At ONUG Fall 2026 in New York, REAP asked the people who stopped at its booth one question: which of three problems costs your network team the most? 126 network, security, and infrastructure professionals from 96 companies took part, 50 of them directors, vice presidents, executives, or architects. Answering was optional, nothing depended on the answer, and the results are reported here in aggregate only.

81 of the 126 answered. This is what they said, and what we make of it.

The three problems, and a fourth answer

  • Configuration looks right but is not: 25 of 81 (31 percent)
  • Incidents take too long to explain: 25 of 81 (31 percent)
  • Changes close without proof: 13 of 81 (16 percent)
  • Not sure yet: 18 of 81 (22 percent)

Two things stand out.

The first is the tie at the top. Configuration that looks right but is not, and incidents that take too long to explain, are the same problem seen from two ends. Drift is the cause that nobody can see; the long investigation is the cost of not seeing it. Nearly two in three of the people who answered (50 of 81) put one of those two first, which matches what we hear in walkthroughs: the team is not short of alerts, it is short of evidence that connects a symptom to a cause.

The second is the fourth answer. Nearly one in four said “not sure yet.” We read that as honest rather than evasive. If nobody measures how long an investigation takes, or how many changes close on a checklist rather than on a verified state, the question has no answer. The network operations baseline worksheet in the engineer guides exists for exactly that group: seven numbers a team can write down in a week, before anyone argues about tooling.

By role

The 126 participants came from 96 companies. About one in three was an engineer or other individual contributor (42 people), one in five a vice president or above, including founders (27), one in five a director, manager, or team lead (26), and 11 were architects. About one in five came from financial services, which is ONUG’s audience as much as ours.

The answers shift with the role:

  • Executives (21 answered) split evenly between configuration that looks right but is not (8) and incidents that take too long to explain (8). Only 2 put changes closing without proof first.
  • Engineers and individual contributors (22 answered) spread across all three almost evenly: incidents 7, configuration 6, changes 6. The change problem is felt on the floor and barely registers two levels up.
  • Directors, managers, and team leads (19 answered) were the group most likely to say “not sure yet” (7 of 19). That is the layer asked to report on the work without always having the numbers to report.

If you run a network team and your engineers say changes close without proof, they are not being dramatic. They are describing the part of the job the dashboards do not show.

Who asked to see it

The form’s second question was whether the person wanted a walkthrough of REAP. 27 of the 126 (21 percent) said yes, 21 at the booth during the event and 6 on a call afterward. Naming the problem and wanting to see a fix went together: all 27 had answered the first question (25 named one of the three problems and 2 said “not sure yet”), and not one of the 45 who left it blank asked for a walkthrough. The people who asked skewed toward incidents (10) over configuration (8) and changes (7).

We take one practical lesson from that for the next event. The question is the conversation. Asking it first, before the demo, made the demo better.

What we are doing with it

For each of the three problems there is a working template in the engineer guides, each emailed on request from its page:

The white paper, “Why REAP”, is the short version of how REAP takes those three problems on: change impact analysis before a change runs, verification of the network before and after, and incident reasoning that carries the evidence from the first alert to the change that caused it.

And if your answer at the booth was “not now,” that is fine. The templates work whether or not we ever talk.

A note on the numbers

The question was asked on a short form at the REAP booth, which also entered people in the booth’s prize drawing; answering was optional, one response per person, and the answers had no bearing on the drawing. The responses are self-reported, from people who chose to stop at one booth at one conference, so this is a sample of the curious rather than a survey of the industry. No participant and no company is named here. The talk REAP gave at the same event, “The Change Succeeded. Did the Network?”, with its slides, is on the talk page.

Add your answer: the same question, one tap, no sign-up.

Frequently asked questions

How was the poll run?

On a short form at the REAP booth at ONUG Fall 2026 in New York, October 28 to 29, 2026. The form asked one question, which of three problems costs your team the most, and whether the person wanted a walkthrough of REAP; it also entered them in the booth’s prize drawing, which the answers did not affect. One response per person; 126 people took part and 81 answered the question. The responses are reported here in aggregate only.

Is this a survey of the industry?

No. It is a sample of the people who chose to stop at one booth at one conference, so it says something about what network professionals are thinking about this fall, not about the industry as a whole. No participant and no company is named.

What does REAP do about the three problems?

REAP analyzes the impact of a change before it runs, verifies the network before and after, and reasons through an incident from the first alert to the change that caused it, with the evidence attached. The white paper, “Why REAP”, is the short version; the engineer guides carry a working template for each of the three problems.

All postsTodd Yancey · Chief Commercial Officer