Identify Conversion Gains From Recent Changes

You are a sales operations analyst at a digital marketing agency. You isolate the impact of sales-process changes from noise and seasonality.
Agency: [AGENCY_NAME] — [SERVICES] | Baseline Period: [BASELINE] | Test Period: [PERIOD] | Changes Made: [CHANGE_LIST] | Data:
[FUNNEL_DATA]
For each change, compare stage conversion rate(s) it was supposed to influence in [BASELINE] vs [PERIOD], call out the lift (or drop), and assess whether the change should be kept, killed, or extended.

- Compare matched stages only; do not mix funnels.
- Show absolute and relative change (pp and %).
- Flag any comparison with 
One block per change: Change | Stage(s) impacted | Baseline rate | New rate | Delta (pp / %) | Confounders | Verdict. End with one-line meta-takeaway.

• [AGENCY_NAME] — Your agency name
• [SERVICES] — Service lines
• [BASELINE] — Baseline period before changes
• [PERIOD] — Period after changes were implemented
• [CHANGE_LIST] — Specific changes shipped (template, script, page, flow)
• [FUNNEL_DATA] — Per-stage conversion data for both periods

Agency: RudderShop Digital — SEO + Web | Baseline: Q4 2025 | Test: Q1 2026 | Changes: (1) new discovery script with 4 qualifier questions, (2) one-pager proposal template, (3) calendar booking added to website | Data:
Q4: Lead->Disc 55% (n=120), Disc->Prop 48% (n=66), Prop->Won 41% (n=32)
Q1: Lead->Disc 71% (n=140), Disc->Prop 44% (n=99), Prop->Won 58% (n=44)

Change 1: New discovery script
Stage: Disc -> Prop | Baseline 48% -> New 44% | Delta -4pp / -8%
Confounders: Q1 inbound volume up 17%; possibly more low-fit leads.
Verdict: INCONCLUSIVE — rerun after lead source mix is stable.

Change 2: One-pager proposal template
Stage: Prop -> Won | Baseline 41% -> New 58% | Delta +17pp / +41%
Confounders: none identified; sample (n=44) acceptable.
Verdict: KEEP and EXTEND — roll out as standard template.

Change 3: Website calendar booking
Stage: Lead -> Disc | Baseline 55% -> New 71% | Delta +16pp / +29%
Confounders: same period as Change 1 lead-volume spike, but conversion still rose.
Verdict: KEEP.

Meta-takeaway: Two of three changes earned their keep; total funnel produced 1.8x more wins (~13 vs 26) on similar effort.

• Hold lead source mix constant when comparing periods, or your verdict reflects sourcing, not your change.
• Use absolute pp change AND relative % — a 4pp lift means very different things at 10% vs 60%.
• If sample is thin, extend the test period rather than rushing a verdict.