Credit Overrun at a Conservation Nonprofit
Riverbend Conservation Trust has 85 ArcGIS Online members and an annual allocation of 12,000 credits. By 28 April it has consumed 8,400 credits, and the remaining 3,600 must last through 31 December. The finance team will not purchase more credits this year. The monthly credit report shows a spike beginning in February, when a scheduled Python notebook started geocoding 25,000 volunteer addresses every Sunday and writing results to a hosted feature layer. Two program dashboards also run a weekly spatial analysis that joins Census data. Separately, 15 field staff depend on Field Maps with offline areas for spring surveys; disabling field collection is not acceptable. The organization has no custom roles yet, and all Creator members can run geocoding and spatial analysis. The new administrator needs to stabilize spending without blocking the spring survey or the public dashboards.
1. What should the administrator check first to explain the overrun, and how should the evidence separate the notebook, the dashboards, and the field workflows?
Read the worked answer
Start with the organization's credit usage report, not with the field apps. Filter to 1 January through 28 April and group by credit category and by item or user. The expected pattern is geocoding credits concentrated on the notebook's output layer and notebook owner, spatial analysis credits on the two dashboard data sources, and little or no credit use tied to Field Maps offline areas unless those areas regenerate tiles or run analysis. Confirm the notebook's schedule and last run, then inspect each dashboard's underlying item to see whether its weekly join is scheduled or manual. This separates a recurring automated cause from a one-time publication cost. Check whether the spike begins exactly when the notebook was created or its schedule changed; if so, the notebook is the primary driver. Do not disable Field Maps first, because the report should show whether offline use is material. The key is to match each credit-consuming capability to the item and owner that invoked it.
2. Recommend immediate controls for the next 30 days that reduce spend while preserving the spring survey and public dashboards.
Read the worked answer
Pause the notebook schedule immediately; the historical outputs remain. Repoint the notebook to a local address locator or limit geocoding to new volunteer records, then run a small test batch and measure credits. For the dashboards, change the weekly spatial analysis to a cached result or run it monthly; if it must stay weekly, restrict it to a smaller extent. Temporarily remove geocoding and spatial analysis privileges from all but a small custom role or group, and move the notebook owner into that group. Keep Field Maps offline areas for the 15 staff, but ask them to refresh areas only when needed. Set a weekly credit check and a threshold alert with finance. These steps target the identified drivers while preserving the spring survey. Avoid simply banning all analysis, because that would stop the public dashboards and likely lead to shadow workarounds.
3. What long-term governance model should the administrator put in place to prevent another mid-year overrun?
Read the worked answer
Adopt a role-based credit governance model. Create custom roles: a field editor who can edit features but not geocode or analyze, an analyst who works in a sandbox group, and a publisher who can publish but not schedule geocoding batches. Require review before any scheduled notebook or geocoding batch is created. Put a named owner and business justification on every credit-consuming item. Build a monthly report by item and owner, and reconcile it with the annual allocation. For recurring geocoding, use a local locator or a vetted address service rather than the world service. For dashboards, cache analysis outputs and refresh on a defined cadence. Document the 12,000-credit budget and the escalation path if 80 percent is consumed before September. This prevents a repeat without blocking legitimate field work.