the past of other industries?