Meeting Notes/2015-12-22 Maintenance Fraction

Wes, Kristen, Joel

Joel: TPG has Q3 goals to help teams measure maintenance fraction, but that only addresses a portion of what I understand Wes to be asking. I understand Wes to be asking, "TPG, can you help Team X answer the question, 'can you do feature Y next quarter?'". To really answer that, though, requires teams to know a lot more than their maintenance fraction (see below for enumerated list).

What role does TPG play in helping the Foundation accomplish this goal?

Wes: I was assuming we would work with only a few teams for this quarter until we have a general solution.

Joel: Yes, we definitely need more time with only a few teams to hammer out details.

How do we get it bigger after Q3? Wes: Open to a lot of things, trying to get a path forward. Kevin said there was a request in the Strategy discussion to ID core and non-core work.

Definition of Core vs Non-Core

Wes: Core is "what are the things you have to do to keep the features in production running?"

Joel: definition varies depending on the context and reason for the question.

Wes: Stratgey team is asking in context of annual planning - things we have to do vs things we can choose to do.

Two contexts: Strategy and Annual Planning - are they the same or slightly different?

Kristen: clarifying: TPG has role to lead this but not to answer on our own, because the answers are in many peoples' heads, not only or primarily TPG. Wes, are you saying that Strategy depends on solving Core/Non-Core, or the other way around?

Wes: Strategy depends on defining Core/Non-Core. We want an FDC-format annual plan by April. Will get a signal from Finance in 2nd week of January. Finance will asks teams to identify core vs non-core work. In February, teams will answer how their work addresses the strategy. Community feedback in Feb/March. So we need a Core/Non-Core definition ASAP. End of January? Kristen: End of January feels very ambitious; maybe we can leverage strategy conversation.

Answering Wes's question involves several new capacities WMF-Eng-Wide. Is TPG working on helping anyone with any of these now?

1. teams have backlogs
2. backlogs are prioritized
3. backlogs are estimated/quantified
4. new work is identified
5. new work is estimated
6. new work is prioritized relative to current backlogs

Wes: this is about planning more than about strategy. "Strategy" is a word that distracts people from the importance of planning.

TPG's plans

Take Maintenance Fraction out of pilot phase and roll out to WMF Eng ASAP

Continue piloting backlog definition/estimation/forecasting with ~3 teams, hoping to work out the kinks and document best practices in Q3, and start rolling out to more of WMF Eng when possible but probably not in Q3.

More SPDPP stuff, like milestones and checklists and implementing high-level phase/process/thinking team by team: continue small piloting to make sure that it is consistent with the other changes, but don't push it out in the near future (not in Q3, maybe not in Q4) because a) TPG is limited in capacity, b) WMF capacity to change is limited, and we don't want to overwhelm ourselves.

Kevin's comments on these notes, not on the meeting itself

My guess is that the link between "defining core" and strategy is that everything the WMF does should either be core (keeping things running), or strategic (supporting one of our strategic goals). It seems possible that work could be both, or it could be neither (in which case it should beware of the looming budget axe).

As far as I know, the strategy process doesn't rely on having a core/non-core definition. But it seems that annual planning does. It was a strategy team action item for Lila to define "core" because she keeps using the term.

In earlier discussions with teams, it seemed clear that they would have a hard time agreeing on a definition of "maintenance". I suspect "core" might be easier, but this might be a case where Wes should propose a candidate definition, and based on discussion and input, to eventually declare the One True Definition. Achieving that by the end of January might be possible.

I don't think TPG should move forward with measuring maintenance fraction until we have a definition in place. Once we have a definition, we should first pilot that.