The meeting that changed The Lifetime Value Co. was not a triumph. It was a demolition. In 2008, founders Josh Levy and Ross Cohen drove from New York to suburban Philadelphia to pitch Beyond.com on BeenVerified, their identity-verification engine. The software was meant to be a driver's license for the internet: prove who you were once, then carry that assurance into dating sites, job boards and online marketplaces. Prospects liked the idea. Nobody bought it.
Beyond finally supplied the missing sentence. A verified member would make every unverified member look suspect. Why would a community introduce a feature that could weaken the value of most of its own network? After 11 months, the founders had spent roughly $550,000 of their original backing. The company was burning about $50,000 a month and had about $50,000 left. The code worked. The incentives did not.
“These companies were never going to cannibalize their own communities.”Ross Cohen, recalling the original model
That objection turned out to be more valuable than another enthusiastic meeting. People did not especially want to verify themselves. They did want to investigate the stranger on the other side of a date, a sale, a phone call or a hire. BeenVerified moved toward the user with the question. What began as identity infrastructure became a consumer search destination, and the parent company eventually became a factory for building focused products on top of public and licensed data.
The company behind the search box
LTVCo is easy to mistake for a background-check business with side projects. A better description is an information-services platform with specialized storefronts. BeenVerified answers broad questions about people and records. NumberGuru and ReversePhone narrow the task to an unknown caller. NeighborWho and Ownerly wrap property records and home-value information around real-estate decisions. Bumper packages vehicle histories for car buyers and owners. PeopleSmart points contact data toward sales teams. MoneyBot5000 searches unclaimed-fund records.
The distinction matters because customers do not wake up wanting “public-data infrastructure.” They want to know who keeps calling, whether a used car has an accident in its past, what a house may be worth or how to reach a prospect. Each brand meets a recognizable moment with its own vocabulary and interface. Behind those front doors, LTVCo can reuse the expensive machinery: ingesting records, joining identities, maintaining search, testing conversion flows, acquiring customers and applying privacy and security controls.
What, exactly, did they do?
First, they abandoned a distribution strategy that required platforms to damage their own economics. BeenVerified went directly to consumers and made public-record search cheaper and faster than traditional investigation services. In 2008, the company advertised checks for small businesses from $10 for a single verification, while more extensive checks could cost far more. In 2009 it launched a consumer iPhone app. Early promotion was not shy: national television spots and even ads on garbage trucks put the service in front of people outside the usual enterprise-buying loop.
Second, they treated the successful product as a test bed. Cohen has described adding a small amount of vehicle-history data inside BeenVerified and watching how people used it. Demand helped validate what became Bumper. This is the useful part to copy: do not begin every adjacent idea with a new company, team and stack. Put a thin version where existing customers can reveal intent. Give it a separate brand only when the behavior earns one.
The repeatable move: question first, brand second
Third, they built around data operations rather than a single interface. An LTVCo engineering note described pulling from multiple public-record sources and evaluating Apache Airflow on Google Cloud because the incumbent workflow tool demanded engineering attention when it failed and made changes difficult to track. That is the unglamorous center of the business. Search quality depends on pipelines, provenance, normalization and refresh cycles long before a tidy report reaches a screen.
The money arrived after the machine worked
The early company took angel backing. Contemporary reporting described a $200,000 angel round followed by $600,000 from a hedge fund; Levy later said the company used a little more than $2 million in angel financing on its path to profitability. The consequential check arrived much later. In December 2020, Morgan Stanley Expansion Capital led a $150 million investment alongside other Morgan Stanley funds and ROCA Partners. It was LTVCo's first institutional fundraise.
The order is more interesting than the amount. By then the company had brands, subscribers, acquisition channels and a global team. The stated uses were concrete: richer data capabilities, more hiring, better user experiences, and additional marketing and partnership opportunities. Capital was poured into a machine with operating history, not used as a substitute for discovering why the first product should exist.
These figures come from company descriptions published at different moments and are not directly comparable. They may reflect portfolio changes, measurement definitions or audience shifts. The honest read is scale, not a smooth growth curve.
A portfolio with sharp edges
The direct alternatives change with the question. People search puts BeenVerified beside Spokeo, Whitepages, TruthFinder, Intelius and PeopleFinders. Bumper meets CARFAX and AutoCheck. Property products compete with public assessor sites and specialist valuation tools. PeopleSmart enters a crowded sales-data market that includes ZoomInfo, Apollo and Cognism. The free substitute is always lurking: a determined user with search engines, government sites and time.
LTVCo's answer is convenience. It collects scattered records, connects them and presents a report around a decision. That convenience also carries the central risk. Publicly obtainable does not mean context-free, perfectly current or harmless. People-search products can surface sensitive information; similar names can be confused; and consumer search reports are not automatically appropriate for employment, housing, credit or other decisions regulated by the Fair Credit Reporting Act. The product must explain permitted use, maintain correction and opt-out processes, protect data and avoid pretending that aggregation eliminates uncertainty.
The moat is not possession of a record. It is the ability to turn messy records into a useful answer before the customer gives up.
The portfolio approach also has a condition: the shared layer must genuinely be shared. If every new vertical needs unrelated data, separate compliance logic, different acquisition channels and a new engineering organization, the factory advantage disappears. A shelf of brands then becomes a shelf of costs. This playbook works when adjacent questions can reuse ingestion, identity resolution, search, billing, experimentation and distribution. It does not work merely because two products both contain data.
The remote company became the operating system
In February 2020, LTVCo asked employees to begin taking devices home in anticipation of an office closure. Six months into remote work, an internal survey found 97 percent of employees thought the company was using Slack well for communication. In 2021, it made the arrangement permanent. Today it describes itself as remote-first, with hubs in New York and Costa Rica.
The culture language is unusually direct: “Try, Fail, Learn, Repeat,” “Question Everything, Respectfully,” and “Teach Each Other to Fish.” There are birthday holidays, professional-development budgets and location-specific benefits. Remote rituals have included Yappy Hour, LTV Cribs and LTV Chefs, which sound less like corporate programming than cable channels invented by a very friendly dog. The company has collected workplace recognition, including Fortune lists in 2025 and repeated Great Place To Work certification.
No perk resolves the hard part of a distributed product organization: preserving dissent, knowledge transfer and fast feedback without allowing calendars to swallow the day. LTVCo's own story suggests a sensible standard. Tools matter, but the operating behavior matters more. Regular founder AMAs, documented workflows, interest-based channels and managers trained for remote teams make the model plausible. Free lunch delivery is pleasant. It is not an information architecture.
The thing worth stealing
The Lifetime Value Co. did not win by correctly predicting its first customer. It survived by noticing that the first customer had no reason to buy. The founders changed who performed the action, who received the value and how the product reached them. Later, they repeated a quieter version of that move: observe a narrow question, validate it inside an existing audience, then package it in a brand built for that moment.
Copy the sequence, not the category. Interview the buyer who says no without being polite. Draw the incentive map before polishing the integration. Test adjacent demand where traffic already exists. Share the difficult infrastructure, but keep the promise specific. Raise expansion money after the engine shows what it can do. And know where the method fails: weak data rights, unreliable records, regulated decisions, expensive acquisition or verticals with nothing meaningful in common.
The original dream was to create a badge people carried around the internet. The more durable business gave people a search box and let their curiosity lead. One asked the market to adopt a new social behavior. The other met a behavior already happening. That difference cost about $550,000 to learn. It has been paying tuition ever since.