Home / Glossary / Conversion Rate
Metrics · the funnel-shape diagnostic · deep dive

Conversion rate. The stage with the lowest conversion is the only stage that matters.

Conversion rate is the percentage of records that progress from one funnel stage to the next — and the single most useful diagnostic for separating "we have a pipeline problem" from "we have a closing problem." Each stage has its own conversion rate (lead → MQL, MQL → SQL, SQL → opportunity, opportunity → close), and the stage with the lowest conversion is the bottleneck — fixing that single stage moves total funnel output more than improving every other stage marginally. The aggregate funnel math hides this entirely: a team can have a healthy-looking overall conversion of 1.5% (lead to close) while sitting on a catastrophic 8% MQL-to-SQL conversion that's the actual operational problem. This essay covers the formula, the funnel-shape diagnostic that reveals four common funnel patterns (top-heavy, middle-narrow, bottom-narrow, even), the source-level conversion analysis that reroutes marketing budget toward the highest-converting channels, and the bottleneck-first operational discipline that improves total throughput.

Category: Metrics · Read time: 9 min · Updated: 2026-05-24 · CONV-1.0
TL;DR
Conversion rate is the percentage of records that progress from one funnel stage to the next. Each stage has its own conversion rate; the relevant diagnostic isn't the overall funnel conversion (interesting only as a headline) but the stage-by-stage breakdown that identifies bottlenecks. The single biggest insight conversion-rate analysis produces: the stage with the lowest conversion is the only stage that matters operationally. Fixing a 12% conversion to 18% adds more revenue than improving four other stages by 2 points each, because the bottleneck stage gates everything downstream. Common funnel shapes and their diagnoses: top-heavy (lots of leads, few opps — top-of-funnel quality problem), middle-narrow (good leads, weak qualification — discovery/sales-readiness issue), bottom-narrow (qualified pipeline, bad close — late-stage closing issue), even (rare — usually means you're not measuring stages cleanly enough to see). Source-level conversion reveals where marketing budget should reroute: a channel producing leads at 0.3% lead-to-close conversion is worse than a channel producing 30% of the leads at 4% conversion, even if the higher-volume channel "produces more leads." The bottleneck-first discipline is iterative: identify the worst-converting stage, invest 80% of improvement effort there, measure the lift, find the new bottleneck (usually the next-worst stage), repeat. Teams that operate this way improve total throughput 20-40% per quarter on stable pipeline volume.

01What conversion rate is

Conversion rate measures the percentage of records that move from one stage in a sales/marketing funnel to the next. Each transition has its own conversion rate, and the funnel as a whole has multiple stage-to-stage conversions you can compute and analyze independently.

The standard B2B SaaS funnel has roughly 5-6 stages with their own conversion rates:

  • Lead → MQL (marketing-qualified lead): does the lead meet the basic criteria to be worth marketing attention?
  • MQL → SQL (sales-qualified lead): does the lead meet the criteria to be worth SDR/AE time?
  • SQL → Opportunity: does the SDR confirm the prospect as a real opportunity worth tracking in pipeline?
  • Opportunity → Proposal: does the deal reach the formal proposal stage?
  • Proposal → Closed-Won: does the proposal convert to revenue? (Equivalent to win rate.)

Different teams use different stage names; the principle is the same. The aggregate overall conversion = Lead-to-Closed-Won, computed as the product of all stage-to-stage conversion rates. The aggregate is almost always small (1-3% is typical for B2B SaaS); the individual stage rates reveal where the leakage actually happens.

The reframe
Aggregate funnel conversion is almost meaningless; stage-by-stage conversion is everything. A team reporting "1.8% lead-to-close conversion" can have wildly different underlying structures — a 30% MQL conversion combined with a 6% SQL conversion is a very different operational picture than a 10% MQL combined with 18% SQL. Same headline, completely different fix.

02The stage-by-stage funnel

What a real B2B SaaS funnel looks like with each stage's conversion rate exposed:

Sample B2B SaaS funnel · monthly · 10,000 leads
Leads
10,000
10,000
MQLs
3,000
3,000
30%
SQLs
600
600
20%
Opportunities
240
240
40%
Proposals
144
144
60%
Closed-Won
36
36
25%

The aggregate funnel conversion is 0.36% (36 wins ÷ 10,000 leads). The stage-by-stage view reveals: the MQL → SQL stage at 20% is the bottleneck — it's not the worst-looking number (the 25% close rate looks worse out of context), but improving the 20% MQL conversion by 5 points (to 25%) flows downstream and produces 9 additional wins per month (a 25% lift on the 36 baseline). Improving the close rate by the same 5 points (25% → 30%) only adds 7 wins. The bottleneck stage gates everything downstream; improvements there compound more than improvements elsewhere.

The operational rule: compute each stage's conversion independently; identify the lowest; invest there first. Most teams chase the most-visible problem (often the close rate) and ignore the actual bottleneck (often earlier in the funnel where lower-level data isn't naturally visible).

03Four funnel shapes + diagnoses

Four characteristic funnel shapes each indicate a different operational diagnosis:

Shape 1
Top-heavy
Lead → MQL: 8%
MQL → SQL: 35%
SQL → Close: 28%
Lots of leads, few become MQLs. The top-of-funnel volume is fine but the lead quality is weak. Marketing is generating top-of-funnel that doesn't meet the basic ICP threshold.
Better lead targeting · refine ad audiences · tighten ICP rubric · re-evaluate freemium funnel quality
Shape 2
Middle-narrow
Lead → MQL: 30%
MQL → SQL: 12%
SQL → Close: 28%
Good leads, weak qualification. Marketing is sending qualified-looking leads but SDR qualification is dropping most of them. Usually a discovery or qualification-skill issue rather than a lead-quality one.
SDR training · qualification framework refresh · MQL-criteria audit · review SDR-to-AE handoff quality
Shape 3
Bottom-narrow
Lead → MQL: 28%
MQL → SQL: 32%
SQL → Close: 12%
Qualified pipeline, bad close. The funnel feeds well-qualified opportunities but they don't convert. Late-stage closing problem — usually indecision (JOLT), value-selling weakness, or competitive losses.
JOLT-style late-stage training · value-selling math · competitive intelligence · win-loss interviews
Shape 4
Even (rare)
Lead → MQL: 25%
MQL → SQL: 25%
SQL → Close: 25%
Even conversion across stages. Rare in real funnels — usually means your stage definitions are too coarse to reveal the actual bottlenecks. Different stages of a B2B funnel structurally have different conversion rates.
Add more granular stages · separate qualification handoff stages · split out late-stage proposal vs negotiation

Each shape calls for a fundamentally different intervention. Treating all shapes with the same playbook — usually "improve close rate" — fixes the wrong thing for three of the four shapes.

04Source-level conversion

Aggregating conversion across lead sources hides where marketing budget should actually flow. A source-level conversion table for a typical B2B SaaS:

Source
Lead → MQL
MQL → SQL
SQL → Close
End-to-end
Inbound (organic)
45%
40%
35%
6.3%
Referral / partner
55%
45%
40%
9.9%
Outbound (signal-anchored)
25%
30%
25%
1.9%
Paid search
20%
22%
20%
0.9%
Outbound (cold)
12%
15%
15%
0.27%
Content download / gated
8%
12%
22%
0.21%

The numbers reveal what aggregate conversion hides. Referral converts 9.9% end-to-end; cold outbound converts 0.27% — a 36× difference. A marketing team allocating budget proportionally to volume (which is the default) would over-fund cold outbound and under-fund referral programs.

The right discipline: compute end-to-end conversion by source, then allocate budget by expected revenue per dollar invested, not by lead volume. A source producing 100 leads at 0.27% closes 0.27 deals; a source producing 10 leads at 9.9% closes 0.99 deals — 3.7× more revenue from 1/10 the volume.

Watch for
The "leads volume = success" trap. Marketing teams measured on lead volume optimize for the channels that produce the most leads, regardless of conversion. The result: budget concentrates in low-converting channels (content downloads, cold outbound) while high-converting channels (referral, signal-anchored outbound) starve. Switch the metric from leads-generated to revenue-per-marketing-dollar to fix this.

05The bottleneck-first discipline

The 6-step operational discipline for improving total funnel throughput:

  1. Compute stage-by-stage conversion every month. Each stage as an independent measurement. Track the trend (3-month rolling average) for each stage; sudden changes flag operational issues earlier than aggregate-only tracking.
  2. Identify the single worst-converting stage. Not the lowest-looking percentage (the close rate is structurally lower than earlier stages) but the stage that's furthest below its industry benchmark or your own historical norm.
  3. Invest 80% of improvement effort on that stage. Most teams spread improvement effort across 4-5 stages. The bottleneck stage gates everything downstream; concentrating effort there produces 3-5× the throughput lift of distributed effort.
  4. Measure the lift weekly. A bottleneck-targeting program should show stage-level conversion improvement within 60-90 days. If 90 days pass with no movement, the diagnosis was wrong — re-identify the bottleneck.
  5. Find the new bottleneck after the lift. Improving one stage often shifts the bottleneck to the next-weakest stage. The discipline is iterative: identify → invest → measure → find next bottleneck → repeat.
  6. Run source-level conversion analysis quarterly. The bottleneck analysis is for stage-by-stage funnel structure; the source-level analysis is for marketing budget allocation. Both are required for full optimization.
  7. Tie marketing comp to revenue-per-dollar, not lead volume. The metric shapes the behavior. Compensation tied to lead volume optimizes for low-converting channels; compensation tied to revenue-per-marketing-dollar optimizes for the channels that actually produce revenue.

06Common mistakes

Mistake 1
Reporting only aggregate conversion rate. Aggregate is a headline; stage-by-stage is operational. The bottleneck lives in one specific stage; reporting only the aggregate makes it invisible.
Mistake 2
Investing improvement effort across all stages equally. The bottleneck-first principle says 80% of effort should go to the worst-converting stage. Distributing effort across stages is operationally inefficient.
Mistake 3
Confusing "lowest percentage" with "bottleneck." The close rate is structurally lower than earlier stages because the population is more qualified. Compare each stage to its own benchmark, not to other stages.
Mistake 4
Optimizing channels by lead volume instead of revenue contribution. A channel producing 1000 low-converting leads delivers less revenue than 100 high-converting leads. Optimize by revenue per dollar, not lead count.
Mistake 5
Failing to redefine stages when they become too coarse. If your conversion looks roughly even across stages, your stages are probably too few. Add granularity — separate "MQL" from "engaged MQL"; separate "proposal" from "negotiation." More stages = more visibility into where leakage happens.
Mistake 6
Computing conversion on the wrong cohort. Lead-to-close conversion measured by month is misleading because deals close on different timelines than they create. Use creation-cohort definition: deals created in period X, measured at maturity.
Try Mama free

Signal-anchored outbound converts at 5-10× the rate of cold outbound across every funnel stage.

Conversion-rate analysis usually points to outbound as the lowest-converting source. Mama's signal-anchored briefs change the math — same SDR effort, dramatically higher conversion at every stage from MQL to close. Fix the bottleneck source; the funnel fixes itself.

Start free trial → Read the win rate deep-dive