返回资讯中心
外部精选
软件工程
#开发者实时流

From the Analytics Department to a Lone Operator with a Laptop

For most of the twentieth century, the path of a business idea from conception to market ran through a complex institutional infrastructure: trend-scouting departments, business-model generators, feasibility analysts. T…

Hacker News 最新Aldus5512026/10/31 · 17 天前30 分钟阅读

以下正文同步自 Hacker News 最新,版权归原站所有,已转换为易读排版。

For most of the twentieth century, the path of a business idea from conception to market ran through a complex institutional infrastructure: trend-scouting departments, business-model generators, feasibility analysts. This system was slow, expensive, and — more importantly — it determined the very structure of entrepreneurship itself: access to launching new ventures belonged only to those who could pay for the entire cycle out of their own pocket. Over the past decade and a half, this construction has been almost entirely dismantled. However, as often happens with technological shifts, the liberation of one participant in the process gave rise to new asymmetries — less visible, but no less severe. This text attempts to trace both sides of the transformation.

Before the era of widespread digitalization, creating a new business resembled an assembly line with a clear division of labor.

The first link was search and generation. This was handled by dedicated people: business scouts, industry analysts, often former engineers or marketers. Their toolkit was “physical”: trade shows, industry journals, patent bulletins, in-depth interviews with customers, borrowing models from more developed markets and adapting them to local conditions. The result of their work was a long list of dozens, sometimes hundreds, of hypotheses.

The second link was feasibility analysis. The long list was handed over to people of a different profession, whose task was to weed out the unsuitable ones. Their product was the feasibility study, a technical and economic justification running 40 to 80 pages: market size and dynamics, a portrait of the paying customer, technical feasibility, required investment and break-even point, legal and patent risks, SWOT analysis.

The third link was the decision. Based on the report, the owner of capital decided whether to launch the project. The cycle took two to four months and cost tens of thousands of dollars, which meant that only one idea out of twenty or thirty ever made it to launch.

It is telling that intermediary businesses operating at the junction of “idea and capital” existed almost from the very beginning of the industrial age. Arthur D. Little, founded in 1886 and considered the world’s first consulting firm, worked in exactly this way: an inventor would arrive with a technical idea (say, a new method of producing cellulose), the firm would assess its applicability, the savings it offered, and the cost of implementation — and would then take that report to industrial giants such as DuPont or General Electric. This function was later institutionalized by McKinsey & Company (from 1926) and the Boston Consulting Group (from 1963), but now from the buyer’s side: the owner of capital hired consultants to find and package investment directions for him. A classic example is the story of Pampers diapers: Procter & Gamble engineer Victor Mills came up with a disposable liner for his own grandson, but turning the invention into a mass-market product required an enormous amount of analytical work to translate the idea of a “convenient thing” into the language of the market, cost of production, and shelf space.

A parallel mechanism existed in the Soviet system: VOIR (the All-Union Society of Inventors and Rationalizers) and the factory-level bureaus of rationalization and invention (BRIZ). A worker or engineer would submit a rationalization proposal, but the idea would only move forward once the economic effect had been calculated — the same feasibility filter, only within a planned economy.

Finally, in the 2000s, attempts emerged to institutionalize the “market for ideas” itself: platforms such as Quirky and Edison Nation accepted ideas from any author, carried out the entire analysis on their behalf (patents, cost of production, distribution), and brought the successful ones to market. The fate of Quirky itself is telling: the company went bankrupt in 2015, demonstrating that mediating between a mass of ideas and the market is, in economic terms, an extremely difficult business to run.

The old model had a characteristic feature that is often overlooked. Consultants and analysts sold time and reports, but they did not invest in the outcome. The fee was received at the moment the report was delivered; if the idea failed in the market, the loss fell exclusively on the owner of the money.

From this followed a rigid rule of entry: only people with capital or access to capital entered business, and moreover with a reserve sufficient for two or three attempts. The logic was simple: whether an idea was viable could only be understood at the third stage of expenditure (rent, personnel, goods, development), by which point the money from the first two stages had already been spent. An idea without money was worth zero — not because it was bad, but because no mechanism existed for testing it cheaply.

Out of this grew a social division into two worlds that almost never intersected with one another. The world of ideas consisted of engineers, inventors, and observant enthusiasts who published their concepts on forums, in specialized magazines, in letters sent to companies. The world of money consisted of entrepreneurs and owners of capital, who read not forums but the reports produced within their own circle. Someone else’s idea held no value for them, because it lacked “packaging” — the figures that established its feasibility. The consultant-intermediaries were, in essence, translators from the language of “I came up with a thing” into the language of “here is the market, here is the margin, here are the risks.” This translation was expensive, which is why the overwhelming majority of ideas — by various estimates, up to 90% of potentially patentable developments — simply remained on paper.

The shift took place roughly between 2008 and 2012 and was caused not by a single factor but by the layering of several: cheap cloud computing, website builders and payment systems, social networks as a channel of nearly free traffic, crowdfunding as a tool for pre-sales, and also the lean-startup methodology, which legitimized the idea of “testing demand before building the product.” The cycle that had once taken months and cost $20,000–50,000 before the first test contracted to one or two days and a few hundred dollars.

The canonical examples made it into the textbooks: in 2008, Dropbox tested demand not with a product but with a three-minute demonstration video, after which the waiting list grew from 5,000 to 75,000 people overnight; Zappos began with Nick Swinmurn photographing shoes in ordinary stores and posting them online, purchasing the merchandise only after an order had already been paid for. Neither company needed an 80-page feasibility study — it was replaced by an experiment costing a few hundred dollars.

The sources of ideas changed as well. In place of trade shows and industry journals came the systematic analysis of public data: reviews on marketplaces and in app stores (a concentrated expression of unmet demand), technological shifts that had made something dozens of times cheaper (speech synthesis, sensors, satellite imagery), and the underused assets of the entrepreneur’s own life. The analytical function contracted from a full report into a short filter: is there an acute pain point, does the market pay for it, can demand be tested within a week, does the author have any protection against copying, does the unit economics work out, what are the regulatory risks, what is the ceiling on scale. With the arrival of generative AI, this combination of “person plus filter plus data” became accessible without any team at all.

The cheapening of testing, however, did not eliminate the need for capital — it merely shifted the point at which capital enters. And this is where the first new asymmetry arises.

The founder now personally performs the entire body of work that was previously distributed across three departments: finding the idea, carrying out the analysis, and packaging it into a working business with its first sales — using their own money and their own time. The early-stage investor steps into a construction that is already finished. The paradox is that the valuation of this construction remains low: what is valued, legally and economically, is neither the idea nor the labor, but the fact that “the author reached sales on their own.” That is why deals in which the founder gives up 50–75% of the company for comparatively modest sums are not an exotic exception but standard practice for angel and pre-seed rounds, especially outside the top venture hubs. An investor working with a funnel of hundreds of such teams dictates the terms; a founder without money for scaling accepts them.

The second asymmetry reveals itself in a crisis. In the old model, failure meant a loss of money but not of status: the owner remained the owner. In the new model, ownership itself becomes the vulnerable asset. The standard instruments of venture contracts — anti-dilution protection, rights of first refusal, bridge loans secured against the founder’s stake, liquidation preferences — are constructed in such a way that at the moment when a company urgently needs money and has none of its own, the valuation collapses down to the size of the cash on hand, and the founder’s stake is bought out for a nominal sum. The situation of “sell your stake for pennies or lose everything” is not hypothetical: one need only recall WeWork, where SoftBank’s emergency rescue in 2019 was carried out at a valuation of roughly $8 billion against a peak of $47 billion just months earlier, or Pebble, the smartwatch pioneer with record-setting crowdfunding totals, which, after running into funding trouble, was sold to Fitbit for a price that did not even cover its debts. In both cases, it was a liquidity crisis, not the absence of a product, that determined the final distribution of ownership.

From this follow two practical rules that today count as basic founder literacy: never give up more than 10–15% for the first money raised, and keep a liquidity cushion sufficient for at least six months of operation without external infusions.

There is also a third, ideological layer to the problem, and it is an interesting one. The formula “you have a bold idea? Turn it into a startup and prove that it’s worth it” sounds like motivation, but functionally it is a transfer of risk. The phrase substitutes a test of endurance for a test of feasibility: analysis is replaced by a year of the author’s life and their personal savings. If the project burns out, the conclusion drawn is that the idea was bad — although more often it is not the idea that burns out but the scheme used to finance it.

Who benefits from this arrangement becomes clear from the flows involved. Investors get pre-validated businesses at a minimal price. Platforms and ecosystems get millions of free “testers” who filter through thousands of hypotheses at their own expense; the handful that survive are then bought up. Even recommendation systems and language models reproduce precisely this startup narrative, because they are trained on a corpus in which stories of “how I raised a startup” outnumber stories of “how I sold a license and now live off royalties” by orders of magnitude.

At the same time, for genuinely new ideas, the startup model is not the only path, and often not even the best one. Historically workable alternatives include: licensing and royalties instead of one’s own production; launching as a service running on someone else’s equipment and infrastructure; and the direct sale of the technology to the R&D division of a large company. This, incidentally, is exactly what Robert Kearns did in his time, the inventor of the intermittent windshield wiper: instead of building a factory, he tried to license the patent to automakers (and then spent decades, ultimately successfully, defending his rights in court against Ford and Chrysler). A d

正文由 探How 从来源站点 RSS 同步,内容未经改写;遇到排版缺失或需要图片、视频时请以原文为准。

打开原文
去社区讨论 →