Interviews Insights from 50+ Network Engineers & Heads of IT

When we founded Lightyear, we set out to build the ultimate IT infrastructure procurement and management solution on the web. Our team noticed that network engineers would spend tons of time on tedious, offline procurement processes (or work with agents) to buy their telecom and had the thesis that we could build software to make this process smoother and faster.

Problem? Although our team has decades of collective telecom knowledge, no one on our team has been a network buyer within an enterprise! Before we could build the best product, we had to study customer needs intimately.

We had to get into our customers’ heads.

And so we set up many, many discussions with network engineers and Heads of IT at companies large and small. In process, not only did we validate our product, but also we noticed a ton of common threads between organizations. Here are some major themes we’ve noticed.

Price discovery and ISP serviceability are data problems

When a company needs a new dedicated internet or WAN circuit, this forces them into a tedious information gathering process that can take weeks to complete. First, the network engineer must determine what internet service providers (ISPs) are on-net / near-net. There are no tricks for this - it often involves phone calls to ISPs, imperfect interpolation with a tool like BroadbandNow, or discussions with a property manager. Regardless of path, time must be spent with each individual vendor to spec out needs. For this to go smoothly, prior relationships are required with individual vendors.

Then comes price discovery. ISPs will take days (or sometimes weeks) to provide pricing. Often, a back-and-forth is required to negotiate the price into the zone of reasonability and validate items like SLA, transport, or install interval. No matter what path you take, a single circuit can take hours of your time over weeks to price and purchase!

This is insanely frustrating, as all of these activities add no value to your operation and take time away from more strategic activities. Any tool we’d build would need to handle these problems with grace, as they’re the ultimate time suck.

Most IT leaders are not a fan of agent-driven procurement

To avoid the tedious process above, one can work with a telecom agent - kind of like a travel agent for telecom procurement. See what we’ve written on agents here. When they work well, they’re great. The agent will handle address validation and competitive bidding for you, for free (they take a commission from the vendor).

However, only one buyer we interviewed (out of 50) enjoyed working with agents. Why? Incentive misalignment causes problems. Agents often prioritize specific providers over others (to get more $$$), play games with billing, and may not do the best job of negotiating for you. Buyers we’ve spoken to just don’t trust agents. Discussions would often involve at least one agent “horror story.”

Further, agents are pretty much out of the picture post-installation. They tend to have an “if it’s billing, don’t touch it” mentality which can bring negative value in a maintenance or downgrade situation.

Building trust with buyers and reducing human incentive misalignment (with software) would be extremely important to our success. We’d also have to ensure our product isn’t transactional in nature - we’d need to support the lifecycle of a customer’s needs regarding a contract.

Implementation is harder than price discovery

Signing a contract is just the start of the battle. Then, installation needs to be scheduled, and it needs to be managed so that it actually happens as contemplated. This is harder than it sounds. ISPs often delay timelines and poorly communicate project responsibilities. A project that involves any construction at all is very unlikely to get done on-time. Problems buyers often outlined involved 1) poor expectations set upfront, 2) poor communication around changes in timeline, 3) poor outlining of specific responsibilities.

For software to be valuable in this context, it’d need to handle project management with a user- friendly front-end: setting expectations upfront, communicating changes clearly when outlined, relaying responsibilities, and escalating customer problems to the ISP if needed.

Vendor organization is the exception, not the norm

The average enterprise has many locations with multiple circuits and services purchased per site. This means tens or hundreds of individual contracts. Unfortunately, the best solution we’ve often seen to organize all of this is a Google Sheet (yes, really), even in some of the biggest companies on the planet.

The most organized companies we’ve come across know what they’re spending and what vendors are used where, but aren’t actively tracking contract expirations, rebidding contracts when they expire, and couldn’t pull contract documentation without spending hours if you asked.

The least organized companies we’ve come across have no idea what is where and couldn’t rebid their infrastructure if their lives depended on it. We’ve come across public companies that fall into this bucket.

Many buyers we spoke to were more excited about a solution to handle vendor organization moreso than a solution to handle procurement. This was extremely important for us to keep in mind.

On-prem and legacy products are still status quo

The tech press will tell you that software is eating the world, that public cloud is killing private cloud which is killing on-prem servers, and that SD-WAN has killed MPLS.

Although all of these nouveau products are fast growing, the vast majority of enterprises we’ve spoken to still have LOTS of legacy systems in place. And they’re still investing in them! You’d be surprised how many enterprises we’ve come across with old, on-prem servers that are looking for a new in-house PBX system or new MPLS circuits.

Even in 2020, major enterprises are only running ~25% of their applications in the cloud. Legacy network services are difficult to transition from, particularly in a large enterprise.

This is good for us - there is a lot of business to be had in helping enterprises transition away from legacy systems to more flexible, cloud-based solutions, but for us to succeed we’d need to ensure we have a deep understanding of the legacy frameworks they’d be transitioning from.

Cool story, so what’d you do with this?

We created a product that addressed some of these issues!

Our software allows you to spec complicated, multi-location networking solutions in <5 minutes and handles address validation and vendor interactions for you. We negotiate down before you see pricing. We organize contracts for you automatically and keep track of expirations, documentation, and specs without you having to lift a finger, and we project manage implementation of new projects for you.

Most importantly, we’ve built a culture of listening to our customers. All of the most exciting aspects of our product have been identified based on a customer interaction.

Check us out - if you’re buying circuits or WAN at some point, fill out our questionnaire. If you’re a network buyer who would be open to talking, I’d love to hear from you - dennis AT lightyear DOT ai.

If you enjoyed this post, check out our interviews with two women in network engineering: Deirra Footman & Serena / SheNetworks!