In January 2025, Hero Bread had a peculiar problem. Its bread was finding customers. Its finance team was still trying to finish October. A company moving into more retail locations was making decisions with numbers from another season. Success had acquired a reporting delay.
- Propeller supplies accounting, planning, and CFO judgment as an embedded team.
- Its sweet spot: growing businesses whose financial complexity has outrun their systems.
- The useful lesson: fix and document the process before automating it.
The bread was selling. The books were late.
According to Propeller’s account of the engagement, Hero’s NetSuite setup was not tracking inventory properly. That made the cost of goods sold difficult to recognize and channel profitability impossible to assess. Controllers had struggled to keep up. The bottleneck was becoming a question about the business itself: which growth was actually paying?
Propeller joined in January, helping implement a new NetSuite instance, repair the balance sheet, and redesign the chart of accounts. It worked across departments, connecting raw materials, logistics, and customer channels. The reported result was a 10-day close, with current unit economics and a common financial picture for leadership.
The story makes Propeller’s role unusually tangible. A finance partner can change the interval between something happening and management understanding it. In a growing business, that interval can become expensive long before anyone calls it a crisis.
Different reporting states, shown side by side. No percentage reduction implied.
The CFO who kept showing up
Chris Fenster founded Propeller in 2008 after seeing a recurring weakness in startup advice. Experts appeared for fundraising, a board cycle, or diligence, then disappeared. Their recommendations survived; their accountability often did not. His answer was to keep a team close enough to see what happened afterward.
Accounting provided the daily connection. The CFO could help with strategy because colleagues were already inside the ordinary work of transactions and reporting. As clients grew, Propeller could change shape, eventually supporting the internal finance departments those companies built.
“Staying turned out to be the strategy.”
Chris Fenster, Founder & Executive Chairman
That history explains today’s preference for the word “embedded.” The firm wants a place in weekly operating rhythms and board preparation. Its current team page lists Raymond Nieuwenhuizen as CEO and Fenster as executive chairman. The company now reports more than 1,500 businesses served cumulatively, including 26 unicorn clients and alumni. These figures describe its client history, rather than a guarantee about the next engagement.
A finance team, assembled to fit
Propeller sits between basic bookkeeping and the expense of building a complete finance department. Its embedded stack combines executive guidance, controllership, accounting support, and financial planning. Specialists can handle inventory, trade deductions, revenue operations, equity matters, and fundraising preparation. Transformation services extend the work into platform selection, implementation oversight, and analytics.
The typical customer, the company says, has revenue between $3 million and $300 million or more. Food and beverage, software, consumer commerce, health and wellness, and Web3 all feature in its market. Their problems differ: a retailer deduction behaves differently from recurring software revenue. Sector experience matters because an accurate ledger still needs interpretation.
Dutch, the veterinary telehealth business, illustrates a less obvious arrangement. Propeller handles accounting, Dutch’s own team leads FP&A, and Propeller CFO Anthony Rosen provides senior strategic support. Founder Joe Spector had worked with Rosen during Hims’ growth. The relationship carried forward; the staffing arrangement was fitted around Dutch’s internal capabilities.
For a founder comparing alternatives, the distinction is practical. A bookkeeping provider may cover the records; an independent CFO may supply senior advice. An internal department brings direct management control. Propeller combines several levels of work under an ongoing engagement, with its own full-time employees. Whether that arrangement earns its cost depends on how much of the combination a business actually needs.
The commercial logic follows the staffing logic. Engagements run month-to-month, while the intended relationship lasts years. Propeller says its combined function can typically cost 20-35% less than hiring a full-time CFO alone. Treat that as the company’s comparison, dependent on scope. A buyer should compare actual roles, hours, deliverables, and responsibility with the equivalent internal team.
The AI experiment that gave back no time
In August 2026, Propeller published a revealing failure. Mabela Zenullari tested whether Claude could perform a real client’s monthly close. The team reconstructed the books before closing and compared the agent’s work with the controllers’ completed close.
Reviewing the agent’s output took about as long as doing the job manually. Reconciliations were being improvised rather than executed against a written standard. When something differed, reviewers could not easily separate a model error from an undocumented human practice.
The response, led by research and development vice president Zain Khoja, was specific: assign controllers general-ledger accounts and document the inputs and ordered steps for each reconciliation. Propeller also described live Amazon merchant and payroll reconciliation tools built with workflows defined first. The failed pilot changed the sequence of work. Automation would follow an explicit process.
Build the process you want to keep
PropellerOS expresses the larger technology bet. The platform aggregates information across ERP, banking, payroll, and CRM systems, adding cash-flow views and peer benchmarking. The underlying Data Lake and Intelligence Engine launched in 2025. The promise depends on connected data and people who can challenge the output.
There are practical limits. A business seeking a three-month assignment is outside Propeller’s preferred relationship; its own guidance points such buyers toward fractional providers. A complicated platform also cannot settle unclear ownership. Someone must know which numbers count, how they were produced, and who approves them.
Its September advice on ERP upgrades makes the same point in ordinary software terms. Selling through several channels can introduce different revenue-recognition rules. More orders can overwhelm a reconciliation that once fitted into Friday afternoon. Those changes require deliberate procedures. Replacing the platform before understanding the procedures can move the confusion into a more expensive system. For prospective customers, this suggests a useful first conversation: ask where the delay begins, which handoffs create errors, and what the team will measure after the change. A demonstration can show a dashboard. A working engagement needs an agreed test of whether the underlying information became more reliable and easier to use.
The reader can copy that discipline without hiring anyone. Choose one reconciliation. Name its owner. Write down the inputs, sequence, exceptions, and review standard. Measure the time spent checking as well as producing it. Hero Bread needed fresher answers. Propeller’s AI pilot needed a repeatable method. Both lead back to the same managerial task: make the work understandable enough to trust.