The first clue that Krush Media is an unusual advertising company is that its origin story begins with a camera. ooVoo, the video-calling app, connected millions of people before video chat became a household utility. Each conversation consumed bandwidth. Each free user generated a bill. Advertising was supposed to close the gap. Eventually, the arithmetic stopped cooperating: according to the studio that later redesigned Krush, streaming costs outweighed what the service could monetize.
That is the kind of failure that can look terminal from outside and oddly clarifying from inside. Jeff Chi had spent years selling and monetizing digital media. Brian Faust worked the corporate side. They bought the company and concentrated on the capability hiding beneath the app - a system for matching advertiser demand with publisher supply. The consumer product receded. In 2017, Krush Media emerged as a programmatic advertising business.
The middleman whose product is fewer middlemen
Programmatic advertising turns a simple act - placing an ad beside a piece of content - into a millisecond auction. A publisher announces an available impression. Buyers evaluate the device, context, audience and price. A bid wins. The ad appears. Between announcement and appearance, however, a long chain of exchanges and resellers can obscure where the inventory began, inflate the price and create openings for fraud.
Krush lives in that chain, but its proposition is to make the chain shorter and more legible. For advertisers and agencies, it offers direct publisher inventory, custom private-marketplace deals and targeting across desktop, mobile, connected television and digital out-of-home. For publishers, it brings advertiser demand, manages auctions and fill rates, and supplies the ad-operations labor that keeps campaigns moving. It is exchange, service desk and customs inspector in one.
A shorter, checked supply path
Its customers therefore arrive with opposite anxieties. A buyer worries about paying for a bot, appearing beside unsafe content or bidding through unnecessary resellers. A publisher worries that an impression will go unsold or sell too cheaply. Krush must persuade each side that helping the other is not a concession. More buyers create competition for the publisher. Cleaner publisher inventory gives buyers more reason to bid.
The clever machine is only half the product. The other half is a person willing to investigate why good inventory looks bad.Krush's operating logic, in plain English
A plain-text file can stop a television auction
The most revealing work at Krush is almost comically unglamorous. A publisher declares its authorized advertising sellers in files called ads.txt or app-ads.txt. Exchanges identify their selling relationships in sellers.json. Another record, the SupplyChain object, shows the hops an impression took. If those declarations disagree, a legitimate stream can look suspicious. Premium buyers may simply decline to bid.
Steven Slaughter, Krush's head of ad operations and yield, described finding exactly this problem in connected TV and mobile apps. Buyers lacked the time to separate errors from fraud. His team, aided by Pixalate's ratings and analysis, helped partners correct integrations, false invalid-traffic signals and broken supply paths. Krush also lists HUMAN, DoubleVerify and Pixalate as verification integrations, using checks before and after bids.
This is where Krush differs from a giant self-service platform. Its technology shapes traffic, enriches requests with data and conducts auctions, but the company repeatedly sells human intervention: campaign trafficking, yield decisions, integration repair and custom deal design. Its public team page lists specialists in publisher accounts, ad operations, sales, product and finance. Inventory teams in the United States, Ukraine and India extend the watch across time zones and holidays.
Giving up a little margin to make the market move
The business earns from the transactions and managed services flowing through its marketplace. One choice is particularly instructive. Krush says it operates on lower-than-industry-average margins so buyers encounter more competitive prices. Better pricing can increase bid rates. More bids can lift fill rates and revenue per thousand impressions for publishers. The company forgoes a larger slice of one transaction in hopes of making the whole marketplace more active.
That bargain works when Krush has useful direct supply, enough differentiated demand and the operational discipline to keep quality high. It is less compelling for a buyer who wants only the broadest possible self-service reach, or for a publisher whose scale and technical staff already justify building an equivalent demand stack in-house. A managed middle layer must keep proving that the revenue it adds or waste it removes exceeds its fee.
The cannabis test
Cannabis advertising makes the ordinary frictions of digital media harder. Rules differ by jurisdiction. Mainstream platforms restrict the category. A campaign needs both compliant creative and inventory willing to carry it. In late 2019, Krush noticed increasing demand from cannabis and CBD brands. The problem was not a shortage of desire to advertise; it was a shortage of vetted, scalable places where the ads could safely run.
In 2020, the company added a cannabis/CBD supply-side offering within Krush X and partnered with CannaVu, part of Trion.io. CannaVu brought specialized advertiser demand. Krush brought direct publisher relationships and the vetting process. The result was a curated lane across web, mobile and connected screens. It is a neat example of the company's broader method: do not pretend the messy constraint has disappeared; make the constraint visible and build the market around it.
The part worth copying
When a product fails, list the capabilities it forced the company to build before discarding the whole thing. ooVoo needed audience data, advertising sales, campaign operations and monetization. Krush turned those backstage functions into the main act. The move is repeatable only when the hidden capability solves a real problem for customers beyond the original product - and when a team has enough credibility to sell it on its own.
The second act is quieter
ooVoo was visible by design. Its users saw one another. Krush is useful partly because most people never see it. The exchange reports handling 14 billion ad requests a day, yet its work is measured in absences: the bot that was not bought, the duplicate reseller removed, the television stream whose corrected file finally attracted a bid.
For brands, that means a route to audiences with more control over provenance and placement. For publishers, it means another source of demand plus people watching the pipes. For Krush, it means occupying a difficult place in the market - the middle - and arguing that a well-run middle can be valuable precisely because it is smaller, checked and accountable. The video call did not survive. The economics lesson did.