The Risks of Going to Market Without a Billing Strategy

Today giant carriers have two things in common from a billing perspective. They all have the same vision of being an end-to-end, 21st century IP carrier providing all network-centric applications. Second, they haven’t a clue as to how they are going to bill for these IP services—or even what they are going to bill for, in some cases.

All the big carrier CEOs and marketing VPs use the same “Wall Street Network” slide at telecom conferences to pitch their IP net-centric vision. The carrier’s customer value solutions are built on a variety of routers that aren’t quite ready, quality-wise, and of course fiber supporting lots of lambdas (optical wavelengths) as illustrated in the figure below.

Ask these carrier executives where’s the billing and other OSSs in their “Wall Street Network” and typically they answer: “As soon as we get the network up and running we shall address billing,” or, “Our billing vendors say they can bill for it when we are ready,” or, my favorite, “We’re just going to bill the customer on gigabytes sent per second!” The shortcomings of these answers are as follows.

1. First Priority Is to Get the Network Up and Running

An IP net-centric carrier with this strategy is headed for failure. Look what happened to the last everything-over-one-network idea, ATM. Every major revenue-producing carrier today has ATM in its backbone networks, but little ATM integrated access to the customer site or ATM to the customer desktop. Why? Billing and the other OSSs were never addressed in a meaningful way by the ATM network engineering community at implementation time (see “The Achilles’ Heel of ATM—OSS”, Billing World, May 1999).

How did this happen? In the early 1990s network engineers saw a problem—the need for higher bandwidth for the emerging Internet backbone networks. Engineering convinced marketing that they could solve this problem with already available fiber and a new technology, ATM multiplexers (MUXs), while creating a network that could do everything: voice, data and video. Also, ATM backbones could support a new networking concept called frame relay. So ATM networks were pushed into service with little attention to billing or service provisioning. Today’s result is that ATM is a flat-rate service limited to providing dedicated circuits (constant bit rate, private virtual circuit services) and low-speed frame relay services. ATM today has little of the flexibility needed to handle high-bandwidth, on-demand applications except for integrated access. Had thought been given to billing and the other OSSs, IP-centric networks may not even be needed, and this subject would be moot.

2. Our Vendors Say They Can Handle IP Billing

The problem with this statement is that it’s an answer to the wrong question. Yes, if a carrier can gather all the customer-generated events in an IP network, filter and correlate them, run them through a rating engine and finally hand them off to a billing system, then the vendor could very well bill for it. The problem is carriers can’t perform event processing, so today the IP services are billed at a near flat rate. (see Rebecca Diamond’s Billing Strategies for the Application Service Provider Marketplace ” in this issue, page 16).

3. We Are Going To Bill for Service by Gigabytes per Second Used

This won’t fly with users, neither commercial nor residential. Today users spend roughly 80 percent of their telecom budgets on voice. Voice expenses can be allocated right down to the specific user and call made. Data expenses are roughly 20 percent of budget on average, and the pricing is essentially flat-rate. Corporate users can allocate data costs via a number of simplistic metrics such as data terminals per department in use and so on. As the expense paradigm reverses, as pitched by the IP-centric carriers, to 80 percent for data and 20 percent for voice, the requirements for more accurate cost allocation metrics for data will go up, but so will the need to bill based on IP QoS as discussed below. In addition, corporate IT departments in this new era of IP bandwidth on demand will require tight authorization by individual employees and applications, as well as on-demand estimates of what network usage will cost, before most IP transactions would take place.

So Why Does a Carrier Need a Billing Strategy Before Going to Market?

The risk for an IP-centric carrier is building a network that it cannot bill for! Billing World’s mantra bears repeating here: “If you can’t bill for it, then it’s just a hobby.” A billing strategy will force carrier executives to face reality. Yes, the “Wall Street Network” creates investor interest and drives stock prices up, but in the long run carriers start believing their own hype—“everything over IP!” The hard reality is this: IP is not a market, but a multiplexing technology that has great economics and provides value added when sending data but is horrible (OK, problematic) when sending live video or voice.

Having a billing strategy before going to market also brings into focus a number of strategic marketing issues. Here are three, for starters.

1. Every IP Application Is Usage-Sensitive

Before you can bill for a network service, a carrier has got to have a price for event rating, among other things. Before a carrier can price something, it has to know the costs of providing service or it could very well come up short financially at the end of the day. Many times I have heard the ridiculous statement, “Since there is so much bandwidth out there and fiber capacity is doubling every 12 months, that bandwidth will soon be free!” This statement compares with saying that the government is going to give farmers free wheat seeds so bread will soon be free at the local grocery store. It requires more than fiber to provide bandwidth or create gigabit networks. It takes ATM backbone switches, gigabit routers, OSSs and very smart people who require continuous and expensive training.

More to the point, bandwidth is not likely to be treated as “free” anytime soon, because it is human (customer) nature to take advantage of services a carrier cannot monitor for billing or cost reallocation. To further examine the ridiculous thought of free bandwidth, go to the top of the “Wall Street Network” chart and examine video streaming. This application has just taken off in the last year because of low-cost equipment and newly available software. Now look the impact video streaming is having on the Internet. When a non-telecom-specific publication like the New York Times runs an article on the front page of the business section entitled “Multimedia Transmissions Are Driving the Internet Towards Gridlock” (Aug. 23, 1999), it confirms that this is truly a real problem.

What’s the cause? Video streaming uses UDP (User Datagram Protocol) for transport as opposed to TCP (Transmission Control Protocol) for data transport. When the Internet gets congested and packets get dropped, TCP slows the terminal output down to near zero until the network frees up, then the terminal increases its output. Video streaming (as well as voice traffic) keeps pumping out UDP packets regardless of whether they get through or not, which leads to network congestion as well as freezing out other data (TCP) traffic. So if a carrier cannot bill for a service like video streaming on a usage basis, bandwidth resources will be gobbled up, or worse, carriers will have to over-provision bandwidth to support QoS. If you cannot bill on a usage-basis, a service like video streaming would become a financial loser.

Fundamentally every resource in the “Wall Street Network” is usage-sensitive—from bandwidth for all applications, through computer MIPS/storage, and very smart human resources to manage content and software on a per-customer application basis. The bottom line is: every resource used must be billed on a usage basis from day 1.

2. The Easier To Bill For, the Higher the Build-Out Priority

Billing folks recognize that billing for IP is simple, once IP network events are collected and rated. So why is usage collection and rating such a hard problem? Native “IP” was designed for routing, not managing; there is no single point to collect information; IP records are short-lived and produce enormous CDR flows, and so on (see “Usage Collection in an IP OSS,” Billing World, March 1999, by Drs. Matthew Lucas and Ori Cohen). As a result no billing vendor can do it all, until mediation devices or IP sniffing technology has progressed.

But while no billing vendor or event gathering software partner can do it all, many are staffing up and picking the low-hanging fruit like ASP support (see Billing Strategies for the Application Service Provider Marketplace page 16) My advice for a starting point on a carrier billing strategy: Take the “Wall Street Network” and list the value solutions (content management, application hosting, etc.) in one column. Then create a matrix of network elements, people, computers and transport resources where usage events are created; then compare to see what billing vendors best fit the package. Also, if the cost of developing billing for a particular service would exceed service revenue for years, place that IP service as low priority, and so on. And of course let’s not forget to look at the legacy circuit-switched/TDM billing systems to check what can be utilized.

The above may sound like a strategic planning no-brainer. But the reality is that rarely a day goes by at Billing World when we don’t get a cold call from someone asking, “We are building an IP network—what vendors perform IP billing?” If a carrier has to inquire about billing after the network is built, it doesn’t have a billing strategy! If you are in this boat, check out our survey of IP billers (see page 28).

3. The Higher the QoS Guarantee, the More Complex Billing Becomes

The key difference between today’s Internet and the “Wall Street Network” vision is the promise of QoS. Ironically QoS is the same differentiator between IP and circuit-switched/TDM networks. IP networks will never provide the same QoS for services provided today by circuit-switched/TDM networks (POTS, private line, etc.). But circuit-switched/TDM networks, on the other hand, cannot deliver the value solutions (content management, hosted applications, etc.) that IP networks promise.

So where does a billing strategy fit into the equation? Billing and discounting IP QoS is a bear (see “Billing and Discounting IP QoS,” Billing World, September 1999, by Drs. Matthew Lucas and John Yin). A carrier must gather data to determine the actual QoS delivered, create a QoS metric to present to the billing system, and create another QoS metric and present it to the customer on a bill in a meaningful way. Developing a billing strategy for IP before entering the market forces a carrier to reexamine what can be delivered regarding QoS. So when it’s time to present a customer or carrier partner a service level agreement (SLA) and prepare the RFP for IP billing or mediation vendors, a sound business strategy will be in place. IP billing and QoS must be understood! If an IP carrier cannot address QoS and billing together for a specific service, then the service is not worth developing.

Of course, billing is only one half of the OSS strategy. Service provisioning and network management are also crucial. Check out Billing World’s sister publication, TelOSSource, to get the scoop on the other OSSs, or check out our new Web site—billingworld.com—for information on how to get a free subscription.