Still Pitching Yesterday’s Business Model
Somewhere right now, a founder is pitching a subscription box company to an investor. The deck is clean, the
Somewhere right now, a founder is pitching a subscription box company to an investor. The deck is clean, the market-size slide is generous, and the unit economics, if you squint hard enough at the projections, almost make sense. Nobody in the room mentions that Blue Apron sold itself for $103 million after being valued at $1.9 billion, or that HelloFresh has been losing subscribers for two years running. The meeting ends with handshakes and a follow-up request.
This is how outdated business models stay alive: not through merit, but through optimism, pattern-matching, and a collective reluctance to say the obvious thing out loud. The models that powered the last decade of startup culture are running out of road, undermined by AI, market saturation, shifting investor appetite, and the basic math of what it costs to acquire a customer who doesn’t actually stick around.
Per-Seat SaaS Pricing
The SaaS pricing model that minted a generation of billionaires ran on beautiful, linear logic: more employees meant more seats, which meant more recurring revenue. For twenty years, it held. Then AI showed up and broke the equation.
In January 2026, Monday.com CEO Eran Zinman announced the company had replaced its entire 100-person sales development team with AI agents. Response times dropped from 24 hours to 3 minutes. Conversions went up. The implication for the per-seat model was brutal: if 10 AI agents can do the work of 100 reps, you don’t need 100 Salesforce seats. You need 10. That’s a 90% revenue compression for every vendor whose business depends on headcount.
The market noticed. In February 2026, what analysts dubbed the “SaaSpocalypse” wiped $285 billion from software stocks in 48 hours. Atlassian reported its first-ever decline in enterprise seat counts. Apollo Global Management cut its software exposure in private credit funds from 20% to 10%. Gartner predicts that by 2030, at least 40% of enterprise SaaS spending will have migrated to usage-based, agent-based, or outcome-based pricing models.
The per-seat model trained an entire generation of SaaS companies to chase adoption metrics that no longer mean anything. User logins. Seats activated. Expansion revenue. Those signals once indicated health. Now they measure inertia. Companies still clinging to pure per-seat pricing aren’t just running an outdated business model. They’re building a product that actively erodes its own revenue as it improves.
The Big Agency Holding Company
On November 26, 2025, Omnicom completed its $13.5 billion acquisition of Interpublic Group and immediately announced what followed: 4,000 layoffs and the retirement of three of advertising’s most storied names. DDB, founded in 1949 by Bill Bernbach. FCB, with roots going back to 1873. MullenLowe. Gone, absorbed into the surviving trio of BBDO, TBWA, and McCann. More than 10,000 advertising careers were disrupted across both groups combined.
The traditional holding-company model, built on layers of creative departments, media buying relationships, and billable hours, is structurally incompatible with a world where AI generates creative variants in minutes and performance data replaces gut instinct. Omnicom is explicitly repositioning itself as an AI-led “Connected Capabilities” platform, chasing at least $750 million in annual cost savings. The heritage, apparently, will not stand in the way of the spreadsheet.
The agencies that survive aren’t selling creativity as a service anymore. They’re selling measurable outcomes, and that changes everything about how they staff, price, and pitch. The multi-layered holding company model with its duplicated functions, expensive real estate, and Byzantine approval chains was always optimised for scale in an era of cheap talent and steady retainers. Neither of those conditions holds anymore.
The Subscription Box
The subscription box had its moment. Birchbox launched in 2010 and looked like the future of retail. Then came the imitators, the saturation, and the brutal economics of shipping physical goods to customers who signed up for a discount and quietly cancelled within 90 days. Industry data shows 44% of all subscription box cancellations happen within that first three-month window. That’s not a retention problem. That’s a business model problem.
Blue Apron is the cautionary tale that keeps getting ignored. Once valued at $1.9 billion, it sold itself to Wonder Group for $103 million after offloading its operational infrastructure to FreshRealm for $50 million and cutting large swaths of its workforce. HelloFresh, which controls roughly 69% of the US meal kit market, has watched its subscriber numbers fall while its marketing costs climb. The company warned investors its earnings for 2024 and 2025 would fall below expectations, citing a “very different operating environment.” When the category leader is struggling that badly, the model deserves scrutiny, not imitation.
None of this has stopped new founders from launching subscription boxes. The market research reports are still bullish, because aggregate market-size projections paper over individual unit economics. Building a subscription box today means competing for customers who are already fatigued, paying customer acquisition costs that rarely balance against a 90-day churn window, and shipping cardboard in an era when sustainability pressure is mounting. The outdated business model persists not because it works, but because it looks like it should.
The SEO Lead Generation Agency
For about a decade, a certain type of agency had a clean, scalable pitch: build a website in a high-intent vertical, plaster it with optimised content, rank on Google, sell the leads to businesses who couldn’t be bothered to rank themselves. Personal injury lawyers. Mortgage brokers. Plumbers. Home insurance. The leads flowed, the margins were fat, and the model required almost no proprietary advantage beyond understanding how Google’s algorithm worked.
Google’s AI Overviews are dismantling this. When an AI summary appears at the top of a search results page, click-through rates for organic results drop by nearly half. A Pew Research study tracking 68,000 real search queries found that users clicked a result 15% of the time when no AI summary was present, and 8% when one was. Business Insider lost 55% of its organic search traffic between 2022 and 2025. HuffPost lost half its search referrals over the same period. Music blog Stereogum lost 70% of its ad revenue in 2025, directly attributing the damage to AI Overviews. Zero-click searches, where Google answers the query without sending the user anywhere, rose from 56% to 69% of all searches between May 2024 and May 2025.
Publishers expect search traffic to fall another 43% within three years, according to a survey reported by Search Engine Land, with one in five respondents expecting losses above 75%. The lead generation agency sitting between a Google search and a local business was always a layer of value extraction rather than a layer of value creation. AI is simply making that layer unnecessary. The founders still building SEO-driven lead gen businesses in 2026 are constructing an income stream on infrastructure they don’t own, optimising for a platform that is actively working to cut them out.
Unstructured Freemium
Freemium was the growth hack that defined the 2010s. Give the product away, build a massive user base, convert a small percentage to paid, and let the math work itself out. Dropbox did it. Slack did it. A generation of founders watched those success stories and copied the playbook without reading the footnotes.
The footnotes are brutal. Dropbox’s conversion rate hovers around 2.5% of its enormous user base. Most SaaS operators generate no meaningful annual contract value from freemium leads. Free users drive real costs: servers, support, infrastructure. And in 2025, with investors demanding efficiency over volume and unit economics over vanity metrics, a dashboard full of free signups carries no weight in a funding conversation.
What doesn’t work is reflexive freemium, launched because the founders used free tiers themselves and assumed their customers would behave the same way. Spotify’s model works because it’s designed around habit formation and specific friction points calibrated to nudge users toward payment. Zoom’s 40-minute limit was the result of actual research into meeting duration. Most founders don’t do that work. They launch a free tier to get signups, watch the conversion rate sit at 1-2%, add more free features hoping to boost engagement, and burn cash without building a revenue model. When a company removes its free plan, it typically discovers the thing it feared most, a drop in signups, is far less damaging than what it didn’t notice: it was subsidising an army of users who were never going to pay.
Hourly Agency Billing
The agency model built on hourly billing has a fundamental misalignment baked in from the start: the client wants results, the agency gets paid for time. The incentive is to bill more hours, not to solve the problem faster. For decades this tension was managed rather than resolved. Now clients are refusing to manage it.
Clients are demanding outcome-based pricing. Agencies are being pushed to package services into defined deliverables with measurable results, rather than submitting timesheets and hoping the client doesn’t ask hard questions. Hourly billing is one of the most persistent outdated business models in professional services precisely because it’s so easy to defend internally — until the client stops renewing. The agencies resisting this shift aren’t just running an outdated pricing model. They’re losing pitches to competitors who’ve already made the transition, and they’re losing talent to AI tools that can do in 20 minutes what used to justify a full day’s billing.
The hours-and-retainer structure made sense in a world where agency work was slow, specialised, and genuinely hard to benchmark. None of those things are true anymore.
Raising Big to Figure It Out Later
Not a single business model, but a philosophy that shaped how an entire category of companies was built. Raise big, spend fast, prioritise user acquisition over unit economics, and trust that revenue will follow scale. It worked for a narrow window between 2015 and 2021, when capital was cheap, public markets rewarded growth stories, and the exit environment was forgiving enough to paper over almost any structural problem.
That window closed hard. Only 11% of startups that raised between 2020 and mid-2025 made it to Series A, and the average time to get there stretched from 1.5 years to 2.5 years. The venture funding environment of 2026 demands profitability paths, disciplined burn rates, and measurable outcomes in quarters, not years. Founders still pitching “we’ll figure out monetisation once we have scale” are finding that the investors who used to fund that conversation have already moved on.
The companies that burned through capital chasing growth without ever building a real business model didn’t just run out of money. They ran out of the market conditions that made their pitch make sense.
The founders still on these foundations usually know something is wrong. The numbers don’t quite work. The churn is higher than projected. Acquisition costs keep creeping up. But changing a business model midstream is terrifying, and copying what worked before is comfortable. Most will keep building until a competitor, an investor, or a quarterly review forces the conversation they’ve been avoiding.
