> ## Documentation Index > Fetch the complete documentation index at: https://docs.mavera.io/llms.txt > Use this file to discover all available pages before exploring further. # Funnel Drop-off → Focus Group Investigation > Pull Mixpanel funnel data, identify worst drop-off steps, and run a Mavera Focus Group to diagnose abandonment reasons per persona ### Scenario Your Mixpanel funnel shows where users abandon your activation flow — maybe 35% leave after the onboarding wizard, or 50% never complete their first project. You know the *where* but not the *why*. You pull funnel data, identify the worst drop-off steps, and run a Mavera Focus Group with open-ended and ranking questions at each abandonment point. ### Architecture ```mermaid theme={"dark"} flowchart LR A["Mixpanel POST /api/query/funnels"] --> B[Step-by-step completion rates] --> C[Identify worst drop-off points] --> D["POST /api/v1/focus-groups"] --> E[Qualitative diagnosis per persona] ``` ### Code ```python Python theme={"dark"} import os, requests, time MP_SA = os.environ["MIXPANEL_SERVICE_ACCOUNT"] MP_SECRET = os.environ["MIXPANEL_SECRET"] MP_PROJECT = os.environ["MIXPANEL_PROJECT_ID"] MV = os.environ["MAVERA_API_KEY"] MB = "https://app.mavera.io/api/v1" MH = {"Authorization": f"Bearer {MV}", "Content-Type": "application/json"} mp_auth = (MP_SA, MP_SECRET) FUNNEL_ID = os.environ.get("MIXPANEL_FUNNEL_ID", "12345") r = requests.get( "https://mixpanel.com/api/query/funnels", auth=mp_auth, params={ "project_id": MP_PROJECT, "funnel_id": FUNNEL_ID, "from_date": "2026-02-15", "to_date": "2026-03-17", "unit": "month", }, ) if r.status_code == 429: time.sleep(60) r.raise_for_status() funnel = r.json() steps = [] meta = funnel.get("meta", {}).get("dates", {}) date_key = list(meta.keys())[0] if meta else None funnel_data = funnel.get("data", {}) if date_key and date_key in funnel_data: step_data = funnel_data[date_key].get("steps", []) for step in step_data: steps.append({ "name": step.get("event", "Unknown"), "count": step.get("count", 0), "avg_time": step.get("avg_time", None), "goal": step.get("goal", False), }) drop_offs = [] for i in range(1, len(steps)): prev = steps[i - 1]["count"] curr = steps[i]["count"] rate = (prev - curr) / max(prev, 1) drop_offs.append({ "from_step": steps[i - 1]["name"], "to_step": steps[i]["name"], "lost": prev - curr, "drop_rate": rate, "avg_time": steps[i].get("avg_time"), }) drop_offs.sort(key=lambda d: d["drop_rate"], reverse=True) funnel_summary = "\n".join( f" {i+1}. {s['name']}: {s['count']} users" + (f" (avg {s['avg_time']:.0f}s to reach)" if s['avg_time'] else "") for i, s in enumerate(steps) ) drop_summary = "\n".join( f" - {d['from_step']} → {d['to_step']}: {d['lost']} users lost ({d['drop_rate']:.0%})" for d in drop_offs ) PERSONA_IDS = os.environ.get("FUNNEL_PERSONA_IDS", "").split(",") if not PERSONA_IDS[0]: for name, desc in [ ("New Free User", "Just signed up on a free plan. Exploring the product for the first time. Limited patience."), ("Team Admin", "Setting up the tool for their team. Needs onboarding to be fast and clear. Accountable for adoption."), ("Technical User", "Developer or analyst. Expects API-first workflows. Impatient with GUIs."), ]: p = requests.post(f"{MB}/personas", headers=MH, json={"name": name, "description": desc}).json() PERSONA_IDS.append(p["id"]) time.sleep(0.3) worst = drop_offs[0] if drop_offs else {"from_step": "Unknown", "to_step": "Unknown", "drop_rate": 0} fg = requests.post(f"{MB}/focus-groups", headers=MH, json={ "name": "Mixpanel Funnel Drop-off Investigation", "persona_ids": [pid for pid in PERSONA_IDS if pid], "questions": [ f"You just completed '{worst['from_step']}' in our product. What would prevent you from continuing to '{worst['to_step']}'? Be specific about what you'd be thinking or feeling.", "Rank these barriers from most to least likely to make you quit: (1) Confusing UI, (2) Too many steps, (3) Need to invite teammates first, (4) Can't see value yet, (5) Technical error. Explain your #1.", "What's the ONE thing you'd need to see or experience to keep going through the onboarding?", "If you abandoned this flow, would you come back later or never return? What would trigger a return?", ], "context": f"""Mixpanel funnel analysis for our product onboarding flow: FUNNEL STEPS: {funnel_summary} DROP-OFF ANALYSIS: {drop_summary} Worst drop-off: {worst['from_step']} → {worst['to_step']} at {worst['drop_rate']:.0%}.""", "responses_per_persona": 2, }).json() for _ in range(24): time.sleep(5) data = requests.get(f"{MB}/focus-groups/{fg['id']}", headers=MH).json() if data.get("status") == "completed": break print(f"Focus Group: {data.get('id')} — {data.get('status')}\n") for resp in data.get("responses", []): print(f"[{resp.get('persona_id','?')}] {resp.get('question','')[:70]}") print(f" → {resp.get('answer','')[:250]}\n") ``` ```javascript JavaScript theme={"dark"} const MP_SA = process.env.MIXPANEL_SERVICE_ACCOUNT; const MP_SECRET = process.env.MIXPANEL_SECRET; const MP_PROJECT = process.env.MIXPANEL_PROJECT_ID; const MV = process.env.MAVERA_API_KEY; const MB = "https://app.mavera.io/api/v1"; const MH = { Authorization: `Bearer ${MV}`, "Content-Type": "application/json" }; const mpAuth = "Basic " + Buffer.from(`${MP_SA}:${MP_SECRET}`).toString("base64"); const FUNNEL_ID = process.env.MIXPANEL_FUNNEL_ID || "12345"; const funnelRes = await fetch( `https://mixpanel.com/api/query/funnels?project_id=${MP_PROJECT}&funnel_id=${FUNNEL_ID}&from_date=2026-02-15&to_date=2026-03-17&unit=month`, { headers: { Authorization: mpAuth } } ).then((r) => r.json()); const dateKey = Object.keys(funnelRes.meta?.dates || {})[0]; const stepData = dateKey ? (funnelRes.data?.[dateKey]?.steps || []) : []; const steps = stepData.map((s) => ({ name: s.event || "Unknown", count: s.count || 0, avgTime: s.avg_time || null, })); const dropOffs = []; for (let i = 1; i < steps.length; i++) { const prev = steps[i - 1].count; const curr = steps[i].count; dropOffs.push({ fromStep: steps[i - 1].name, toStep: steps[i].name, lost: prev - curr, dropRate: (prev - curr) / (prev || 1), }); } dropOffs.sort((a, b) => b.dropRate - a.dropRate); const worst = dropOffs[0] || { fromStep: "Unknown", toStep: "Unknown", dropRate: 0 }; const personaIds = []; for (const [name, desc] of [ ["New Free User", "Just signed up free. Exploring, limited patience."], ["Team Admin", "Setting up for team. Needs fast, clear onboarding."], ["Technical User", "Dev/analyst. Expects API-first. Impatient with GUIs."], ]) { const p = await fetch(`${MB}/personas`, { method: "POST", headers: MH, body: JSON.stringify({ name, description: desc }) }).then((r) => r.json()); personaIds.push(p.id); await new Promise((r) => setTimeout(r, 300)); } const funnelSummary = steps.map((s, i) => ` ${i + 1}. ${s.name}: ${s.count} users`).join("\n"); const dropSummary = dropOffs.map((d) => ` - ${d.fromStep} → ${d.toStep}: ${d.lost} lost (${(d.dropRate * 100).toFixed(0)}%)` ).join("\n"); const fg = await fetch(`${MB}/focus-groups`, { method: "POST", headers: MH, body: JSON.stringify({ name: "Mixpanel Funnel Investigation", persona_ids: personaIds, questions: [ `You completed '${worst.fromStep}'. What prevents you from continuing to '${worst.toStep}'?`, "Rank barriers: (1) Confusing UI, (2) Too many steps, (3) Need teammates, (4) No value yet, (5) Error. Explain #1.", "What ONE thing would keep you going through onboarding?", "Would you come back later or never return? What triggers a return?", ], context: `Funnel:\n${funnelSummary}\n\nDrop-offs:\n${dropSummary}\n\nWorst: ${worst.fromStep} → ${worst.toStep} at ${(worst.dropRate * 100).toFixed(0)}%.`, responses_per_persona: 2, }), }).then((r) => r.json()); let data; for (let i = 0; i < 24; i++) { await new Promise((r) => setTimeout(r, 5000)); data = await fetch(`${MB}/focus-groups/${fg.id}`, { headers: MH }).then((r) => r.json()); if (data.status === "completed") break; } for (const resp of data.responses || []) { console.log(`[${resp.persona_id}] ${(resp.question || "").slice(0, 70)}`); console.log(` → ${(resp.answer || "").slice(0, 250)}\n`); } ``` ### Example Output ```text theme={"dark"} Focus Group: fg_funnel_mp_3k9a — completed [New Free User] You completed 'Create Project'. What prevents you from 'Invite Teammate'? → I don't want to invite anyone until I've proven to myself this tool works. Asking me to invite teammates before I've done anything feels like a commitment I'm not ready for. Let me play with it solo first. [Team Admin] Rank barriers → #1: Can't see value yet. If there's no "aha" before the invite step, I can't justify pulling my team in. Show me a demo result with sample data before asking me to bring others. [Technical User] What ONE thing would keep you going? → An API key on the second screen. If I can hit an endpoint from my terminal within 2 minutes of signup, I'll finish onboarding. If the only path is clicking through a wizard, I'll bookmark it and forget. ``` ### Error Handling The Funnels API requires a pre-defined funnel. Create funnels in Mixpanel → Funnels, then grab the funnel ID from the URL or API. If building programmatically, use the Insights API with `funnel` analysis type. Funnel counts vary significantly by date range. Use at least 30 days for statistical significance. The `unit` param (`day`, `week`, `month`) controls aggregation granularity. 3 personas × 4 questions × 2 responses = 24 total responses. Allow 60–120s for completion. The loop provides \~120s. *** ## What's Next Back to Mixpanel integration overview Cluster profiles into behavior-based personas Retention strategies from cohort behavior Full reference for POST /api/v1/focus-groups