Level 1: The Excel Foundation
Level 1: The Excel Foundation describes what workforce management looks like at the first level of the WFM Labs Maturity Model™ — and the minimum path out. Level 1 is the manual baseline: scheduling lives in spreadsheets, forecasting leans on recent history and instinct, and real-time management means supervisors improvising when problems surface. The defining observation is that Level 1 operations are manual, not broken: competent people compensating for missing instruments. The gap is rarely technology — the ACD already exports interval data and a spreadsheet can run queueing math — it is instrumentation and discipline. This page covers operating at the level and advancing from it; for Level 1's place among the five levels, see the model page and The Maturity Model as a Transformation Framework. It is part of the Adaptive Concepts series.
The operating reality
Four patterns recur wherever the level appears, typically in operations up to roughly one hundred agents:
- Activities, not roles. Forecasting, scheduling, and real-time monitoring are chores shared among supervisors and a spreadsheet-fluent lead, not named roles with accountabilities and handoffs. When service slips, fixes arrive as one-off favors.
- Artifacts, not loops. Plans exist as tabs and emails; nothing connects what was expected to what arrived, so the operation cannot tell normal pattern drift from real disruption until backlogs make the difference obvious.
- Definitions vary. Service level appears on wallboards, but abandons are treated inconsistently, occupancy is described as "busy" rather than governed against a band, and shrinkage is a remembered percentage rather than a policy.
- Averages hide the day. A green daily service level can conceal red intervals customers actually lived; without 15–30-minute visibility, two very different days end with the same number.
The death spiral
Level 1's characteristic failure is a loop. A demand spike or handle-time jump triggers a crisis response — all hands, breaks slid, coaching and training canceled. The canceled development degrades quality; degraded quality extends handle times; extended handle times create the next crisis. Nobody is negligent, and the loop tightens anyway. The load-bearing insight is that canceled training is not a saving: it is borrowing capacity from next month, and the debt compounds. Operations that protect development with an explicit guardrail — cancel only past defined thresholds, and re-book within 48 hours — are not indulging their people; they are servicing the debt before the interest accrues.
The instruments that stabilize it
Stabilizing Level 1 requires publishing a small set of instruments and then letting them drive behavior:
- An interval service-level target (for example, 80% answered in 30 seconds, measured every 15–30 minutes) with a written abandon treatment — excluding short abandons under a documented threshold is the common middle ground.
- An occupancy ceiling (commonly around 88–90%): sustained occupancy above it predicts fatigue and quality drift. There is deliberately no floor at this scale — in small queues, low occupancy is what hitting the service level costs, not slack; an operation carrying 8 erlangs needs eleven agents to make 80/30 and will run near 73% occupancy doing it. Treating that as waste, and cutting the heads, is how small operations talk themselves out of their own service level (see The Occupancy Trap).
- A shrinkage policy naming components — breaks, meetings, training, PTO, unplanned absence — with target rates, applied to required staff to produce scheduled staff.
- A queueing baseline. Erlang-C translates the interval promise into required bodies. In a worked case — 200 calls in a 30-minute interval at 415 seconds handle time, targeting 80/30 — 52 agents meets the target (at 88.7% occupancy, brushing the ceiling), 51 delivers roughly 73%, and 50 roughly 65%: the "Power of One," in which the queue's non-linearity turns one agent into a seven-point service swing.[1] Dividing the 52 by (1 − shrinkage) then yields scheduled heads — at 30% shrinkage, 74.3, rounded up to 75. Koole's treatment of the occupancy–service trade-off is the standard deeper reference for this arithmetic.[2]
The instruments matter because they change the register of conversation: without them, improvement arguments sound like opinion; with them, a miss becomes a diagnosable lever — volume, handle time, or staffing — rather than an autopsy.
The manual rhythm
Level 1 does not need a platform to run repeatably; it needs a rhythm with written handoffs — the cadence Cleveland's planning process formalizes at larger scale:[3]
- Weekly — a one-page brief: the interval forecast, its assumptions, known events, and two named uncertainties.
- Daily — a roster built to the brief, breaks placed against peaks, development blocked with its guardrail.
- Intraday — a short ladder of pre-agreed micro-moves applied in order before anything is canceled: reclaim one back-office task, nudge two breaks by five minutes, unlock a small flex pool.
- End of day — a two-line variance log: which intervals deviated and which lever moved, feeding the next forecast.
The rhythm's structure is a loop of handoffs: forecasting hands assumptions to scheduling, scheduling hands constraints and protections to real-time, and real-time hands variance notes back to forecasting. Its governance fits in one line: if it changed the plan, it goes in the variance log; if it keeps happening, it becomes a standard procedure; if a procedure is ignored twice, fix the procedure or the goal. Ready-to-copy one-page procedures for the weekly forecast, roster build, and intraday control are at Level 1 Process Templates.
The four promises
Interpersonal design at Level 1 is not a soft add-on; it is the mechanism by which the plan becomes a day people can sustain, and it anticipates the Service-Profit Chain's causal logic with no new software:
- Predictability — schedules posted on time; breaks moved in small pre-agreed nudges with a stated reason, not slid wholesale.
- Fairness — published rules for PTO, swaps, and overtime; exceptions logged once and applied the same way next time.
- Voice — preferences captured and visibly reflected in the roster, without promising what cannot be delivered.
- Development — coaching protected by guardrail, and re-booked within 48 hours when variance genuinely forces a cancellation.
Adherence under this design is a support mechanism — the way the team keeps its promises to customers and to each other — rather than surveillance; context precedes consequence when a window is missed.
Signals of the crossing to Level 2
These are readiness signals — evidence the operation can carry a platform — as distinct from the pressure signals (volume growth, burnout, cost) that usually prompt the purchase conversation; pressure without readiness buys shelfware. The shift is felt before any software contract is signed: definitions are posted and used in daily conversation; the weekly-daily-intraday cadence has survived six to eight weeks of peaks without collapsing; and the pain has concentrated — manual schedule edits, adherence visibility, or multi-skill coordination is now the named bottleneck rather than general confusion. At that point the platform case is one page, not a feature list: today's failure mode ("manual edits consume six hours a week; late changes create two red intervals most days"), tomorrow's capability, measurable effect — anchored to the operation's own intervals, shrinkage policy, and occupancy band so the savings are not imaginary.
Maturity Model Position
The essential Level 1 fact within the model's progression: its exit is a set of agreements rather than a purchase, and the habits it installs — posted definitions, protected development, rules over panic — are the inputs every later level assumes (see The Maturity Model as a Transformation Framework for why a capability built without its inputs is being skipped-to, not built early).
See Also
- Adaptive Concepts — the series this page belongs to
- Level 1 Process Templates — the copy-ready procedure pack
- Erlang-C — the staffing baseline and its assumptions
- The Occupancy Trap — why occupancy is a consequence to govern, not a target to chase
- Variance and Volatility in Service Operations — the uncertainty taxonomy behind the rhythm
