Goal Automation Throughput AnalysisPayments & Transfers

Task prompt

Product leadership is questioning whether our savings Goal automation is actually working as advertised. They've heard anecdotal reports that money isn't making it into users' Goal accounts reliably, and they want hard numbers before the next board deck. Two things worth knowing about how this data gets written down, so you read the amounts the way our finance team does. Our disbursement rows only get stamped with a separate settled amount when the money that moved came out different from the amount that was asked for; when a disbursement completed and that settled figure is blank, it's blank because nothing changed and the full requested amount landed. And a partial transfer doesn't count as completed for our purposes. Start with the delivery picture on Goal transfers: - How many Goal disbursement attempts have actually completed successfully - What's the real success rate for Goal disbursements specifically - The total dollar amount that landed in Goal accounts through those successfully completed transfers — count only the completed ones here, since I'm asking about the partials separately below - The average amount that landed per successful Goal disbursement, again across those completed transfers only - Averages get skewed by a few big movers, so also give me the median amount that landed across those same completed transfers - Staying with those completed transfers: how much less actually landed than we originally asked to move - On the Goal transfers that only went through partially, how much of what we asked for never made it - How much customer money is still in flight — the Goal transfers that are still pending, meaning they haven't completed and haven't errored out, so nothing has landed yet - Which year saw the most Goal disbursement activity - And our single best calendar month for Goal deliveries: how much money actually landed in that month? Again from the completed transfers only Then the part leadership will actually push back on. Every disbursement traces back to the automation that created it, and well over half of our automations have been deleted by their owners at some point. Judge deletion from the automation record itself, not from the rule row underneath it, and give me: - Out of that landed total — the completed-transfer figure from above, not counting the partials — how much of it came from automations that have since been deleted - Of the money still in flight, how much is queued against automations that are already deleted — money we're apparently still planning to move on behalf of a rule the user has switched off Last ones, for the risk section. Still looking only at those completed transfers unless I say otherwise: - What is the smallest number of customer accounts whose combined landed total reaches at least half of everything that landed in Goals? - Same idea, but for three-quarters: the smallest number of customer accounts whose combined landed total reaches at least 75% of everything that landed - And once more at 90%: the smallest number of customer accounts whose combined landed total reaches at least 90% of everything that landed - How many distinct customer accounts were active in that best calendar month above — accounts that had at least one completed Goal transfer land that month - Of the money that landed in that best calendar month, how much of it came from automations that have since been deleted? - And a reliability cross-check: how many customer accounts have both at least one completed Goal transfer that landed money AND at least one Goal transfer that errored out? Leadership wants to know whether failures and successes are concentrated in the same households - For those same dual accounts — the ones with both a successful landing and an error — how much total money have they landed through completed Goal transfers? That's the dollars sitting in the households that have also seen a failure If a handful of households account for most of this, leadership needs to hear that before anyone calls it a mass-market win. Return a single JSON object with keys: goal_fulfilled_count, goal_success_rate_pct, goal_actual_throughput_usd, goal_avg_actual_usd, goal_median_landed_usd, goal_fulfilled_shortfall_usd, goal_partial_shortfall_usd, goal_pending_in_flight_usd, peak_goal_year, peak_month_landed_usd, goal_landed_from_deleted_automations_usd, goal_pending_on_deleted_automations_usd, goal_accounts_for_half_of_landed, goal_accounts_for_three_quarters_of_landed, goal_accounts_for_90pct_of_landed, peak_month_account_count, peak_month_landed_from_deleted_usd, goal_accounts_with_error_and_landed, goal_landed_from_error_prone_accounts_usd