From PwC to founder: what an auditor's brain does in a startup
Years at PwC left me with one habit that matters more than any technique: asking where a number came from. It has saved me from my own pitch deck.
By Ali Asif,
The first thing I do with any number is ask who told me.
I picked the habit up at PwC and never put it down. I hold a CPA and an ACCA. The letters matter less than the reflex they trained. A figure is a claim, somebody made it, and somebody else can check it. When I rebuilt a client's numbers as a junior there, the corrected figure moved by a number large enough that nobody in the room argued with it. I can't tell you more about that, because the work was confidential and I'd keep it that way even if it weren't. I can tell you what stayed with me. The error wasn't dramatic. It sat in a place nobody had looked at since someone else set it up, and everyone had been quoting it.
I was junior, and the correction was big. What got it taken seriously, as far as I can reconstruct it, was showing the source for every line rather than sounding sure.
Where it shows up now
In the pricebook. At Northlane we keep every supplier quote we've received. At my last count, 77 of the 236 rows were past their validity date. They're flagged, and an agent can't quote them as current. The easy move is to say all rates are fresh. The auditor in my head won't allow it.
In my own product claims. For ProCul, the accounting product I'm building, the temptation is to say it closes books. What I can stand behind is narrower. It has passed its tests against a synthetic company with planted errors, and it has no customers yet. I write the small, checkable sentence and label the rest as unproven.
In a bid. Each price in a tender response traces back to a document: a quote, a prior RFQ, a dated benchmark. A rate without a source gets a flag, not a pass.
In my own sentences. The NTDP line I deleted before sending is the cleanest case. I'd written "invited" about a program where I'd only applied. The test I now run before anything goes out is the same one I'd run on a working paper. Can I point to the evidence?
What the habit doesn't do
It doesn't make me fast. Auditors are paid to be skeptical, and a startup is paid to move. Early on I over-checked things that didn't matter. I'd spend an afternoon reconciling a figure that was never going to change a decision. A founder needs to know which numbers are worth the afternoon. Mine are the ones that go outside the company: a price, a claim to an investor, a number in a bid.
It also doesn't protect me from the thing it's worst at, which is optimism. I can verify a number all day and still be wrong about whether anyone will buy. The auditor in me checks what happened. He's useless at telling me what comes next.
And there's a pull in the other direction that I try to watch. Because I came from a firm where accuracy was the product, I can mistake carefulness for progress. Some weeks the spreadsheet is cleaner and nothing has shipped.
The transfer, if you're making the jump
If you're an accountant thinking about starting a company, I'd keep the reflex and drop the pace. Three things carried over for me, and two things I had to unlearn.
What carried over: asking for the source, keeping a record of why a decision was made, and a fairly thick skin about being questioned. Having your work reviewed line by line makes a hard meeting feel more survivable.
What I had to unlearn: waiting for complete information, and treating a draft as an embarrassment. A first version of a bid or a page or a product can be wrong in a way that costs nothing, as long as it's labeled draft and a person looks at it before it leaves the building.
My regulated-trust ventures sit in that overlap. They're built on the idea that people will pay a neutral party to say, "I checked, and here is how." That's what the job taught me to sell.