The ISP Failure Checklist: 4 Questions Before You Sign

The IT director had a signed contract, a construction date, and a vow: new fiber into the Tampa branch by end of the first quarter. In June, the provider hadn't even pulled permits. By August, the emails went unanswered, and the office was sharing a cable line that turned to sludge during every afternoon thunderstorm.

It's not an uncommon tale. With multi-site business internet, you are not actually purchasing access to bandwidth - you're actually buying the operator's track record of implementation, over a number of years, at all of the locations on your list. The brochure doesn't reveal this. The record does.

What questions should you ask an Internet Service Provider before signing a multi-year contract?

The four questions: How many markets is the provider simultaneously building? How long have installs actually taken at similar sites? Who owns the last mile? And what do you do when day two the circuit goes down? Doug Dawson, from the consulting firm POTs and PANs, observes that ISP launch plans have 1,500 to 2,000 discrete tasks, so if your provider isn't ready to give you details on operations, they probably haven't gotten through the list.

What new broadband ISPs don't do - and how the bill gets sent to your branch office

Dawson enumerates the scenario he continually observes. The single most common failure mode, he states, is overexpansion: "The most common cause of ISPs failure is the expansion of engineering, permitting and field staff beyond a single market to a level where they can't complete a market in a normal manner."

Dawson: the second pattern is timing optimism. New players often try to contract their first customers six months after starting construction, which is just not realistic given the permitting backlogs, supply chain lead times and the straightforward challenges of recruiting technicians in a new metro.

The third is scale. A launch plan that contains 1,500-2,000 line items has a long tail of clutter, boring little jobs that are easy to push off once a schedule starts to fall behind. As Fails and Pan reports, the cost breaches and the sluggish customer acquisition then cause teams to cut corners and those shortcuts cause the network to go down first. And in today's world of social media, public reviews, and blog posts, that reputational damage can endure far longer than the outage itself.

This is what you care about. That's not the provider's problem alone. If your provider is six months behind on its own build, then your branch opening wriggles in along with it. If they're skimping to get their work done on cash, then your circuit is the one that flaps at 4 p.m. on a Friday. You pick up that execution risk the day you sign.

The four-question checklist for any provider pitching your building

In the first sales call, leverage these The answers differentiate operators from optimists more rapidly than any speed test.

Question Weak answer Strong answer

What markets are you currently building? "We're building out everywhere." One sentence listing each market and the local crew and spectrum or fiber assets.

How long has it installed at locations just like mine in the last 12 months? "Commonly 30 to 90 days." A prepared window with the weather reported - authorizes, permits, riser access, rooftop line of sight.

Who owns the last mile and the routing? The incumbent. The incumbent. Do they own it?, with a mention of the physical path end to end.

And on day two, if the circuit fails: "Open a ticket" A named SLA, a monitored failover path and a 24x7 team with engineering power.

Another question to ask if you have more than three sites: can I view all my sites in a single portal, on a single bill, under a single SLA? Disparate vendor management isn't a procurement issue. It's the cause of 4 hour triages versus 20 minute triages.

What an operating record actually buys you

Dawson's article, in one reading, is a case for boring longevity. A provider that has already suffered the costly consequences is living with the 1,500 item list. Permitting is a process, not a shock. The field techs are out there in the marketplace.

And that's the story WiLine tells about itself at the table here. We've been building and managing our own network since 2002 as a Silicon Valley fixed-wireless pioneer, and we own the fiber, the wireless and the routing gear-rather than resell someone else's. Due to our ownership of this infrastructure, there's no reseller markup in the middle, and support engineers can chase down an issue all the way through rather than logging a trouble ticket with a carrier.

It also changes the install math. Fixed wireless on FCC-licensed spectrum doesn't require a trench, so it takes 5-15 days to place in a 10G symmetrical link, instead of depending on a street cut and a construction permit. When deploying to multiple locations, that translates to opening a branch on time versus opening a branch with a Wi-Fi hotspot in the lobby.

And it provides you with a secondary physical route. A rooftop wireless link can't be affected by a backhoe a couple of city blocks away. When you add SD-WAN and sub-second failover, it's the diversity route that ensures the site remains online even when the pipe in the street goes down.

That footprint: 600+ U.S. markets, 12 metro markets including San Francisco, Los Angeles, Phoenix, Chicago, Atlanta, Tampa and Philadelphia, an 800G optical backbone, and 3,500+ business customers. In January 2026, Network Infrastructure Magazine named us Top Business Internet Solution for 2026. It's not about the plaque. It's about that we built a network that is part of an infrastructure that has stood the test of two decades of the failure modes (crises, disruptions, wars, bubbles, dinosaurs) that Dawson described.

Your next step

Prior to your next renewal, grab your site list and give each location a grade on three criteria:

1. Who physically owns the last mile?

2. Is there a second path that's not shared with the first?

3. How fast can a new site truly be lit?

The majority of IT teams will have at least one branch where the best answer is "we don't know."

Begin with these. Then pose these four questions to each doctor in your running - yes, to your current family doctor, too.