Sign in to Pulse
Enter your email — we'll send a magic link, no password needed.
Link sent to . Check your inbox.
Access pending
You're signed in, but your account isn't on the Pulse allowlist yet.
Ask vamsee@scalability.club to add you in tasks.members.
What's moving — auto-generated insights
Computed on every load from the live data behind all tabs. "Last week" = most recently completed Mon–Sun.Rider funnel — all-time
Every rider who has ever entered the funnel, by the furthest stage reached to date. Read as two owned halves: Funnel conversion = Lead → Onboarded (acquisition) and Business success = Onboarded → Activated (activation) — always quoted together, never as a bare "conversion".This week vs last week
Last week = most recently completed Mon–Sun.BD vs Non-BD, all-time
Onboarded attempts (attempt-grain), split BD vs Non-BD.Everything below this tab is the detailed, analyst-level breakdown — cohort tables, retention triangles, methodology notes — behind the same numbers shown above.
Methodology — attempt-week cohorts & grain (click to expand)
Attempt-week cohort (2026-08-21 rework). Each row is a real lead attempt, not a contact — bucketed by the week that specific attempt was created. A contact who tries, fails, and retries weeks later shows up in both weeks; the attempt that actually converts gets the credit in its own week, not retroactively folded into the first one. Once a contact has ever Won (reached Activated), every later lead they generate is a reactivation and is excluded here entirely — see
Grain: every count on this tab is attempt-grain (flow — cohorts, conversion). Population/stock counts (how many people) are contact-grain and live on the Summary tab headline — never compare the two directly (27,492 attempts ≠ 26,438 people).
Segment: every attempt carries a BD/Non-BD tag from creation (valid Bounce Daily booking dated before that lead) — so BD + Non-BD = total, at every stage.
docs/Funnel.md for that as its own dimension. Because reached_installed ⇒ reached_lead, reached_kyc_done ⇒ reached_installed, etc. hold per attempt, each column is ≤ the one before it — nothing here can read over 100% or increase step to step.
Grain: every count on this tab is attempt-grain (flow — cohorts, conversion). Population/stock counts (how many people) are contact-grain and live on the Summary tab headline — never compare the two directly (27,492 attempts ≠ 26,438 people).
Segment: every attempt carries a BD/Non-BD tag from creation (valid Bounce Daily booking dated before that lead) — so BD + Non-BD = total, at every stage.
Lead Cohort Funnel
Rows = week the attempt was created. Each column = how many of that same attempt-cohort have reached that stage by now.Full funnel by segment — BD vs Non-BD
Every attempt is tagged at creation (BD = valid Bounce Daily booking dated before that lead), so BD + Non-BD = all attempts at every stage. Respects the week-range filter.BD vs Non-BD — Onboarded (attempt-grain)
BD = had a valid Bounce Daily booking before this lead's date (per-lead tag). Non-BD = no earlier valid booking — says nothing about lead source.Non-BD: deposit tier split
₹99 vs ₹399 sign-up deposit tiers (deposit is independent of the BD tag).BD vs Non-BD by city
Weekly Rider Mix — New / Retained / Resurrected vs Churned
Active = delivered ≥1 order that week. Counts are riders (people, contact-grain), not attempts.Mix detail — absolute & %
Avg active days/week = stickiness signal. WAR = distinct riders active (≥1 delivered order) that week. Computed as (avg daily distinct riders ÷ WAR) × 7. Weeks with <7 days of order data are marked partial.
Active Riders Deep-dive — avg active days/week
DAR/WAR detail
Retention triangles are anchored on Active week — same weekly-active population as the mix above. W+0 = full active headcount that week. W+k = of that same group, how many are also active k weeks later. Rider activity data only goes back to , so earlier cohorts show blank cells.