Value stream mapping

Bottleneck analysis

Bottleneck analysis is licensed CC BY 4.0. Attribution: Katafacts (katafacts.com).

1 · What it is

What it is

Every value stream has two distinct constraint questions, and they don't always have the same answer: which stage caps throughput (the one with the largest process time — the capacity constraint, in the theory-of-constraints sense), and which stage chokes flow worst (the one with the largest wait time — where deals sit stalled longest). A bottleneck analysis answers both directly from your stage data, the same computation the sales funnel value stream map runs as part of its fuller picture, pulled out on its own for when the question is specifically 'which stage do I fix first,' not a full documented map.

2 · When to use it

When to use it — and when not to

Use it when

You already suspect the pipeline has a problem stage and want it named with the arithmetic behind it, not just a gut feeling confirmed.
You need to decide where to focus limited attention first — one specific stage, not a general 'speed up the funnel' initiative.
You have per-stage process and wait time and want the capacity constraint and the largest wait contributor identified without building a full map.

Not when

You want the overall flow-efficiency percentage, not a specific stage — that's the funnel PCE calculator.
You want waste findings, a Pareto ranking of every stage, or countermeasures — this deliberately only surfaces the two constraint findings, not the fuller picture the sales funnel value stream map builds.
Every stage in your pipeline has roughly the same process and wait time — there's no single bottleneck to name, and forcing one would misdirect the fix.

3 · How to fill it in

How to fill it in

Scope: Just the one-line description at the top. Keep it to one comparable pipeline segment, same discipline as every other VSM-family artifact — a blended segment produces a bottleneck finding that describes neither.
Process steps: List each stage in order with its process time and the wait time after it. This is the only real intake — paste rows straight from a CRM export if you have them.
Lead time · Process time · Flow efficiency: The capacity constraint (largest process time) and the largest wait contributor are identified for you the moment you submit — that's this tool's whole output.
Waste — the eight wastes: Not part of this tool's output. The full sales funnel value stream map runs the complete eight-wastes breakdown; this one only names the two constraint stages.
Where the wait concentrates: Not part of this tool's output as a ranked list — you get the single largest wait contributor directly, not a ranking of every stage. Use the full sales funnel value stream map if you want every stage ranked.
Countermeasures: Not part of this tool's output — there's no observation-notes field feeding a countermeasure narrative here. Once the constraint stage is named, the sales funnel value stream map or a deal A3 is where you'd work through what to do about it.

4 · What good looks like

What good looks like

The example below is the same six-stage outbound funnel as the full sales funnel value stream map, because the constraint findings this tool surfaces are exactly the ones that map computes as part of its fuller analysis. What's different is the path: this page only asked for stage names, process time, and wait time.

Same example, as a downloadable xlsx workbook.

Download .xlsx

Value stream map — current state

Mid-market outbound pipeline — current-state analysis

Jordan Ellis, Sales Manager · 2026-03-16

Team: Priya Anand, RevOps, Sam Cole, Sales Enablement · Sponsor: Renee Okafor, VP Sales

Scope

Scope runs from a lead reaching SQL (sales-qualified) status to closed-won or closed-lost, for outbound-sourced mid-market deals only. Excludes pre-SQL marketing nurture and post-close onboarding — those are separate value streams with their own owners.

Process steps

Step	Process time (hours)	Wait after (hours)	Operators	Defect %	WIP after
Lead qualification	2	24	1	40%	—
Discovery call	3	96	1	35%	—
Solution demo	4	120	1	25%	22
Proposal & pricing	6	168	2	30%	—
Negotiation	5	72	1	15%	—
Contract signed	2	0	1	—	—

Lead time · Process time · Flow efficiency

Lead time

502 hours

Process time

22 hours

Flow efficiency

4.38%

2 + 3 + 4 + 6 + 5 + 2 (process time) + 24 + 96 + 120 + 168 + 72 + 0 (wait time) = 502 hours

22 / 502 × 100 = 4.38%

Capacity constraint

Proposal & pricing

6 hours · 1.20% of lead time

Largest wait contributor

Proposal & pricing

168 hours · 33.47% of lead time

WIP flags

Discovery call — Wait time here is 19.12% of total lead time, above the 15% flag threshold — this is where WIP is piling up.
Solution demo — Wait time here is 23.90% of total lead time, above the 15% flag threshold — this is where WIP is piling up.
Proposal & pricing — Wait time here is 33.47% of total lead time, above the 15% flag threshold — this is where WIP is piling up.

Waste — the eight wastes

Defects82.6 %
Rolled throughput yield across the 5 steps with a recorded defect rate is 17.40% — about 82.60% of units need rework or scrap somewhere in the stream, so first-pass output is lower than any single step suggests.
(1 - 40/100) × (1 - 35/100) × (1 - 25/100) × (1 - 30/100) × (1 - 15/100) × 100 = 17.40% yield — from Lead qualification (40%), Discovery call (35%), Solution demo (25%), Proposal & pricing (30%), Negotiation (15%); defects = 100 − 17.40 = 82.60%
OverproductionQualitative
Not quantified from step timings. On the next walk, look for producing more than the next process pulls, or running ahead of takt — batches built early, reports nobody reads.
Waiting480 hours
Work sits idle for 480 hours in total across this value stream, queued after 5 of 6 steps.
24 + 96 + 120 + 168 + 72 + 0 = 480 hours
Non-utilized talentQualitative
Not quantified from step timings. Look for operators' process knowledge or improvement ideas going unused — people closest to the work with no route to change it.
TransportationQualitative
Not quantified from step timings. Look for unnecessary movement of material or information between steps — handoffs, re-keying, files shuttled between systems.
Inventory22 units
22 units of work in progress sit in queues after 1 of 6 steps — every one of them is finished effort not yet paid for.
Solution demo (22) = 22 units
MotionQualitative
Not quantified from step timings. Look for unnecessary operator movement within a step — reaching, searching, walking to fetch what the work needs.
Extra-processingQualitative
Not quantified from step timings. Look for doing more to the product or information than the customer actually values — extra approvals, duplicate checks, unused detail.

Where the wait concentrates

Ranked by cost

Proposal & pricing168
Solution demo120
Discovery call96
Negotiation72
Lead qualification24
Contract signed0

Countermeasures

Pre-approve standard pricing tiers for deals under a set contract-value threshold so they skip the two-approver review entirely; keep the full review only for genuinely custom terms.
Linked finding: Proposal & pricing — largest wait contributor at 33.47% of lead time, and the capacity constraint at 6 hours of active work per deal
Impact: high · Effort: medium · Owner: Priya Anand · Due: 2026-04-06
Cap the number of deals allowed to sit between Solution demo and Proposal at any time, pulling the next demo only when a proposal slot frees up, instead of running demos as fast as reps can book them.
Linked finding: Solution demo — 22 units of WIP queued, wait time 23.90% of lead time
Impact: medium · Effort: medium · Owner: Jordan Ellis · Due: 2026-04-13
Tighten the SQL handoff criteria with marketing so fewer poorly-fit leads enter the pipeline only to drop out at qualification — add one firmographic checkpoint before a lead is marked SQL.
Linked finding: Defects — 17.40% rolled throughput yield across 5 stages; Lead qualification is the single largest contributor at 40% drop-off
Impact: medium · Effort: low · Owner: Sam Cole · Due: 2026-04-06

5 · Common mistakes

Common mistakes

Treating the capacity constraint and the largest wait contributor as the same stage without checking.
They frequently aren't the same stage — one caps how much work can move through per unit time, the other is where already-started work sits stalled longest. Fixing the wrong one for the problem you actually have wastes the attention this tool is meant to focus.
Naming a bottleneck from a pipeline segment that blends genuinely different deal types.
Same scope discipline as every other VSM-family artifact — a blended segment's 'bottleneck' is an artifact of the blend, not a real constraint in either underlying segment.
Acting on the bottleneck finding without re-running the analysis after the fix.
Removing one constraint reliably surfaces the next one — a bottleneck analysis is a snapshot, not a one-time diagnosis, the same theory-of-constraints principle that applies on a shop floor.

6 · What it connects to

What it connects to

upstream

downstream

Sales funnel value stream map
Once the constraint stage is named, this is where you'd build the fuller picture — waste findings, a full Pareto ranking, and countermeasures tied to the finding.
Funnel PCE calculator
If the question shifts from 'which stage' to 'how bad is it overall,' this is the more direct next step.
Deal A3
Once a specific constraint stage is named, a deal A3 is where you'd run root cause analysis on why that stage specifically is capping throughput or choking flow.

7 · Where AI helps

Where AI helps

Judgement — stays yours

Deciding whether the capacity constraint or the largest wait contributor is the more urgent one to act on first
Judging whether a named bottleneck is structural or a one-off caused by unrelated circumstances

Analysis — AI helps

Nothing here — identifying the constraint stages is a deterministic calculation, not an analysis judgement call

Drudgery — automated

Identifying the capacity constraint and the largest wait contributor from stage data, with the arithmetic shown
Exporting to xlsx in the house format

9 · Rate this kata