In today's world of technology we don't question that Electricity and Water should be metered and charged according to usage, but with data plans the Holy Grail of data plans is the "Unlimited" plan, and every modern, sophisticated consumer expects to eventually be able to be able to purchase unlimited data or receive it for free. This expectation is so pervasive that it has worked its way into the strategic planning for telecommunications operators around the world.
Let's take a look at a world where we apply the same expectations to Electricity and Water.
Let's begin with the utilities. Electricity and water both require extensive infrastructure with finite resource inputs that are purchased by the utilities on a use basis. For example, to produce more electricity a utility must purchase more fuel, or for a water utility to pump more water they must pay to expand their access to water resources, wells, reservoirs, etc., as well as pay for more electricity, labor, and fuel to operate their infrastructure. As output scales up to meet increasing demand, so do the costs.
Now factor in a flat rate fee regardless of usage for consumers and businesses. How will that fee be derived. The simplest method would be to take the total cost of delivery and divide it into two groups - a certain percentage will be paid by dividing one percentage evenly between all businesses regardless of size or usage, and the remainder would be divided equally between consumers regardless of usage or income. Every utility user would be able to use as much or as little as they want to or need without incurring additional charges.
It doesn't take much thinking to realize that this model is 1) unfair, 2) unsustainable, and 3) flawed.
So why does it seem to work with data? Operators must still contend with extensive infrastructures and increasing input costs as demand and usage scale upward. Electricity, spectrum, upgrades, site leases, cell site densification, competition, etc. Perhaps the truth is that it doesn't work but customer expectations are driving the industry into it. As unlimited plans prevail the industry will enter a new "lean" period where operators will be experimenting with new business models to shore up falling revenues and rising costs.
A couple years ago I had to give a presentation on the business case for unlimited, but this business case only existed for the minority operator that had an excess of underutilized capacity and difficulty bringing a metered product to market, or a monopoly and the ability to charge like a public utility. The audience was skeptical about unlimited and for good reason, I honestly wasn't that into the idea myself. I had graphs and stats to support the business case, but it was obvious that unlimited data would lead to lean times for all. A quick look at the exponential growth of data usage is a good indicator that unlimited data would lead to flat growth during a time when the consumed resource was being consumed at exponentially growing rates.
The only real highlights from the presentation that garnered positive head nods were the technologies that might make unlimited plans more workable - 1) content filtering - block torrent traffic, 2) intelligent caching of video, software updates, and popular content, 3) traffic optimization, and 4) rate limiting. With the possible exception of #1, these would be good practices for any operator. Torrents in a usage based model could drive higher revenues, but users who generally use torrents tend to be lower income users or are not the primary subscribers who pay for the service (students or children).
At that time of the presentation I worked for a minority incumbent operator that had technological and knowledge/skill hurdles that acted as a barrier to implementing usage based plans. To further the unfortunate-ness of the situation, my employer was also not a monopoly. The way forward for my employer should have been to leverage the underutilized capacity of their new fiber network to up-sell business customers, where demand was high, in order to derive enough revenues to subsidize residential and wireless unlimited usage at limited speeds, while also implementing the previously mentioned technological solutions for improving subscriber experience and managing IP input costs more efficiently. Again, unfortunately, these last and final recommendations were ignored. The results were predictable... and predicted.
At the same time that I was giving the presentation I was also promoting internally the idea of moving away from unlimited services to metered services, countering the retention of the unlimited services by offering "right sized" plans that would encourage users to ration their usage while providing the buyer with a better user experience in exchange for their purchase. Unlimited services were predictably slower, suffered congestion related outages, and other service quality issues. Metered usage would have to be prioritized over the unlimited traffic as a premium level service in order to compete with the unlimited service, and in a price sensitive market - usage based packages can be sized accordingly to offer metered access at lower enough prices to induce buyers.
Onward and upward, conclusions from my personal experience are that the move to unlimited needs to be strategized well beyond subscriber acquisition numbers. Projecting out the future cost of inputs and upgrades, considering the implementation of technologies to mitigate or reduce the increase in inputs and input costs, and developing a business model that accounts for decreasing revenue per subscriber and lower returns per additional input. Major U.S. telecoms and service providers are looking outside of the traditional telecom and ISP basket to add value and increase their margins per subscriber as a result and reaction to the impact of unlimited plans. Video is an easy early target, but service providers are diversifying into home security, apps/entertainment, advertising, and "social data" products are as well. In 2017 and 2018 much of the products and services in these sectors will be only affordable to Tier 1 and Tier 2 providers, but going into 2019 and 2020 we should see commoditized versions of these products being marketed to to the Tier 3 and smaller operators at more affordable price points. Given this timeline, it is probably best for smaller markets to give the "unlimited" business models time to mature before jumping onto the unlimited bandwagon, fortunately there is an unlimited amount of space on the bandwagon, if you can manage to hang on.