Anamap Blog
Product Data Interview: Daniel Koeck
Interviews
Updated 2026-11-18

What should you build first if you want developers to provision their own virtual machines? Technology product manager Daniel Koeck expected the answer to be more control over where those VMs run. Customer configuration logs and follow-up interviews pointed to something developers would rarely see. Infrastructure teams needed a way to enforce placement rules before they could let anyone provision a VM on their own.

In this installment of our Product Data Interviews series, Daniel explains how that discovery changed the roadmap, why a handful of rules can matter more than thousands of configurations, and how he learns from users when production telemetry is limited.
Connect with Daniel on LinkedIn.
Daniel’s written responses have been edited for clarity and organized into our ten standard interview topics.
1. Which product decisions do you own?
I oversee decisions about how teams consume infrastructure services, from multi-tenancy and Kubernetes to provisioning virtual machines. A big part of the job is bringing together the needs of two groups. Infrastructure teams are responsible for delivering a secure, trusted platform. Site reliability engineers (SREs) and application teams expect a cloud-like experience where they can provision what they need on their own.
2. When you’re evaluating whether something is working (or worth building), what signals matter most to you?
One of the clearest signals is when SREs or application teams start looking for alternatives outside their organization. They need enough speed and flexibility to build and run their workloads. When developers start going around internal IT to get that, it’s a strong sign the internal service has become too rigid and needs to be simpler to use.
3. Walk me through a recent product decision you made that involved data. What did the process actually look like?
We were deciding how to simplify placement for self-service VM deployments. Our first assumption was that end users needed more control over where their VMs ran, including which hosts or availability zones they could use.
We collected configuration logs from existing customer environments and followed up with interviews to understand why those configurations existed. Some infrastructure teams needed to put Windows and Linux VMs on separate hosts to comply with licensing requirements. Others needed to route particular VMs to hosts with faster outbound network access. Those rules had to be enforced automatically, without asking the person provisioning a VM to know about them.
That changed the order of the roadmap. We built the ability for infrastructure teams to configure and enforce their placement rules first. That gave IT the guardrails to open up self-service provisioning while we worked on giving end users more placement options.
4. How confident do you generally feel in the data available to you when making product decisions? What tends to increase or reduce that confidence?
I trust a metric when I know what it represents and how it relates to the customer’s definition of success. With a new feature, one of the first things I ask is whether the activity came from a test environment or from production. A spike can reflect people trying the feature in a sandbox. I use that signal to decide which customer conversations to have, so I can find out what people were trying to do and whether they’re ready to use it in production.
5. What’s the most frustrating or time consuming part of getting the insights you need to make a decision?
Getting automated telemetry from production is the hardest part in enterprise infrastructure. Customers run critical workloads in regulated, on-premises, or air-gapped environments. For good security and compliance reasons, they may turn telemetry off or restrict what can be collected. That makes customer relationships and post-deployment check-ins especially important when we need to understand what’s happening in production.
6. How self-serve is data access for product managers in enterprise infrastructure?
The sources I can pursue most directly are qualitative. I can run a structured customer interview, ask design partners for feedback, or work with field teams that see how customers deploy their systems. Sometimes thirty minutes with a platform architect gives me more context for a decision than I could get from a chart.
7. What's the hardest thing about turning data into action rather than just more dashboards or reports?
The hard part is understanding what the data tells you to do first. In the placement example, the configuration logs gave us something concrete to investigate. The interviews explained which rules IT had to enforce before it could open up provisioning. Once we understood that dependency, we could put the provider controls first and give end users more placement options in a later phase.
8. Are there product metrics or definitions that teams regularly interpret differently?
One example is what teams infer from the number of rules in an environment. Core IT might configure only five or six provider rules, while users can create thousands of their own. If you go by volume, it’s easy to conclude that user-level rules matter far more. Those few provider rules can be what lets IT offer the service without taking on significant licensing or financial exposure. You have to know what a rule does before the count can tell you anything about its importance.
9. What’s a surprising or overlooked source of product insight that you think more teams should pay attention to?
The customer of your customer. In enterprise software, PMs often spend most of their time with platform owners and infrastructure teams. Those teams are responsible for architecture, security, and governance. The SREs and application developers who use the service every day have different workflows. When feedback passes through several layers of an organization, some of that context gets lost, and it becomes harder to see where the friction starts.
10. What advice would you give another PM at a startup trying to make better product decisions with data?
Ask your platform-owner champions if you can join an internal onboarding session or a workflow review with an application team. You get to hear directly from the people who use the product every day and ask more specific questions about where they get stuck. The platform owner gets a clearer picture of what their internal users need, so they can make the service work better for them.