One Model, All Purposes: Sal Tirabassi on What a Fractional CFO Actually Does

🎧 Listen: Unscripted Small Business on Castos

“Fractional” is one of those words that can mean anything, so host Jeremy Rivera opens by asking whether it’s just a fancy term for part-time. Sal Tirabassi’s answer is more specific than that: permanent part-time, sometimes one person, sometimes a three-person team layered underneath a strategic CFO who works directly with the owner.

“The objective for them is really that they get an A-grade service for much less than having to build out a whole team on their own. So they get it without the overhead.” — Sal Tirabassi

Tirabassi runs CFO Pro + Analytics, serving founder-owned, family-owned, and sometimes venture-backed businesses between $3 million and $100 million in revenue — the full stack from strategic CFO work down through FP&A, accounting, audit, and bookkeeping.

Why this model exists now

The service isn’t brand new — Tirabassi has met fractional CFOs who’ve done it for 18 years — but the market changed. “One thing that definitely has made the whole model much more viable is the whole advent of the communication systems that we now are accustomed to because of COVID.” A business anywhere can now shop a national market for the specific expertise it needs.

For bootstrapped or lightly capitalized companies, the arithmetic stacks: it costs less than full-time, it carries no benefits or office overhead, and — he argues — “you’re getting a much better service on average,” because you’re buying someone who has seen more situations than you could afford to hire outright.

Analytics is not an add-on — it’s the philosophy

Tirabassi has an MBA and a graduate degree in engineering, and he brought a rule with him into every CFO seat he’s held: one source of truth, all running through finance.

“All data that is owned by the organization should be owned, cleaned and managed by one part of the organization. And my worldview is that it should be done by the finance team.” — Sal Tirabassi

In a huge company that’s nearly impossible. In an emerging business that hasn’t figured it out yet, the CFO can genuinely corral it all. When he ran finance in-house, the largest part of his team were operations analysts assigned to specific areas of the business and told to treat those areas as clients: show up at their weekly staff meeting with their KPIs and whatever insight the data supports.

That translated directly into the services business. For clients who want it, they’ll stand up a business intelligence platform and then operate it — typically Power BI in Microsoft shops. His timeline: “you can be up and running with a business intelligence system in like six weeks, with some really good information coming out.”

A practical note for anyone considering it: you don’t want reporting queries hitting production systems, so a nightly replication is usually fine. “You’re using day-old data, which is generally good enough for most applications.” And a formal data lake only becomes necessary at hundreds of millions of rows — below that, Power BI is happy connecting to a hodgepodge of repositories.

The first three priorities on a new engagement

1. Scrub the bookkeeping. Most clients arrive with a financial planning problem — cash flow management, or simply “I need to know what’s going to happen in the next 90 days.” Before you can forecast, you have to know how bad the historicals are, because the forecast is worthless if you can’t tie actuals back to it. He uses a tool called Bunker to plug into QuickBooks or NetSuite and get a fast read on how clean the chart of accounts really is. “A lot of times it is” messy — and knowing exactly how messy is the deliverable.

2. Learn mechanically how the business works, then build one model. Not the sales pitch version — the mechanics. How do they acquire customers? How do they market? What do salespeople get paid? How do they service customers? What goes into product development? That feeds a driver-based model, and here he states a best practice worth pinning up:

“You should only have one financial model and it should serve all your purposes. It should give you your monthly budget, it should give you your long-range forecast, it should be what you share with your board and with your investors, and it should have the ability to flex so that you can show upside and stress cases — and you’re just operating in one system.” — Sal Tirabassi

Gathering the inputs takes time; with strong Excel modelers, a good draft takes a few weeks.

3. Find out where the executive team’s head actually is. Why do the stated priorities matter to them, and can you deliver a better version than they asked for? The diagnostic he’s running: is this an organization that whipsaws and can’t commit — in which case the intervention is a planning process and a meeting cadence — or one that’s organized and simply starved of good data? “It’s gonna be somewhere in between those two bookends.”

The engagement he walked away from

Asked whether he’s ever been handed a sinking ship, Tirabassi describes an entrepreneur who was “way out over his skis in terms of how he had allocated capital” — asset rich, cash poor, on top of a complicated set of books that made cash management punishing.

They tried cleaning up the books. Then he told the client to fire him.

“You don’t really need us. What you really need to do is start shedding some assets… The CFO function is going to be an expensive layer that you don’t really need to spend the money on right now, because you bought too many things. You need to sell some of them and retrench, and then go back and hit it hard with all of your financial planning in a much more thoughtful way.” — Sal Tirabassi

Where the threshold is

Jeremy pushes for the weight classes — how small is too small? Tirabassi splits the answer.

Bootstrapped, under ~$3M revenue: hard to justify his level of service. What would fit is a few hours a month — help picking a good bookkeeper, then a light monthly financial review.

Over ~$5M revenue: this is where he says it becomes obvious. Roughly $120,000 a year for a professional part-time finance function is “worth its weight in gold,” and non-VC-backed owners tend to grasp it fastest, because they’re already living inside their own cash flow every week. “If somebody can help me understand this cash flywheel better, it’s actually going to make me more successful.” Hot Girl CFO’s Meaghan Wall makes the same case from the other side of the desk — cash flow is queen, and it is also baby.

Up to ~$100M: still valuable, but the shape shifts — more consultative strategic finance, less hands-on, because even at that scale you may not be able to hire a full-time CFO who’s seen as much.

Venture-backed small companies are a different animal entirely. There’s cash in the bank, so it’s not about turning the flywheel — it’s planning and insight so they look professional and can defend their story. “What’s the messaging going to be to investors? What’s the data that you have to back it up?” His view: get the fractional CFO in before you go out to raise, not during.

Practical AI, from someone who has to be right

Tirabassi uses Claude, ChatGPT, and Perplexity daily and says he doesn’t really go to search engines anymore — which he thinks has serious implications for where ad dollars eventually land. His concrete use cases are the useful part.

Excel and Power BI formulas. “Give me the formula that’ll always pick the third word in the cell… those left, right, mid functions are extremely complicated. You just go there, it sets itself to say, pretend A2 is the cell you want, and then it gives you the formula. You know what you want to do and you can build it yourself, but why bother?”

Complex accounting questions — cross-referenced. He’ll ask about handling accrued revenue on a mid-month recurring transaction under ASC 606, then run the same question through two or three different models. “I know a lot about accounting, but I’m not a CPA. And so I want to compare and cross-reference.” Sometimes it’s to check other people’s work; sometimes it’s to walk into a conversation with the accountants able to guide it.

Recovering data out of your own old charts. This one made Jeremy sit up. Take a screenshot of a chart in a report you produced months ago, upload it, and get the underlying numbers back as a table you can paste into Excel — no hunting for the original spreadsheet.

“The trick is that whenever you make new charts, always select the feature where it includes the data point for each bar — because that’s what the GPT looks for in order to rebuild the table.” — Sal Tirabassi

He uses the same trick on third-party data found on the web: pull the numbers, rebuild the chart in your own colors, and footnote the source.

And a discipline most people skip: keep chat threads separated by purpose. He keeps one dedicated “technical support” thread in Claude. Drop a blog outline request into it and “you’re mixing the awareness that the LLM has… otherwise you could veer off into certain areas that end up with really strange answers.”

On the low-code side, he points at tools like Zapier as genuinely useful for owners now: pull in Google campaign data, join it to CRM data, push it into a repository you can run BI on, all without involving IT. “If you have four hours on a weekend, there’s a lot of really interesting stuff that you can do to create small systems that will make you more scalable and measurable.”

Bonus: the messaging-sprawl fix

Asked whether every client drags him onto a different platform, Tirabassi plugs a tool he wrote about on his own blog in a post called From Many Messaging Apps to One: Beeper, which pulls WhatsApp, LinkedIn, Telegram, Facebook Messenger, Slack, and Discord into one desktop and phone app. “I would probably die without it.” (iMessage and Microsoft Teams are the notable exceptions.)

One compliance note for service providers: with a client regulated under HIPAA, you don’t get a choice — you have to work inside their platform and their mailbox.

The most satisfying part of the job

Jeremy closes by asking what question he should have asked. Tirabassi picks his own, and answers it.

“The most satisfying thing is you give them an aha moment and it really changes the way they think about things… you did all this work in the background and you came up with these critical insights that may be completely opposed to their perception, or could be completely validating it — and you build a certain type of trust through that experience.” — Sal Tirabassi

Connect with Sal

Sal Tirabassi runs CFO Pro + Analytics. He writes about analytics and execution effectiveness at tirabassi.com, publishes longer technical pieces on Substack (free, roughly every couple of weeks), and is on LinkedIn.