Uncovet, a startup that wants to be the recommendation engine for indie designer clothing and accessories, has raised $1.3 million from Javelin Venture Partners, Siemer Ventures, and L.A. angel investor Paige Craig.
A graduate of Science, Uncovet is an accessories and home décor site that is building a style graph to personalize product recommendations. With more than 500,000 monthly visitors, the startup uses Facebook Connect to allow users to invite friends to the site and can also take a user’s Facebook Likes into account when recommending products.
The result is a personalized and curated group of products that act as a style graph. The company, which launched last year, also features “Daily Finds” and will be launching its first native iPhone app early June (half of the startup’s daily visits already come from mobile device. Android and iPad apps will be released later this year.
Uncovet shoppers visit the site more than six times per month and will open an Uncovet email nine times per month. Uncovet delivers a more engaging experience by targeting specific consumer segments with products that resonate with their specific tastes. Uncovet is a next-generation version of Urban Outfitters; filled with unique items from independent designers, targeting a young, trend-conscious audience, and leveraging a new world of social data to fuel growth and loyalty,” says Michael Jones, CEO of Science.
Google continues to increase the reach of its Google+ platform, and today the company is launching a new mobile content recommendation service powered by Google+. These recommendations will appear as small widgets at the bottom of the screen as users browse a news site that has enabled this service. Google’s launch partner for this service is Forbes, but others can implement these recommendations by just adding a single line of code to their mobile sites. Recommendations, Google says, can appear regardless of whether users are signed in to Google+.
As Seth Sternberg, Google’s product manager for the Google+ platform told me last week, the team set out to create an “awesomely seamless experience to find more content” on the mobile web. On mobile sites, he argues, publishers often see high bounce rates because users have a hard time finding interesting additional content to read on a site once they have finished reading an article.
These recommendations, Sternberg told me, are based on social recommendations on the site from your friends on Google+ (only if you are signed in, of course), what the story you just read was about, the story’s author and some of Google’s “secret sauce.”
The new Google+-based recommendations, interestingly, only appear once a reader slides back up on a page. This, Google’s analytics show, is a pretty good indicator that a user has finished reading a post (even if there is still more text left on the page). The recommendation widget then slides up from the bottom and one extra click brings up more relevant items for the page. The other option is to show the widget after a user scrolls past a configurable CSS entity.
Publishers will be able to manage the recommendations widget from their Google+ publisher accounts. From there, they can decide when exactly the widget should appear and manage a list of pages where the widget shouldn’t appear, as well as a list of pages that should never appear in recommendations.
Desire2Learn is a 10-plus year old Canadian company that makes learning software — a learning management system to be precise — and here’s why, in spite of that description, it shouldn’t bore you to sleep. In a space that’s traditionally been controlled by bigs like Blackboard and Moodle, Desire2Learn has quietly managed to carve out its own growing share of the market.
Last September, the Waterloo-based company raised a sizable $80 million round from NEA and others, and today has over 700 clients and more than 10 million people across higher education, K-12, healthcare and beyond are using its learning software.
Although the company doesn’t disclose financial information, we’ve heard that its institutional contracts are now translating into millions in revenue, which along with its raise, has allowed it to expand its staff from 600 to over 750 over the last year. In turn, the company has been ramping up its focus on acquiring EdTech talent and is rumored to be planning an IPO in the U.S. at some point down the road.
While Desire2Learn has established a solid base, it’s strategic M&A that can help lead the way forward, fighting off a flattening growth curve and leading to better products. The company has been acquiring with more frequency of late, including two back-to-back in January and March.
Desire2Learn acquired course recommendation engine, Degree Compass in March and is already putting its tech to use to continue expanding its learning platform. This week, the company announced what it called “the biggest update to its Learning Suite to date” — an update in which Degree Compass’ tech plays a central role, not only by expanding its toolset but by potentially changing the way students navigate their academic career.
To do this, Desire2Learn wants to bring predictive analytics into play in education. But why? Well, first and foremost because, today, if students want to figure out whether a course is right for them — or how well they might perform in that course — they’re hard pressed to find a good answer. They can ask fellow students, check websites that rank faculty based on nebulous criteria or try to find surveys, but none of these options are ideal.
With its new analytics engine, Desire2Learn aims to change that by giving students the ability to predict their success in a particular course based on what they’ve studied in the past and how they performed in those classes. The new, so-called “Student Success System,” was built (in part) from the technology it acquired from Degree Compass; however, while Degree Compass used predictive analytics to help students optimize their course selection, the new product aims to help both sides of the learning equation: Students and teachers.
On the teacher side, Desire2Learn’s new analytics engine allows them to view predictive data visualizations that compare student performance against their peers so that they can identify at-risk students, for example, and monitor a student’s progress over time.
The idea is to give teachers access to important insight on stuff like class dynamics and learning trends, which they can then combine with assessment data, to improve their instruction or adapt to the way individual students learn. In theory, this leads not only to higher engagement, but also better outcomes.
For students, they use Desire2Learn as they normally would, using it to view course materials, take quizzes, submit homework and chat with their peers. The platform then collects and analyzes each student’s personal data and, by drawing from a wider set of inputs, the engine can more accurately predict which classes students will perform best in and what their grades will be.
The system is currently operating at about 90 percent accuracy when it comes to predicting performance by letter grades, CEO John Baker tells us — a number which should improve as the engine accumulates more data, he says.
In addition to its predictive analytics, Desire2Learn is also making some significant updates to its mobile app, including new integrations with Dropbox and SkyDrive to allow students to engage with learning resources in the same way they do outside the classroom. What’s more, Desire2Learn is moving into Patbrite’s territory through ePortfolio and its new tool which allows students to build portfolios based on their in-school projects, grades and achievements in a way that’s applicable to life after school and finding a job.
Essentially, the tool allows students to move their academic resume to the cloud so they can take it with them after they graduate, which the company is incentivizing by offering 2GB of free storage.
Basically, what we’ve come to realize, the Desire2Learn CEO tells me, is that the company’s initial approach to business (or academic) intelligence was off track. “Students and teachers don’t necessarily want more data, they want more insight and they want that data broken out in a way that they can understand and helps them more quickly visualize the learning map,” he says.
When I asked if building and adding more and more tools and features would dilute the experience and result in feature overload, Baker said that the company doesn’t want to build a million different tools. Instead, it wants to become a platform that supports a million tools and allows third-parties that specialize in particular areas of education to help develop better products.
Through open-sourcing its APIs, Desire2Learn along with Edmodo and an increasing number of education startups are beginning to tap into the potential inherent to the creation of a real ecosystem. Adding predictive analytics tools gives Desire2Learn another carrot with which they hope to be able to draw both teachers, students and development partners into its ecosystem.
Online real estate company Trulia just released its earnings for the first quarter of 2013, reporting that its revenue grew 97 percent year-over-year to $24 million.
Despite the growth, the company still posted a net loss of $2 million. On a non-GAAP basis, it lost 2 cents per share. Analysts had predicted a loss of 1 cent per share with revenue of $21.08 million.
Total traffic has grown too, from 20.6 million unique monthly visitors during this period last year to 31.4 million this year. And it had 11.4 million uniques on mobile. (Trulia has changed the way that it counts mobile traffic, so we can’t offer an apples-to-apples comparison — previously it was just usage of downloaded apps, but now it also includes traffic to the Trulia website from tablets and other devices.) And the number of subscribers has grown 42 percent year-over-year, to 27,920.
In the earnings press release, CEO Pete Flint said:
Trulia achieved an excellent start to 2013. We achieved another quarter of record revenue, driven by strong execution in both our Marketplace and Media businesses. Trulia’s mobile traffic continues to expand at a rapid rate, while our subscriber base grew by approximately 3,500 during the quarter.
In the past quarter, Trulia also launched a new recommendation engine called Trulia Suggest. Since the launch, Trulia says users have performed 2 million “likes” or “hides” on properties in the company database.
As of 4:45pm Eastern, Trulia is up 6.68 percent in after-hours trading.
Verious, the mobile component marketplace which launched out of TechCrunch Disrupt SF 2011, is today debuting a new service which it calls a “code recommendation engine” for developers. The service aggregates the code and content from 10,000-plus sources, including around 3 million dev and design resources. This means that the engine scours a much larger database than the 2,000 mobile app components Verious.com previously offered.
“Last fall, as we were refining our product roadmap, we spoke to a number of developers who told us that they were constantly searching for content to help accelerate their programming efforts – beyond pre-built components,” explains Verious founder and CEO Anil Pereira. He says that sometimes the content they were looking for was more for planning purposes, while other times it was to serve a more immediate need.
“These activities spanned mobile, web and all other platforms, and what we found in every case is that developers started by searching and then ended up with fourteen browser windows open with content from various sites to determine what was the best path forward to solve their immediate coding challenge,” Pereira says.
Pre-built mobile app components, like those that Verious today offers, were only one part of the equation, so developers would turn to Google search instead. Pereira says the company realized they could help fill this need, by bringing all programming content into one search service.
In fact, Google once served this very same vertical with its Code Search Engine, but announced back in 2011 that it would shut that service down. However, code.google.com still lives today, though others have reported encountering 404′s at times. Whatever Google’s intentions, it doesn’t sound like it will actively develop Code Search going forward. This has led developers to turn to several alternatives, many of which are listed here, and some which offer larger databases than what Verious does today. However, many of those services are focused on narrower verticals – like open source code, for example.
To build its new engine, Verious used APIs from sources like YouTube, Scribd, GitHub, Slideshare, Stack Overflow, and others, while also going after the long-tail of development and design-related content from blogs, online tutorials, and other niche programming sites.
The challenge in building such a resource wasn’t only the size of this database (as detailed above), but the other efforts that had to be made in order to scrub and normalize the search results, index listings, assign tags to content, and make the system capable of fetching new content from feed sources daily, while also being able to add new sources on a regular basis.
Because of the value the team saw in this type of research tool, Verious also made a key decision about the company, too: it decided to move beyond being a “mobile only” service, and instead attempt to include every possible programming language, platform and framework. Pereira says Verious now has over 170 of these.
The new product launched into private beta this December, allowing users to save items found in search results to their collection on Verious. These online collections can either be kept private, or shared with others on a developer’s team. Other content, such as personal links and bookmarks, can be saved to these collections, too.
The final piece was building the recommendation engine. This proprietary technology looks at several factors, including popularity, usage, downloads, views, followers, favorites, and/or ratings, etc., as well as how “finished” a code snippet may be.
“The closer a piece of content is to helping a developer find code they can use, the higher a weighting it is given,” Pereira explains. “Since every item has attributes that have it fall on a particular scale, we then translate each scale into a common scale and that is how we output ‘top 10′ recommendations for each query.”
Something of a by-product of all this work is another feature that works like an “About.me” page for developers. Because Verious had been pulling in blog and website feeds for the search engine, they were able to create developer profiles along the way, linking to all of a developer’s content and code-related activities on places like GitHub, YouTube, StackOverflow and more. Of the 200,000 active developers Verious has spotted, it has managed to manually create over 1,000 of these profiles to date, and is now automating the process for the rest.
The updated version of Verious.com, which has now been revamped to showcase its new focus, is live now for everyone.