If you were building a PMM team from the ground up, e.g. in a startup, which responsibilities would you want to own vs what would you expect the PM team to manage?

The approach will typically vary depending if the startup is pre or post PMF. Pre-PMF you need to be pretty open to a broad set of workstreams falling under PMM. It will also depend if you have a wider marketing team or not. PMM can often help define what other marketing roles are required. When hiring a PMM team pre-PMF it's first worth thinking about how you're going to structure the team; will it be around product features or workstreams e.g. Competive intel, lifecycle, messaging, etc. or geography. For post-PMF you'll usually want to align closer to the product team.

I believe PMM should own how we articulate the value of our products or solutions to the market. That would include enabling internal teams with clear messaging and positioning and owning how we take the product to market. My expectation is that PM would own the technical documentation of the product and the enabling of internal resources on how the product works technically. 

Obviously this is pretty high level and there are a lot of resopnsibilities underneath this that would require ownership. I'd recommend you work closely with your product counterpart(s) to document ownership and align on what PMM owns vs. PM. You can then use that document as you bring on new PMs and PMMs into your organization. 

One of the things I have noticed during my time at various companies is that PMs may not get the opportunity to interact and engage with customers as much as they like. Too many times, they are very development focused, and roadmap is led by HiPPOs. If this is the case in your orginazation, I would make customer engagement and understanding the primary role for a PMM. You can then take that insight and knowledge and feed it back to the Product team, hopefully building a strong relationship. 

That being said, I agree with Judy in that every company is different so it is important to understand where the gaps currently are, and how PMMs can add value. It is easy to take on work that Product feels that they are responsible for, so rather then having that battle later on, it may be best to start with filling the gaps and taking on further responsiblity as you build trust. 

This is something I have experience doing since I was one of the first monetization product marketers at Instagram and the first product marketer at Slack. I'll say that it was different at both companies, so it can vary. It really matters what your company's priorities are and what each individual is focused on and where their strengths lie. 

PMMs do not support PMs 1:1 in small companies (wouldn't that be nice!), so it's best to be strategic about where a PMM can help fill in gaps early on. To me, this is really at three phases: 1) defining the MVP (through an understanding of customer needs, competitive overview, and market sizing), 2) determining if an alpha/beta is needed and managing those, and 3) launch and post-launch support. 

I also believe strongly that a PMM should help define success metrics. If it's too early to set concrete KPIs, it's important to start building the foundation for what you'll measure and why.

Philosophically, I think it comes down to the "what" (both), the "why" (PMM) and the "how" (PM). The interesting part is the ways in which the "how" and the "why" inform each's take on the "what" when collaborating. 

When it comes to that "what" I think, in general, PM has the domain expertise on functional aspects of it, while PMM defines it from a business need perspective.