Saturday, May 11, 2019

Upside from "Digital Transformation" Will be Hard to Find

One key problem when assessing the success of “digital transformation,” broadband or “economic development” programs is that there are so many simultaneous potential drivers that we cannot ever be sure what contribution any single factor has provided.

For example, one might note that European telecom companies (and their equity values) have not done as well as “technology” or “media” firms since 2007. But that also is likely true in many other regions. We cannot separate regulatory frameworks, firm decisions and strategies, underlying economic growth levels or differences in any region’s density of firms in any technology, media or telecom segment that is showing higher growth rates.

Beyond that, even examining results at firms such as Telefonica, which most observers would agree has had an aggressive posture on “digitalization,” there is a known productivity paradox, where investments in technology do not produce measurable productivity gains, even after a decade of widespread deployment.

In fact, a recent study has shown no correlation between broadband and economic growth.

One can argue that such gains will eventually show up (though it could take a decade or more). One can argue the gains will, in fact, not show up. One might even argue we cannot measure the gains. The point is that “automating existing processes,” always the first step in applying any information technology, might not produce measurable gains.

After time, when entire business processes are retooled, it might be possible to show gains. But that takes time. We sometimes forget that Amazon, for example, was founded in 1994. But not until 2015 did Amazon pass Walmart (once the largest U.S. retailer) in terms of market value.

Some 12 years later, Amazon’s total sales were less than $11 billion, at a time when U.S. retailers had sales of $252 billion, but almost no net income. That is by choice, of course. The point is that a long passed before Amazon became a fearsome and feared competitor, using a “digital” business model.



Measurable gains (revenue, profit, sales) might be hard to quantify, in the retail parts of the connectivity business.

No comments:

Costs of Creating Machine Learning Models is Up Sharply

With the caveat that we must be careful about making linear extrapolations into the future, training costs of state-of-the-art AI models hav...