Lead with the promise, not the plumbing

This week BMC Helix invited me to spend an hour with their field team and give them the honest practitioner view. No pitch, no theatre, just what I actually see when I walk into a client. Credit to them for that, because it is a braver thing to do than it sounds: hand an independent the microphone and ask them to say the uncomfortable parts out loud. Here is the short version of what I said.

I have sat in a lot of boardrooms. The moment I say “metrics, logs and traces,” I lose the room. The person signing the cheque does not care how the tool works. They care about three things: who owns it when it breaks, how fast you can tell them what changed, and what the noise is costing them in engineering time.

Lead with the promise, not the plumbing. The plumbing is a conversation for the people on the ground.

The boardroom line: the buyer signs for what sits above it, the machinery sits below. From my field enablement talk, August 2026.

I have not walked into an estate with fewer than five observability tools in five years. The busiest had eight. They overlap, they compete for the same journeys, and the politics of who owns which data slows everything to a crawl while the regulator waits.

The fix is never another tool. On one payments estate we went from four tools to two on the journeys that actually mattered, and the problems that had the regulator circling quietly went away. Not because the second tool was cleverer. Because the data finally sat in one place and somebody owned it.

The last four clients have all asked me the same thing: help me point the fire hose somewhere else. OpenTelemetry has changed how buyers think. Being genuinely open, not just “OTel compatible” on a slide, is now the difference between the vendor who stays and the vendor who is gone in eighteen months.

Openness costs you something in the short term. What it buys you is trust, and trust is the only thing that survives a renewal.

Everyone wants to talk about AI, and that is what is getting vendors through the door. But AI is a beautiful wrapper on top of the same old problem. Put it on a poorly instrumented, poorly governed estate and all you get is bad answers faster.

AI is not Harry Potter. It cannot conjure structure that was never there. Foundations first. They are not glamorous, and they are the whole game.

For twenty-five years observability has been treated as a cost centre. It does not have to be. One outcome a CTO can feel, one honest number against “if you had not had this, you would have lost that,” and you stop being a line on the invoice. You become the reason they sleep at night. That is the shift, and it is the whole argument of the book.

Observability was never about seeing more. It is about seeing what matters, and the outcome has to be felt by the customer. Show them the worst dashboard, not the best one. Tell them the truth.

Everything else is plumbing.

My thanks to BMC Helix, and to Swayam Prakash Dash for the invitation and the good questions.