Advanced Lean Training KPI Targets and Baselines: What to Track First

If you're doing advanced Lean work, you're past the "track everything" phase - and you're past vanity metrics. The fastest way to earn leadership trust is to:

1.       Pick a small set of KPIs that actually move the business,

2.       Build a baseline you can defend, and

3.       Set targets that are ambitious and credible.

This page gives you a practical sequence for what to track first, what "baseline" really means, and how to set KPI targets without guessing.

To start advanced Lean work, track a small "starter set" tied to flow and customer impact: end-to-end lead time, on-time delivery (or schedule attainment), first-pass yield/defect rate, and work-in-process (WIP). Build a baseline using 4-8 weeks of stable data (or at least 30-50 observations), with clear operational definitions and stratification by product family, shift, team, or channel. Set targets by linking to business needs (cash, capacity, service level), checking current capability (run charts/capability), and focusing improvements at the constraint step. Use realistic ranges (e.g., 15-30% lead-time reduction in the first wave, then stretch), review weekly, and adjust targets only when scope changes - not when results get uncomfortable.

The KPI logic for advanced Lean (what matters first)

Advanced Lean KPIs should answer four questions in this order:

1.       Is value flowing faster? (time + WIP)

2.       Is the customer getting what they need, when they need it? (delivery + quality)

3.       Is the system stable and capable? (variation + rework)

4.       Is the business winning? (cost + cash + capacity)

When teams skip this sequence, they often start with cost ("labor efficiency") and accidentally create local optimization: people look busy while lead time and customer pain stay the same.

What to track first: the "Lean Starter Set" (pick 5-7)

The starter set (recommended)

End-to-end lead time

·         Definition: request-to-delivery (or order-to-cash), end to end

·         Why first: flow is the point of Lean; lead time exposes waste immediately

WIP (work in process / queue size)

·         Definition: items waiting or in progress at key steps

·         Why first: WIP drives lead time and hides problems

On-time delivery / service level

·         Definition: % delivered by promised date/time (define "promised" carefully)

·         Why first: direct customer impact; aligns ops + commercial teams

First-pass yield (FPY) or defect rate

·         Definition: % done right first time (or defects per unit/ppm)

·         Why first: rework destroys flow and capacity

Rework loops / escalation rate (choose one)

·         Definition: % requiring rework, returns, corrections, approvals, escalations

·         Why first: tells you where variation is leaking into the system

Constraint throughput (advanced Lean "secret weapon")

·         Definition: output rate at the bottleneck step (units/hour, cases/day, tickets/day)

·         Why first: improving non-constraints rarely moves system results

COPQ (cost of poor quality) (if finance-ready)

·         Definition: scrap + rework labor + warranty/returns + expedited shipping + concessions

·         Why first: translates quality and delay into money leadership understandsOptional add-ons (only if they support your first wave)

·         Changeover time (if you're fighting batching)

·         OEE (only when equipment is truly the constraint and data is clean)

·         Inventory turns / days of inventory (for cash + flow)

·         Customer complaints / NPS drivers (when service perception is the constraint)

KPI selection rules (so you don't overload the dashboard)

Use these rules to keep KPIs tight and useful:

·         One KPI per outcome category: Time, Delivery, Quality, Cost, Capacity

·         1-2 leading indicators max (e.g., WIP, constraint uptime, queue aging)

·         A KPI must be actionable weekly: if you can't act on it in 7 days, it's not a steering metric

·         Define ownership: every KPI has a single accountable owner (not "the team")

·         Avoid composite metrics early: start simple, then mature later

How to build a baseline you can defend (fast + clean)

A baseline isn't "what last month looked like." It's a measured reference point with definitions, scope, and context.Step-by-step baseline method

1.       Lock the scope

·         Which value stream / product family / service line?

·         Start and end points (where the clock starts/ends)

2.       Write operational definitions

·         Lead time: calendar time or business hours?

·         OTD: promised date vs requested date vs scheduled date?

·         Defect: what counts and who decides?

3.       Choose the baseline window

·         Ideal: 4-8 weeks of stable operations

·         If operations are volatile: use the most recent stable window and document what "stable" means

4.       Get enough observations

·         Aim for 30-50+ data points per KPI (more is better)

·         If volume is low, use weekly aggregates but keep raw detail stored

5.       Stratify early

·         Baseline averages can lie. Break results by what actually changes performance:

·         product/service type, channel, shift/team, supplier, region, customer segment

6.       Use run charts before you use averages

·         Plot the KPI over time

·         Look for shifts, trends, seasonality, and special causes

·         Document known events (downtime, promotions, policy changes)

7.       Capture baseline as a "baseline card"

·         For each KPI, store:

·         definition, formula, source system, data owner, frequency

·         baseline window dates

·         baseline value + spread (min/max or percentile)

·         known exceptions

Deliverable you want: a one-page baseline card per KPI that makes debate unnecessary.

Setting targets (credible, not wishful)

Targets should come from business need + capability + constraint focus, not hope.Target-setting ladder (use in this order)

1.       Business need: What must improve for revenue, cash, capacity, or customer promise?

2.       Voice of the customer: What service level or defect level is required?

3.       Current capability: How capable is the process today (variation, stability, rework)?

4.       Constraint leverage: If we improve only one step, which step moves the whole system?

5.       Feasibility & timebox: What can the team realistically change in 30/60/90 days?Practical target patterns (typical starting ranges)

These are not guarantees - use them as planning bands for the first improvement wave:

·         Lead time: 15-30% reduction in the first wave (often more once batching is attacked)

·         WIP: 10-25% reduction (paired with pull limits)

·         OTD/service level: +5 to +15 percentage points (if baseline is unstable)

·         FPY/defects: 20-50% reduction in defect rate when top rework drivers are known

·         Constraint throughput: +5-15% without major capex when waste at the bottleneck is removed

·         COPQ: 10-30% reduction when rework and warranty are measurable and attackedDon't set targets without these two guardrails

·         Guardrail 1: quality must not decline while chasing speed

·         Guardrail 2: safety/compliance must not be traded for output

KPI review cadence (how advanced teams run it)

Weekly (30 minutes): "Flow & stability huddle"

·         Lead time trend, WIP trend, OTD trend, FPY/defects trend

·         Top 1-3 causes of misses + countermeasures (not explanations)

Monthly (60-90 minutes): "Business impact review"

·         COPQ, capacity gained, backlog/aging, customer escalations

·         Validate whether the constraint has moved

Quarterly: "Target refresh"

·         Reconfirm scope, adjust targets only for scope changes or step-change learning

·         Add/remove KPIs only if a KPI is no longer driving decisions

Common mistakes (and what to do instead)

Mistake: Starting with 20 KPIs

Do instead: Starter set + one constraint measure

Mistake: Using only averages

Do instead: Run charts + stratification

Mistake: Targets based on "industry benchmarks" with no context

Do instead: Capability + business need + constraint leverage

Mistake: Measuring OEE everywhere

Do instead: Measure it only where equipment is truly the constraint and data integrity is high

Mistake: Changing definitions mid-project

Do instead: Freeze KPI definitions for the project phase; version them if you must change

Quick-start checklist (copy/paste)

In the next 48 hours:

·         Pick 1 value stream (start/end points written)

·         Choose 5-7 KPIs from the starter set

·         Write operational definitions (one paragraph each)

·         Pull 4-8 weeks of data (or stable window)

·         Create run charts for Lead Time, WIP, OTD, FPY/Defects

·         Identify the constraint step and add constraint throughput

·         Set 30/60/90-day targets using business need + capability

·         Schedule weekly KPI huddle