The last few decades have produced many successful marketplaces. We went from goods marketplace pioneers such as eBay and Amazon to simple service marketplaces such as Uber, Lyft, Doordash, Upwork, Thumbtack, TaskRabbit, and Fiverr. But why haven’t we seen many successful B2B service marketplaces?
Some would argue that companies such as Upwork, Thumbtack, Fiverr, or TaskRabbit are horizontal B2B marketplaces in the sense that they provide access to suppliers of different services. But while businesses do indeed transact with freelancers on such “horizontal” marketplaces, for most service verticals these are limited-value, one-off transactions. They fail to enable long-term business collaborations.
Horizontal services marketplaces don’t provide much value beyond matching clients with quality service providers. In other words, they don’t facilitate collaboration between buyers and suppliers, never mind provide ways for the two parties to collaborate more efficiently over time as they engage in follow-on projects.
In essence, the model these marketplaces were built around is not much different from the likes of Craigslist, which put a convenient UX on traditional classified advertisements.
In their article “What’s Next for Marketplace Startups?,” Andrew Chen and Li Jin found that there aren’t many successful service marketplaces because those offerings are complex, diverse, and difficult to evaluate. It’s challenging to define a successful transaction in a service marketplace because it’s harder to quantify success.
One reason is that several service providers must often work together to complete a single job for a buyer, requiring a complex workflow from end to end. As a result, it’s difficult for marketplaces to not only mediate service delivery but also make it significantly more efficient for buyers and suppliers. If both the buyer and suppliers don’t see a significant efficiency gain other than being initially matched, why would they continue using the marketplace?
(Image via Getty Images / Lidiia Moor)
The $50 billion translation industry is a prime example of complex B2B services marketplaces. On the supply side are roughly 50,000 small agencies around the globe responsible for more than 85% of this $50 billion industry. (Note we are referring to agencies here as suppliers, though they play on both sides.)
On the demand side are businesses that need to translate text from one language into another. Plus about 1,500,000 freelance linguists work in this industry, many of whom are more specialized than professionals in other industries.
Anyone can find and hire a translator on Fiverr or Upwork. Both provide a vast selection of language translators. However, the quality and cost of the translation depends on the translation tools available to the translator as well as their subject expertise.
Neither Fiverr nor Upwork provide computer-aided translation (CAT) and collaborative workflow solutions for users of their platforms. Additionally, neither provides an effective way for all parties to collaborate and continuously improve the efficiency and quality.
But the problem with traditional marketplaces goes even further: Multiple translators and reviewers are usually needed to complete a single job for a customer. Multi-language translation projects are even more complicated. Such projects require multiple service providers and cost estimates, in addition to project management tools.
This is why building a B2B service marketplace is difficult. Service marketplaces must not only connect buyers and suppliers, but also provide tools to enable an efficient and collaborative workflow that reduces wasted time and effort.
In addition to the problems already outlined, traditional marketplaces experience another issue that prevents them from growing and retaining market participants: Buyer and supplier attrition.
Many business services are based on regularly recurring engagements. In some cases, a buyer and a service provider interact daily, requiring a different workflow than gig-marketplaces are built around.
Buyers and suppliers have little motivation to continue interacting on a platform with no workflow automation solutions. They lack a way to improve service efficiency and quality, automate collaboration, payment, paperwork, and other basic processes required for a business.
This is why many traditional marketplaces suffer from slow network effects and high attrition. (A network effect is what happens when a platform, product, or service delivers more value the more it is used.
Think Facebook, eBay, WhatsApp.) Why wouldn’t companies work directly with service providers outside of a marketplace after they were introduced? What incentives keep the service transaction on the marketplace? These are critical questions to answer when building a marketplace.
Traditional marketplaces target broad services, making it nearly impossible to provide workflow solutions for buyers and suppliers. Going forward, successful service marketplaces will be developed relying on an industry-specific SaaS workflow. This will focus buyers and suppliers on longer-term projects and interactions that serve the unique needs of collaborations and transactions in a specific vertical.
Image via Getty Images / OstapenkoOlena
In “The next 10 Years Will Be About Market Networks,” James Currier, Managing Partner at NFX Ventures, defines a new era of service marketplaces, which he calls market networks.
A market network is a platform that combines elements of an n-sided marketplace, a network, and workflow solutions. An n-sided marketplace is one that requires coordination of multiple supply-side parties to provide a complex service for a single buyer.
Market networks enable multiple buyers and suppliers to interact, collaborate, and transact on the same platform. They provide users with industry-specific workflow solutions that enable efficient, ongoing collaboration on long-term projects. This reduces costs and leads to a higher quality of services and increased overall value for all users.
But how do you actually build a successful market-network platform? While the answer to that varies from company to company, here is our approach. We were able to build a market network for the translation industry that combines the components: network, marketplace, and workflow solution.
The first step to building an effective complex market network is to develop a workflow that is easy for users to embrace. It might not seem like much, but this increases productivity by enabling teams to perform tasks that were previously impossible.
Trash is a new startup promising to make it easier for anyone to create well-edited videos.
Social video is an area that CEO Hannah Donovan knows well, having previously served as general manager at Vine (the video app that Twitter acquired and eventually shut down). She said that in user research, even though people had “really powerful cameras in their pockets,” when it came to editing their footage together, they’d always say, “Oh, I’m not technical enough, I’m not smart enough.”
Donovan, who also worked as head of creative at Last.fm, said she “got curious about whether we could use computer vision to analzye the video and synthesize it into a sequence.”
The result is the Trash app, which comes with a straightforward tag line: “You shoot, we edit.”
Donovan demonstrated the app for me last week, shooting a few brief clips around the TechCrunch New York office, which were then assembled into a video — not exactly an amazing video but much, much better than anything I could have done with the footage. We also got to tweak the video by adjusting the music, the speed or the “vibe,” then post it on Trash and other social networks.
Donovan founded the company with its Chief Scientist Genevieve Patterson, who has a Ph.D. from Brown and also did postdoctoral work with Microsoft Research.
Patterson told me that Trash’s technology covers two broad categories. First there’s analysis, where a neural network analyzes the footage to identify elements like people, faces, interesting actions and different types of shots. Then there’s synthesis, where “we try to figure out what are the most cool and interesting parts of the video, to create a mini-music video for you with a high diversity of content.”
The app should get smarter over time as it gets more training data to work with, Patterson added. For one thing, she noted that most of the initial training footage used “Hollywood-style cinematography,” but as Trash brings more users on-board, it can better adapt to the ways shoot on their phone.
It’s staring that on-boarding process now with what Donovan calls a “creator beta,” where the team is looking for a variety of creators — particularly talented photographers who haven’t embraced video yet — to try things out. You can request an invite by downloading the iOS app. (Donovan said there are plans to build an Android version eventually.)
Trash has raised $2.5 million from sources as varied as the National Science Foundation, Japan’s Digital Garage and Dream Machine, the fund created by former TechCrunch Editor Alexia Bonatsos. Donovan said the startup isn’t focused on revenue yet — but eventually, it could make money through sponsorships, pro features and by allowing creators to sell their footage in the app.
And if you’re wondering where the name come from, Donovan offered both a “snarky response” (“I don’t give a damn and I don’t take myself too seriously”) and a more serious one.
“We believe that one person’s trash is another person’s treasure,” she said. “With filmmaking, as you know, there’s a lot of things that get left on the cutting room floor. That’s one of the product concepts, in the longer term, that we want to explore.”
View this post on Instagram
You shoot, we edit. Throw some video clips into TRASH and we’ll use AI to edit them together with music in one tap. Share your videos to TRASH and elsewhere, watch and react to videos from other creators. Link in bio to download TRASH from the app store today #gettrashed
A post shared by TRASH (@thetrashapp) on
The landmark study greatly increases the efficiency of surface delivery of nutrients and pesticides to plants. Currently, when crops are sprayed with stuff that’s supposed to help them grow faster or better, the vast majority of that (up to 95 percent, according to CMU’s engineering blog) will just end up either as concentrated deposits in the surrounding soil, or dissolving into ground water. In both cases, accumulation over time can have negative knock-on effects, in addition to being terribly inefficient at their primary task.
This method, described by researchers in detail in a new academic paper, would manage to improve efficiency to nearly 100% absorption of nutrients and pesticides delivered as nanoparticles (particles smaller than 50 nanometers across – a human hair is about 75,000 nanometers wide, for context) sprayed onto the leaves of plants, which then make their way through the plant’s internal vascular system all the way down into the root system.
Using this method, agricultural professionals could also greatly improve delivery of plant antibiotics, making it easier and more cost-effective to treat plant diseases affecting crop yields. It would be cheaper to delivery all nutrients and pesticides, too, because the big bump in efficiency of uptake by the plants means you can use much less of anything you want to deliver to achieve your desired effect.
This research could have huge impact in terms of addressing growing global food supply needs while making the most existing agricultural land footprint and decreasing the need for potentially damaging expansion of the same.
SpaceX is going to launch a Falcon Heavy rocket for only the third time ever tonight, should all go according to the current mission plan. The launch, set to take place during a four-hour launch window that opens at 11:30 PM EDT (8:30 PM PDT) tonight, will lift off from Launch Complex 39A at Florida’s Kennedy Space Center.
On its first-ever nighttime launch, Falcon Heavy’s STP-2 mission will carry a cargo made up of a number of payloads from commercial customers, as well as from the U.S. Department of Defence, the National Oceanic and Atmospheric Administration (NOAA) and NASA. The mission involves putting 24 different spacecraft into orbit, along three separate orbital paths. One of the is an experimental research satellite for the Air Force Research Laboratory, and NASA’s payload includes four different experimental craft, which the agency detailed last month.
It’ll also carry LightSail 2, a crowdfunded spacecraft spearheaded by Bill Nye’s Planetary Society, which will make its way through space using the literal solar wind beneath its massive sail. SpaceX is also re-using Falcon Heavy boosters for the first time, with side boosters used on the Arabsat-6A mission flown in April, and it’ll attempt to recover all three first-stage rockets via landings at Cape Canaveral and aboard its drone landing pad barge.
The launch will be streamed live above, with the feed getting started around 15 minutes prior to scheduled launch window opening.
Comscore’s name is usually in the news because of its widely-cited research and stats around media traffic and other analysis charting digital consumer behavior. More recently, it’s been coming up for another reason: ongoing corporate upheaval and its tumbling stock price. Today comes the latest development in that story: the company announced that it has raised $20 million, with the option of increasing the sum to $50 million, from a firm called CVI Investments.
“This transaction strengthens our balance sheet and positions us to pursue our refocused growth strategy while providing the flexibility to better apply resources to meet our business objectives, and ultimately drive long-term value for our stockholders,” Dale Fuller, Interim Chief Executive Officer of Comscore, said in a statement.
As explained in the 8-K, the money is coming in the form of a share purchase that is expected to close around June 26.
Comscore did not give more specifics about how it plans to use the funding, but it comes at a tricky time, with the stock today at one point dipping to a 52-week low at $7.39/share. Earlier this year, it lost both its CEO and its president, and then this month its COO departed after less than a month with the company. Counting its current interim CEO, it has been through five CEOs in the last five years. In May, the loss-making company also announced that it would be reducing headcount by 10%, or 180 people, as part of a restructuring and effort to move into profitability.
Comscore competes with the likes of Nielsen in measuring media consumption and patterns of digital consumers, but that is not its only challenge.
The company, and others like it, have traditionally been a key component in the world of advertising, as they provide an inportant, third-party assessment of audience data, necessary for helping to plan media spend and campaigns. But the rise of adtech and marketing tech, and a new array of places where ad inventory is placed beyond websites, has created a new level of more granular measurements and customer demands, so part of the challenge for Comscore has been to build new products to meet those new scenarios.
Its most recent series of executive departures and workforce reductions have not been the first faced by the company: it has also been the subject of an SEC investigation into its accounting practices, having admitted in 2018 that it overstated revenues by some $127 million resulting from a long-term WPP partnership. Prior to that, longtime CEO Gian Fulgoni left the company over the same problem.
Last year, it was reported that Comscore had engaged Goldman Sachs to reach out to parties potentially interested in acquiring it, including strategic acquirers operating in a similar space and buyout firms. The talks were never confirmed and nothing ever materialised at the time.
The company’s market cap is now at around $460 million, having seen its share price decline drastically since 2015.
With increased pressure on Chinese manufacturers like Huawei and ZTE, Shenzhen-based drone giant DJI has no doubt had cause for concern of late. In late-2017, the U.S. government’s Immigration and Customs Enforcement office raised concern that the company’s camera-equipped flying machines could be sending data back to China.
A few weeks back, the Department of Homeland Security similarly raised warning over commercial drones from China. In a hearing entitled “Drone Security: Enhancing Innovation and Mitigating Supply Chain Risks” last week, meanwhile, the National Defense University’s Harry Wingo told the Senate Transportation Subcommittee, “American geospatial information is flown to Chinese data centers at an unprecedented level. This literally gives a Chinese company a view from above of our nation.”
DJI fired back in a letter provided to TechCrunch, noting,
Because the drone industry is becoming an increasingly critical engine for small American businesses as well as the entire U.S. economy, it is essential that decisions affecting key components of the industry are based on fact. We are deeply concerned that, left unchecked, the unsubstantiated speculation and inaccurate information presented during your Subcommittee hearing will put the entire U.S. drone industry at risk, causing a ripple effect that will stunt economic growth and handcuff public servants who use DJI drones to protect the public and save lives.
The letter also breaks down some of the finer points of the discussion as follows,
DJI drones do not share flight logs, photos or videos unless the drone pilot deliberately chooses to do so. They do not automatically send flight data to China or anywhere else. They do not automatically transmit photos or videos over the internet. This data stays solely on the drone and on the pilot’s mobile device. DJI cannot share customer data it never receives.
DJI’s professional pilot app has a built-in setting to disconnect all internet connection, as an extra precaution for pilots performing sensitive flights. Unlike some technology companies, DJI does not sell or monetize customer data.
DJI embeds password and data encryption features in the design of our products. This provides customers with secure access to the drone and its onboard data. In cases when U.S. drone users do choose to share their data, it is only uploaded to U.S. cloud servers.
DJI operates a global Bug Bounty Program so the world’s security researchers can identify unforeseen security issues, and we hire independent security experts to test our products. These are just some of the steps we take to assure high-security users they can use our products with confidence.
With increased speculation, the company is looking to assemble some of its products stateside. A warehouse in Cerritos, California is set to be repurposed to build those drone models sold in the U.S., in order to better comply with government regulation.
The company tells TechCrunch,
DJI is committed to investing in America and providing U.S. government workers, first responders, and public servants with customized solutions that meet their unique security, safety, and procurement needs. As part of our long-term commitment to America that began in 2015 with our research and development facility located in Palo Alto, we are opening a new production facility in California and filing for compliance under the U.S. Trade Agreements Act. This new investment will expand DJI’s footprint in the U.S. so we can better serve our customers, create U.S. jobs, and strengthen the U.S. drone economy. We look forward to working with U.S. Customs and Border Protection in its review of our application.
DJI hopes that assembling products in California will help the company better comply with the Trade Agreements Act, a move that comes as preps the release of the DJI Government Edition, a repurposed Mavic Pro drone designed for use by government agencies.