Week 21 Day 2: Horizon 1 -- What We Are Doing Now

Horizon 1 is the only horizon that should occupy the team's daily attention. It includes the work that is actively being executed, the blockers that need to be cleared, and the decisions that need to be made this week. Everything in Horizon 1 has a specific owner, a specific deliverable, and a specific timeline. If a piece of work does not have all three, it is not ready for Horizon 1.

The most common mistake with Horizon 1 is overloading it. Leaders pack everything they want done 'soon' into the current horizon, which defeats the purpose of the model. Here is the constraint I use: Horizon 1 should never contain more than three major workstreams at any time. Not three tasks -- three workstreams. Each workstream may contain dozens of tasks, but the team should be able to name the three things they are focused on right now without checking a document. When I audit teams with execution problems, the first thing I check is the Horizon 1 load. Almost always, it contains six to ten workstreams that the leader considers 'current.' The team is context-switching between all of them, completing none at the pace the leader expects. The fix is simple and painful: move everything beyond three to Horizon 2. Tell the team explicitly: 'These are our three priorities. Everything else waits.' Watch the team's velocity increase within a week -- not because they are working harder, but because they have stopped paying the context-switching tax.

The three-workstream limit is supported by research on 'work in progress limits' from Lean manufacturing (Ohno, 1988) and their application to knowledge work through the Kanban method (Anderson, 2010). Anderson's research across 131 software teams found that teams with explicit WIP limits completed work 30-50% faster than teams without limits, despite starting fewer items simultaneously. The mechanism is Little's Law from queueing theory (Little, 1961): cycle time equals work in progress divided by throughput. Reducing WIP directly reduces cycle time without requiring any increase in effort or capability. The context-switching tax is quantified by research from the American Psychological Association (Rubinstein, Meyer, and Evans, 2001), which found that task-switching costs range from 20-80% of productive time depending on task complexity, with complex cognitive tasks at the high end of that range. Weinberg (1992) in 'Quality Software Management' estimates that each additional project a knowledge worker takes on reduces their effective productivity per project by roughly 20%, meaning that five simultaneous projects yield only about 5% effective time on each. The instruction to move overflow to Horizon 2 rather than simply deprioritizing reflects what Goldratt (1984) calls 'managing the constraint' in Theory of Constraints -- the insight that system throughput is determined by the bottleneck, and adding more work to a constrained system reduces total output.