"It's difficult to fail if you actually talk to customers."
—Paul Kudlow, TrendMD
As he was going through medical school the last few years, Paul Kudlow pictured his future: talking to patients all day, providing solutions, and working all hours. Now, as he leads a successful startup, he still provides solutions and works all hours—but it's customers not patients that he talks to all day.
"It was still a massive transformation," Kudlow told me last month. "I went through medical school and started residency—and then this came out of nowhere."
This is TrendMD, a kind of Outbrain or Taboola for the medical world. TrendMD enhances content discoverability for readers by providing publishers with strong incentives to display relevant links to third-party content. They took that model and designed an article recommendation widget that's embedded in places doctors and other researchers use. Content producers can also promote their links on sites where the TrendMD widget sits.
TrendMD is one of the featured Models of Excellence at SIIA's Business Information & Media Summit (BIMS), Nov. 14-16 in Fort Lauderdale, where Kudlow will be speaking.
For Kudlow, it started off as an academic and research project and grew into something much bigger. The underpinning came from the frustration of not being able to keep up with the latest literature and research. He thought, "What's the best way to get research to the people who can make use of it?"
So far, so good. "After launching in May 2014, we're up to over 2500 medical journals and content sites, and we're in use by about 260 publishers," Kudlow said. "The recommendations themselves are seen by about 40 million unique readers a month, and we generate almost 400 million recommended articles a month."
Was that 400 million? "Yes, each widget displays anywhere between 6-10 recommended links. And there's around 90 million widget loads or views a month."
Thus it's no mystery why Kudlow is putting off being a practicing doctor. "I started loving [the work on TrendMD]," he said. "I have the option to finish my residency in the future. But that may be unlikely."
Interesting revenue model
"We don't engage directly with doctors," Kudlow said. "A doctor may go to a site that he or she knows like BMJ or Medscape. They'll be reading an article, and at the bottom of the page they'll see us, "recommended by TrendMD." There will be articles on that site and links to external articles. TrendMD makes money when doctors click on those external links [on a cost-per-click basis]." (Kudlow prefers the term "external links" to "sponsored content.")
"We try to live on places where doctors go," he said. "And provide value there. Doctors and researchers are stubborn [so] we don't want to change their habits. [They tend to] go to their journals and that's it. They rely on search engines like Pubmed and Google Scholar, but the problem with search engines is that you need to know what you're looking for. Our discovery engine recommends based on your previous reading habits."
The model can get a little complicated. Each click on a link that directs a reader outside the publisher's journals earns that publisher 0.5 traffic credits. Traffic credits can either be reinvested to promote a publisher's own content and attract new visitors (one credit per visitor) or be redeemed for a monetary value of $1 per credit to increase revenue. In general, publishers earn 7–10 traffic credits per 1,000 loads of TrendMD on their sites.
In an effort to maximize the number of visitors they receive, many publishers in the TrendMD network maintain their credit balances by purchasing weekly subscriptions at $1 per traffic credit. Publishers can opt to display only recommendations from their own sites, or they can opt in to the TrendMD network and include recommendations to articles from other publishers.
The TrendMD staff stands at 14 now, with 12 "full-time-ish." They have offices in Toronto and San Francisco—and some virtual employees. "Scaling an organization of this size is completely different," he said. "It used to be me and one other person toiling in the field and talking to customers. We got good at that. But now it's a growing team. Every stage of the game brings new challenges that come up, exciting but difficult. We're growing the sales team now.
"What's really interesting is that, as a doctor, you may not know what treatment options are out there. Unless you search for it you're never going to find it. Traditionally they find it through word of mouth. We're quite selective about who gets into the network."
No entrepreneur playbook
TrendMD uses many factors, in addition to relatedness, to determine the best links for each reader. These include "collaborative filtering (similar to Amazon's 'people who bought this item also bought that item') and user clickstream analysis (the Netflix approach, basing recommendations on the users' interests expressed through their online history).
"Unlike medicine, there's no playbook for startups," Kudlow said. "You kind of invent it as you go and see what fits. We offer value to readers. We're distributing content so it can get to the readers and give value to the authors."
Before TrendMD, he said there was no way to push this kind of content to readers. "Often we heard journals say that they get new readers through good SEO or posting content online," Kudlow said. "That's a bit like saying we printed the journal out it in one library hoping that people can read it."
Though both involve working long hours, Kudlow see big differences in his career choice.
"I say to people, in medicine the whole goal is learning rules and following rules. If you break rules in medicine, [the result can be tragic]," Kudlow said. "With entrepreneurialism and startups, it's the opposite mentality. You need to challenge the status quo. If you don't, you can't break into [your] market and you can't change things. Fundamentally being a good doctor and being a good entrepreneur are very different."