So, You Want To Run A Wireless ISP? — continued

The basic product
The basic "product" for a wireless ISP is obviously wireless connectivity and Internet bandwidth. To achieve higher margins, you will need to augment this basic service set with professional and managed services, as well as equipment and installation revenues. All of these "products" need to be test marketed and packaged before the sales force risks their credibility by making any real effort to sell them.

In this market, pricing decisions are very difficult because there are plenty of competitors. In fact, the huge loss of market capitalization in the communications marketplace is primarily due to the effects of ruinous competition—DSL being the most obvious example.

At NextWeb, we defined our competition as the T1 line—not DSL or cable access—and used T1 local loop plus Internet access as our benchmark. Our "elevator pitch" was that we could offer 30 percent more bandwidth at half the cost, and we could install in hours not months! Because we needed to convince customers to switch from a solution that is presumably working now, to an unknown wireless product, we needed a compelling reason to switch—not just performance, not just price, but both performance and price!

Promotion is one of the hardest marketing areas to get right. It is always expensive and the results are very difficult to measure. However, if you don't do any advertising, direct mail, press releases, or marketing events, how will you build your brand to support premium pricing and attract customers?

NextWeb started with some advertising and heavy direct mail in our local market with disappointing initial results. However, it is vital to be persistent and stick with it. Only then will your marketing program start to take shape as an effective sales tool.

Speaking of sales, the fastest way to ramp up revenue is with a direct sales force—in the field and on the phones. Success is only achieved with a properly selected, trained, and motivated team of sales professionals that have a sufficient database of leads. The appropriate "sales DNA" is critical to a sales team and you will need to cycle through any non-performers.

There are many ways to generate qualified sales leads. We've utilized everything from digital terrain databases to reverse phone directories, as well as sales lists and databases. We ended up establishing a trained telemarketing staff to qualify prospects and set up appointments for our account executives to close. It was tempting to outsource part or all of our sales efforts. But setup costs were high and the sales ramp-up extended. In the long run, we knew keeping sales functions in-house would be in our best interest.

Everything is in beta
No one wants to build a business on unproven technology—it's too risky. But we had to compete against the "big boys" who are talking "quality of service," "customer provisioning," and other "state-of-the-art" capabilities.

If you think operating a wireless ISP is based on proven technology—think again! As a senior wireless executive said not so long ago, "There are too many science projects out there!" Understand that everything any wireless ISP wants to sell is really in "beta testing." After all, the technology works in the lab—most of the time—and fulfills some basic expectations. But making 802.11x technology work in the field is another matter. Expect to roll up your sleeves and dig right in to the guts of this stuff to figure it out and make it perform.

At NextWeb, we started with Adaptive Broadband—they were the leaders in UNII-band point-to-multipoint and we wanted to keep out of the congested 2.4 GHz ISM band. We learned asynchronous transfer mode (ATM) technology first and set up our demo systems. We'd been in the wireless manufacturing, installation and service businesses for many years—we thought that if anyone knew how to make this stuff work, it should be us.

We wanted to produce the reliability of dynamic configuration and maximum throughput with low latency—and rival ATM QoS. After six major system software revisions from Adaptive, we settled for building our network with static configurations and no QoS—just to build a reliable network! Many wireless ISPs simply went out of business because they couldn't get the product to work in the field like it did in the lab. Eventually Adaptive filed Chapter 11 and recently sold its assets to Moseley, remerging under the Axxcelera name.

Next, we looked at Cisco's fixed wireless UNII product. But their streaming video demo didn't stream, and the costs were astronomical compared to our business model. We ended up selecting Aperto Networks for a second vendor of our point-to-multipoint UNII-band service. Aperto offers an IP-based network with QoS capability that would be easier to configure than ATM and we hoped to use IP routers for network redundancy. We have Aperto in service now, but there is a costly price tag attached to building a mesh network, so we still don't have the redundancy we hoped for; and QoS is still to come.

For our backbone, we designed a high capacity OC-3 wireless network to interconnect our base stations and avoid paying PacBell leased line charges. Just to interconnect our base stations, we were quoted DS-3 prices of $4,000 to $5,000 a month and OC-3 prices up to $7,000 a month. We selected 18 GHz OC-3 radios from DMC-Stratex because of their excellent reputation and local service. MHS links were out of the question, due to the cost, and we didn't want to go unlicensed for three reasons. First, OC-3 bandwidth wasn't available at the time. Second, we needed maximum reliability on the backhaul. And third, we didn't want to give up any channels in our 5.8 GHz UNII band for base stations—customer premise equipment (CPE) links.

We had hoped to find a way to build a redundant OC-3 ring, but without expensive Add/Drop multiplexers there was no way to do it at the physical layer. When we looked at implementing it at the ATM layer, all the software was still in development and there were significant configuration issues. Finally, we were surprised to find that the OC-3 microwave was significantly less reliable than earlier-generation microwave radios we were used to, and we lost DMC's convenient local service when they moved it to Europe!

When it comes to wireless vendors, it pays to be skeptical! For example, no matter the hype of non-line-of-sight (NLOS) gear, if there is a 20 dB signal loss through trees and you only have a 10 dB margin, it won't work—period! Test everything in the lab first, before deployment. Talk to other providers about their experiences and make sure you have access to a high level of vendor support—you'll need it, because everything is in beta.

Supportive roles
Without credible customer service and technical support you will lose your customers to rival services. You will need 24x7 instant support to explain to customers why their connections went down or why aren't getting the bandwidth they expected. Your wireless ISP will need to respond to Web hosting glitches, e-mail account additions, and DNS updates from any customer at any hour. When a virus like Code Red hits, you will need to get everyone notified and patch the breach as soon as possible.

Our solution was to set up a virtual network operations center/technical assistance center (NOC/TAC). Everything is on-line, securely accessible from anywhere—including our phone system, network monitoring and control, trouble ticket system, and all our servers and network elements. Whenever any piece of equipment goes down the duty engineer is paged and can telnet directly into the unit. Similarly, all emergency phone support calls are routed to the duty engineer, wherever they may be. We have a service level agreement (SLA) on our web site, and a secure area where we monitor individual customer usage and performance.

It is important to set up a priority system for operations and engineering staff to follow. For example, the top priority at NextWeb is an out-of-service customer—we define a response time and the work required to meet it. The next priority is any new customer installation and configuration. Our goal is to meet the in-service promise date given by sales. Finally, network maintenance and upgrades, and company support systems have the lowest priority.

Lessons learned
There are many good aspects of running a wireless ISP business. Everything is under your own control—you own the network and consequently, you own your customers. You can sell direct and provide a wide range of value-added services.

The problems facing a wireless ISP business are not unlike those experienced by many wired start-ups. Because you are a small fish in a very big pond, the competition is fierce, customer expectations are unreasonable, and it all costs more than you think. Building the network is tough, but selling and keeping customers is just as difficult. But if you have a plan in hand, you're ready to meet the challenges on a day-to-day basis—just be prepared to adjust the plan.

—End

< Back to Page 1