What Pulley's Shutdown Teaches Us About Startup Financial Metrics
Updated: Oct 5
On September 16, TechCrunch reported that Pulley — the cap-table management startup backed by General Catalyst, Stripe, and Founders Fund — is shutting down on December 8. The company raised more than $50 million. It is now helping its customers migrate to Carta, the very competitor it was built to beat.
Read that again: $50 million from three of the smartest names in venture capital, and the endgame is handing your customer list to your rival.
I spent years as a CFO in technology companies before moving into tax and accounting practice. I have sat through board meetings where this kind of ending becomes inevitable — usually 18 months before anyone says it out loud. The startup financial metrics that predict it are not secret. They are just not the metrics founders like to talk about. This article is about the numbers that actually matter, using Pulley as the freshest case study we have.
The Case in 60 Seconds: What Happened to Pulley
Pulley was founded in 2020 by Yin Wu, a serial entrepreneur and former Microsoft engineer, to manage cap tables — the ownership records of startups. Solid market, real problem, credible team, top-tier investors. The company announced no official reason for closing. But a former employee offered a theory worth every founder's attention: Pulley's real competitor wasn't Carta. It was spreadsheets — because "startups can easily use AI to keep, improve, and maintain their spreadsheets."
Sit with that. A venture-funded software company may not have been out-executed by a rival. It may have been out-executed by a $20-per-month AI subscription that made its core value — structuring ownership data — something customers could do themselves.
Both threads of this story deserve unpacking: why $50 million guarantees nothing, and why product-market fit is not a permanent achievement.
Funding Is Not Revenue: The Math Founders Skip
Here is the accounting reality that gets lost in fundraising announcements. Money raised sits on the balance sheet as cash — with an obligation attached. Revenue runs through the income statement — money customers pay because the product is worth it. Confusing the two is the most expensive category error in startups.
What does $50 million actually buy? Runway. At a burn rate of $1.5 million a month — a modest number for a venture-backed software company with engineering salaries, sales teams, and cloud costs — $50 million is roughly 33 months of life. Less, in practice, because burn grows with headcount. Funding buys time to build a machine that turns a dollar of spending into more than a dollar of durable gross profit. If that machine never materializes, the runway simply measures the distance to the cliff.
The data says this is the norm, not the exception. In CB Insights' well-known post-mortem analysis of failed startups, running out of cash or failing to raise new capital appears in 38% of failures, and "no market need" in 35% — and the second usually causes the first. Startups rarely die because the bank account empties by surprise. They die because the unit economics never justified refilling it.
Valuation deserves the same skepticism. A valuation is the price of the last transaction in your stock, negotiated between two parties with their own incentives. It is not an appraisal of the business. Plenty of companies have carried nine-figure valuations straight into asset sales.
The Five Startup Financial Metrics a CFO Reads First
When I evaluate a company — as an operator, an advisor, or an accountant reconstructing what went wrong — the amount raised is close to the last number I look at. Here is what I read instead, and what "healthy" roughly looks like for a software business:
1. Customer acquisition cost (CAC), fully loaded.
Not just ad spend — sales salaries, commissions, tools, and marketing overhead, divided by new customers won. Founders routinely understate CAC by half because the sales team's payroll hides in a different budget line.
2. Net revenue retention.
Of the customers you had a year ago, what do they pay you today, including upgrades and churn? Below 100% means your bucket leaks faster than growth can pour; the strongest SaaS businesses run 110–120%. For a cap-table product priced per stakeholder, churn happens invisibly — every portfolio company that dies or gets acquired takes its subscription with it.
3. Gross margin.
Revenue minus the direct cost of delivering it. Software should live at 70–85%. If humans must touch every account — onboarding, support, compliance review — you may be running a services firm with software pricing, and the fundraising math stops working.
4. Burn multiple.
Net cash burned divided by net new annual recurring revenue added. Burning $6 million to add $2 million of ARR is a burn multiple of 3 — tolerable for an early-stage company, alarming past Series B. Under 1.5 is strong. Above 3, sustained, is the metric equivalent of a check-engine light.
5. Months to break-even at current trajectory.
Not a hockey-stick model — the honest version: if growth and burn continue as they are, does the line ever cross? A company that cannot sketch a credible path to breakeven is, by definition, planning to be rescued — by the next round, or by an acquirer.
None of these five requires a data science team. They require clean books and the willingness to look. That's the unglamorous part: if your bookkeeping can't produce these numbers, that's the first fix — I've written before about why bookkeeping is the non-negotiable foundation of every financial decision that follows.
The AI Lesson: Product-Market Fit Has an Expiration Date
The spreadsheet comment from Pulley's former employee points at something bigger than one company. A large slice of the software economy consists of products whose core value is structuring data that used to be messy — cap tables, expense categorization, contract summaries, report formatting. That is precisely the work generative AI is commoditizing fastest.
Product-market fit is not a trophy you win once. It is a lease that technology shifts can terminate early. The uncomfortable exercise for every founder — and every business owner, tax practices included — is to ask: which part of what my customers pay me for could AI make nearly free within two years? If the honest answer is "the main part," the strategy conversation needs to happen now, while there is still runway to reposition — not after the renewal rates start telling the story for you.
Notice what this means for the funding question, too. Capital cannot buy protection from this. Pulley's investors were as sophisticated as they come. Fifty million dollars did not stop the ground from moving under the product.
What to Check in Your Own Business This Quarter
You do not need to be venture-backed for any of this to apply. Five checks, in order:
Compute your fully loaded CAC — every sales and marketing dollar, not just ads — and compare it to the gross profit a customer generates in their first year.
Measure revenue retention: what did last year's customers pay you this year?
Pull a real gross margin, with delivery labor included in the cost line.
If you're pre-profit, calculate your burn multiple and months of runway at today's spending.
Write down, in one sentence, what AI would have to do for your customers to stop needing you — then check the app store for who's building it.
And keep the tax layer connected to the strategy layer: growing companies with engineering spend should be modeling the restored R&D expensing rules under Section 174A — cash recovered from the IRS is runway that costs zero dilution. Your baseline numbers, including what the IRS actually has on file, are a login away in your IRS Business Tax Account.
Where it gets complicated enough to warrant help: when the metrics disagree with each other, when investors and the P&L are telling different stories, or when you suspect the books themselves are hiding the answer. That is CFO-level work, and it is exactly what we do at TaxesZenPro — we read the business, not just the return.
Because $50 million raised is not profit. It is money someone still has to turn into a business — and the five numbers above tell you, early and unsentimentally, whether that is happening.
Comments