Unmanaged switching

Description

  • Switching tasks burns time in ramp-up. Cue unrealistic expectations and disappointment, especially if multiple stakeholders are asking for switches but aren't aware of each other

Disproof

  • An established system for deciding any change to in-flight work, and a way of communicating it

Consequences

  • If switching happens faster than the actual time cost of switching (invariably at least half a week), you get this lovely thing we call ‘thrashing’, where literally nothing gets done
  • People assuming that a one week detour means a one week delay, undercompensate their own scheduling. Dependencies get fouled up, marketing pushes get misaligned, a whole bunch of nasty ensues
  • A one-off won’t kill you, but repeats will torch internal and eventually customer trust
  • As per usual, either move slow and die, or burn everyone out trying to maintain speed

Causes

  • People who, in 202X, still make it part of their personality to not understand how engineering works whilst working at a product company and somehow still seem have jobs
  • When product teams have some kind of prioritisation cadence with individual departments, where someone weak-willed gets cornered by someone on a warpath
  • If you’re in the business of making switches that you know to be wrong but aren’t skillfully managing and preventing, it’s a psychological safety and skill issue. Effectively, you’re prioritising the shoutiest, but in-flight

Approaches

  • Changing direction before you hit ‘done’ can make sense if the priorities of the business have drastically changed
    • People should only be mad if this happens for a bad/avoidable reason
    • But if the original target is no longer viable, don’t waste more time on it
  • If the diversion is only meant to be temporary, be extremely quantified about the cost of stopping, contextualising elsewhere, and then re-contextualising on the existing problem. This is often forgotten by anyone abstracted away from building (investors, non-product teams, sufficiently disconnected leadership)
  • Train people in charge of defending prioritization to not buckle like wet cardboard, and also how to not piss people off when saying no