Section 1
Key takeaways
• Carta tracked a record 966 startup shutdowns in 2024, up 25.6% from 769 in 2023, and admits the real number is higher because it only sees its own U.S. customers . • In CB Insights' post-mortem of 431 venture-backed companies, the median one died just 22 months after its last raise, with 70% having run out of capital . That is your real retainer horizon, not the 36-month contract. • For a service firm, any single client at 10%-plus of revenue already meets the definition of client concentration risk ; a client above 20–25%, or a top three above 50%, reads as a structural threat to a buyer . • Startup-heavy books fail on both axes at once: high per-client mortality and high concentration, the exact profile that turns your EBITDA from earnings into "a lottery ticket" . • Manage your book the way a fund manages duration: by expected lifespan, not by logo prestige.
Section 2
Why does a startup client look like growth and behave like churn?
Start with the concrete case, because the abstraction only lands once you've seen it on a real P&L. Picture a 12-person branding-and-growth studio doing roughly $1.6M a year on retainers. Last year the team landed three logos they were thrilled about: a seed-stage fintech, a Series A consumer app, and a "pre-seed but the founder is ex-FAANG" AI tooling company. Combined, those three accounts represent about $480K of annual revenue, call it 30% of the book. The deck looks incredible. The case studies write themselves. Sales uses these names on every discovery call. Now run the actuarial math the founder didn't run. CB Insights examined 431 venture-backed companies that shut down since 2023 and found the median one died just 22 months after its last fundraise . Not 22 months from founding, 22 months from the last time money hit the account. Your shiny new client signed a 12-month retainer, but the clock that actually governs your revenue is the runway clock, and it is often shorter than the contract. When the cash gets tight, marketing and outside services are the first line items a board cuts. You are not a fixed cost to a startup. You are discretionary spend with a renewal date that the market may not let them reach. This is the reframe most owners miss: a startup client can love your work, praise it in every meeting, refer you internally, and still churn, because the churn decision isn't about you. It's about their burn rate. You can do everything right on delivery and still watch the revenue evaporate, because the failure mode is upstream of your service entirely. We dig into why "great work" is a weak retention moat in the discovery and qualification work that decides who you let onto the book; the short version is that who you sign determines your churn long before how well you serve them does.
Section 3
The numbers behind the mortality rate
It would be convenient to dismiss 2024 as a one-off correction. The data doesn't allow it. Carta, the cap-table platform that quietly sees the financial vitals of a large slice of U.S. venture-backed companies, recorded 966 startup shutdowns in 2024, up from 769 in 2023, a 25.6% increase and a record for the dataset . And Carta is explicit that this undercounts reality: it only sees companies that use Carta, so it is "missing a good chunk" of the closures happening elsewhere . Before you write the whole cohort off as a bubble unwinding, sit with the nuance the data's own author insists on. As Peter Walker, Carta's head of insights, put it: "Yes, shutdowns increased from 2023 to 2024 in every stage. But there were more companies funded (with bigger rounds) in 2020 and 2021" . That is not reassurance. It means a large, over-funded cohort is still working through its runway, and the shutdowns are the lagging echo of money raised years ago. For a service firm signing startups today, the implication is blunt: a meaningful share of your prospective startup clients are operating on borrowed time they themselves can't fully see. Now the root cause, which matters because it tells you the mortality is structural, not cyclical. CB Insights found that "ran out of capital" was the cause of death in 70% of failures, but flagged it as "the final cause of death, not the root problem" . The deeper driver, present in 43% of the failures, was poor product-market fit . That distinction is everything for your purposes. A company that runs out of money because it never found product-market fit isn't going to be saved by one more raise. It is structurally unstable, and a structurally unstable company is structurally unstable revenue for you. You are attaching your recurring income to an entity that, four times out of ten, never figured out whether it had a business at all. And this isn't only a venture phenomenon you can sidestep by signing slightly later-stage companies. Zoom out to the whole economy and the curve stays grim: Bureau of Labor Statistics business-survival data shows roughly 50% of new businesses fail by year five and 70% by year ten . "Young company" is a risk category whether or not there's a venture round attached. The 90% founder-failure framing that spooks salespeople isn't hype, it's the long-run base rate of betting your book on the unproven.
Section 4
The trap compounds: mortality times concentration
Here is where a startup-heavy book stops being merely risky and becomes structurally dangerous. The two failure modes multiply. The first is the mortality you've just seen. The second is concentration, and chasing prestige logos quietly drives it up. Big, well-funded startups write big checks. They're the accounts that balloon past the others. So the same instinct that loads your book with high-mortality clients also tends to make one of those high-mortality clients your largest line item. You've concentrated your revenue in precisely the accounts least likely to survive. The thresholds here are not aggressive. For a service business, any single client at 10% or more of revenue already meets the standard definition of client concentration risk, and the classic trigger event is mundane: a new CFO arrives, does a vendor review, and your retainer is suddenly on the table. With a startup, that "new CFO" might instead be a down round, a board mandate to extend runway, or a pivot. Same outcome, faster timeline. Push past 20% and you're in the zone buyers and prudent owners treat as a structural threat. For professional-services firms, a single client above 20–25% of revenue, or a top three above 50%, is a red flag that compresses what your firm is worth. The cleanest articulation of the stakes is this: "If your firm's EBITDA is $1M, but $400K of that is attributable to one client who could leave, the buyer isn't really buying $1M of earnings. They're buying $600K of earnings and a lottery ticket" . Layer the two facts together. In our 12-person studio, the three startup logos are 30% of revenue, and the fintech alone is about 12%, already over the concentration line on its own. The base rates say a meaningful fraction of that 30% may not renew, and the largest single slice sits in the most volatile account. This is not a book that churns one client at a time. It's a book engineered to lose a cluster of revenue in a single bad quarter, because the same macro conditions, tightening capital, cautious boards, hit all your startup clients simultaneously. Their failures are correlated. That's the part that turns a manageable problem into a cliff: you don't get to lose them one at a time and backfill calmly. They go together. And the cost isn't only this year's revenue. It's the enterprise value of your own firm. The exact profile a startup-heavy book creates, high per-client mortality plus high concentration plus correlated risk, is the profile that compresses a service business's valuation to that "lottery ticket" . You are not just risking next quarter's cash flow. You're capping what your company is worth the day you'd want to sell it.
Section 5
The BGA framework: The Retainer Half-Life
Stop optimizing for logo prestige. Start managing your book the way a bond-fund manager manages duration risk: by the expected lifespan of the revenue you hold. The framework is called the Retainer Half-Life, the idea that every client carries an expected lifespan, and your job is to know the half-life of your book and deliberately lengthen it. 1. Assign every client a half-life tier. Bucket your book into three: Short (high-mortality, early-stage startups, pre-product-market-fit, anything living off a runway clock), Medium (funded but established, profitable scale-ups, growing SMBs), and Long (cash-flowing, established buyers with their own customers and recurring revenue, the "boring" companies). Use the base rates as your anchor: treat an early-stage venture client's realistic horizon as the ~22-month median-to-death, not the contract term . The contract is what you hope for; the half-life is what you plan around. 2. Measure the weighted half-life of the whole book. Take each client's share of revenue and multiply it by its tier's expected lifespan, then sum. A book that's 40% Short, 30% Medium, 30% Long has a dangerously low weighted half-life even if it looks impressive on a logo wall. The number itself matters less than tracking whether it's rising or falling quarter over quarter. If your weighted half-life is shrinking, you're getting more fragile no matter how fast revenue is growing. 3. Overlay concentration on the Short tier specifically. No single Short-tier client should exceed the 10% concentration threshold , full stop, because that's the revenue most likely to vanish on someone else's timeline. Reserve any tolerance for higher concentration (and even then, stay well under the 20–25% danger line ) for Long-tier clients whose survival you can actually underwrite. Concentration in survivors is a calculated bet; concentration in the high-mortality tier is just exposure. 4. Set a book-mix target and enforce it at the point of sale. Decide the mix you want, a defensible default is a majority of revenue in Medium and Long tiers, with Short capped at perhaps 20–25% of the book as your "upside allocation." Then enforce it in qualification, not in hope. If you're already over your Short-tier cap, the next shiny startup isn't a yes-with-enthusiasm; it's a no, or a yes only at terms that price the risk. This is a qualification rule, which is why it lives upstream in how you score and gate demand rather than in delivery. 5. Price and structure for the half-life you're underwriting. For Short-tier clients you do take, change the deal shape: shorter commitments with explicit renewal gates, upfront or milestone-weighted payment to de-risk the runway problem, and a setup/onboarding fee that ensures you're profitable even on a client who churns at month eight. Don't sell a startup the same patient, low-deposit annual retainer you'd give a stable enterprise, you're carrying more risk, so the terms should carry more protection. The deal-structuring moves that make a fragile client survivable sit in building protection into the offer itself. 6. Backfill toward survivors on purpose. Lengthening your half-life is a demand-generation problem, not just a screening one. If your inbound is all startups, your marketing is fishing in the highest-mortality pond. Deliberately build a pipeline of established, unglamorous buyers, the regional manufacturer, the multi-location services business, the profitable B2B company nobody posts about. Run a quick honest read on your current exposure with the growth diagnostic, then retarget positioning and outbound toward the durable end of the market. Boring is a feature. A practical rule of thumb to operate this: review your weighted half-life and Short-tier concentration once a quarter, the same cadence you'd review cash. If either is moving the wrong way, the fix is in who you sign next, not in working harder to retain clients whose failure was never about your work.
Section 6
You're running The Retainer Half-Life right when…
You're running it right when you can name your book's weighted half-life off the top of your head, and it's been flat or rising for three straight quarters even as revenue grew. When no single high-mortality client sits above 10% of your revenue, and the accounts that do sit near your concentration ceiling are established buyers whose survival you'd bet your own money on. When a hot seed-stage logo lands in your inbox and your first move is to check your Short-tier cap rather than to celebrate, and you can say no, or say yes only at de-risked terms, without flinching. When your marketing is deliberately fishing in the boring, durable end of the market rather than the buzzy, fragile end. And when the question "what happens to us if our three best-known clients all fail in the same quarter?" has a calm answer instead of a sick feeling, because you engineered the book so it can't take you down with them.