How complex is your Salesforce, really — and what would simpler be worth?

We make CRM simpler. We remove bloat with sharp insights and tooled methods — and the cost falls with the bloat.

Licensing analysis: the gap is not the opportunity — the qualified ladder is

Two quantities, on one scale. On the left, what the analysis measures: the named-license funnel from the contract down to the people who actually wrote business data. On the right, what is actionable — built rung by rung, each rung with its own threshold, its own owner and its own remedy. Subtracting one side from the other would give an enormous number nobody could defend; adding up the qualified rungs gives one a CIO can take into a renewal.

In the demonstration estate’s flagship org: 2,817 seats provisioned on the contract, 2,482 assigned, 1,936 logged in over the trailing 30 days, 622 with any recorded operation in the two-week event-log window, and 437 who wrote to a business object. The rungs above that read: 335 unassigned entitlement — the buffer, the position you already hold; 359 named users with no login in 90 days, exempt profiles excluded — a longer, safer threshold than the 30-day window, deliberately, because the number you can defend at a renewal is the conservative one; 92 read-only consumers who log in and write nothing, better served by the corporate BI tools you already run than by a CRM seat; and 185 non-business activations — system noise, plus integrations that were never people, which is simply the measured difference between the two bottom levels of the funnel. Four rungs, four different decisions, and not one of them is a deletion list.

Your numbers will differ. Measuring them is the only way to know by how much.

Desktop app for Mac and Windows, read-only via the Salesforce CLI. No Stood cloud: the analysis is local by default, and team sharing runs on your own GitHub or S3, with publisher and reader roles (security).

Everything resolves to the business process

Most org-analysis tools measure something real but unanswerable. A count of metadata components tells you the org is big. An aggregate health score tells you it is unhealthy. Neither tells you what to do on Monday.

Stood measures one unit: the business process — an object and a record type, with its stages, its automations, its users and its share of the contract. Structure rolls up to it. Usage rolls up to it. Cost rolls up to it.

The unit matters because a business process has an owner and a budget. A metadata component has neither. When the analysis says Concierge Intake Case: 34,558 records, complexity 24, no activity in 90 days, someone in the building is responsible for that sentence, and it can carry a line in next year’s plan. In the demonstration estate, that resolution turns 254 business processes across five orgs into a worklist: 126 with no activity in the period, 47 with no records at all.

Five lenses on one unit

Processes, volumes and stages. Which processes still live at all — a question most orgs have never been asked. The Mapping view lays each business process out as swimlanes, with active/total counts on every stage, so a pass-through stage or a bottleneck is visible, not inferred. Structure includes what hangs off the process: its custom and related objects, with volumes beside them, so a relationship built and never adopted reads as clearly as one carrying machine-written debris. The demonstration estate resolves to 254 business processes across five orgs: 126 with no activity in the 90-day snapshot window, 47 with no records at all — and, in the flagship org alone, 2,721,733 ghost records: records whose stage value no longer exists in the picklist definition, invisible to every list view, process and report grouping, yet present and counted.

Structure and complexity. Each business process gets a complexity score built from seven inputs: stages, profiles, layouts, lightning pages, validation and assignment rules, triggered flows and Apex triggers. The score is the plain sum of the seven counts — no weights, no model — so anyone can recompute it and see exactly what the structure costs to understand. Opportunity at 149, with 35 stages, 55 profiles and 42 pages, is checkable arithmetic, not a verdict.

Code and executions. The Apex view builds a dependency tree per trigger, tagging classes that carry record-type conditions, DML and SOQL, down to the source line. What matters is what ran: cold classes are detected from event log files, 380 of the flagship org’s 1,150 Apex classes with no observed execution in a two-week window — 33%. Audits relate to specific periods, and that is what makes the lens sharp enough to name bloat honestly. In/out call volumes come from the same logs. Underneath, in the flagship org: 17,643,827 backbone records and 108,576,580 related records — counted, not estimated.

Real usage. A CRM has to actionate concrete business; Stood Flows detects whether it does. Usage is measured from real business data operations in event log files — not just from logins — and it carries both dimensions: what people wrote, and whether they come back. The KPIs table shows, per process, who created and modified records in the window, and flags rows with no activity in the period. Cohort analysis buckets every connected user as daily, weekly, monthly or infrequent, month by month, with login seasonality overlaid year on year — which is what stops a two-week window from lying.

Governance and FinOps. The wrap-up: multi-org dashboards of portfolio scorecards, versioned baselines with deltas against a chosen snapshot, and the SELA reconciled line by line — in the demonstration contract, 102 lines across 9 orgs, a $6.57M/yr catalogue capped at $4.21M/yr, each SKU mapped to the org that consumes it. Financial engineering runs on simulations: the deactivation-policy view computes seats freed under your own rules, with exempt profiles excluded, and deactivates nothing.

The full arc, in the order the work runs: Analyse metadata → Usage & costs → Diagnostics → Govern — from the Stood Flows overview.

Diagnostics with named remedies

Stood Flows does not emit a grade. It emits insights — each with a name, a detection method and a remedy. Issues are what a human analyst or an agentic co-worker raises from them, qualified against business value, technical constraint and job weight, then tracked and snapshotted. No threshold engine is inventing work for you.

Zero-record record types · Zero- and low-activity processes · Ghost records · Stage & status equilibrium · Profile visibility sprawl · Related custom objects · Apex dependencies · Integration & I/O suspects.

An entry set, not a fixed catalog — it grows as new signals prove useful in the field.

Each diagnostic has its own page — the diagnostics index covers what a diagnostic is and how findings are qualified. The set is not a fixed catalog; it grows as new signals prove useful in real estates.

Extract, count, sum — then verify, then detect

A score nobody can decompose settles no argument. Each business process is sliced and structured from a defined set of metadata that Stood Flows extracts, counts and sums, and every item is individually verifiable against the org, through Setup or the CLI. That mechanic is what makes red zones detectable: when every number decomposes into counts, an outlier has nowhere to hide — and the number is a measurement rather than a rating. The same honesty applies throughout the site: where a figure is a simulation, the page says so.

Read the methodology →

Security and privacy by design

Nothing to breach, nothing for procurement to escalate: there is no cloud in the path — and, for the procurement reviewer who needs it in their own words, no vendor in the data path at all. SaaS scanners ingest your org’s metadata into their own tenancy; Stood has no cloud and never holds yours — read-only at three layers, no connected app, nothing sent to us. The analysis is local by default, and when a team shares it, sharing runs on your own GitHub or S3, with publisher and reader roles. The full architecture, and what is transmitted in each mode, is published on one page.

Security and trust →

Start with a Quickstart. Prove value fast. Then run it monthly.

Stood engagements begin as a Quickstart — a couple of hours with one of our experts, walking the first steps on your own org. Not the whole arc: two hours shows a first scan, the shape of the analysis and what your own event logs can support, which is enough to judge whether the rest is worth doing. Then a first scoped project, three orgs over three months say, that establishes a baseline, a qualified worklist and a rhythm. The rhythm is yours and continues monthly whether or not you buy services from Stood; snapshot over snapshot, the technical-debt curve bends. That is how CRM gets simpler — bloat removed with sharp insights and tooled methods, and the cost following it down.

From the blog: field notes on bloat

Most of what we know about bloat we learned removing it, and we write it down as we go. Recent notes on where complexity hides, what it costs to keep, and how estates actually get simpler:

All field notes →

Documentation: installation, org permissions and the metadata fetch tree at flows.stoodcrm.com.