Your address will show here +12 34 56 78
Blog, eCommerce

The hyper-competitiveness of the e-commerce industry has turned it into one of the biggest drivers of technology innovation. Today, the need for differentiation and the pressure to discount is driving e-commerce companies to look more closely at big data innovations such as predictive and prescriptive analytics, as well as artificial intelligence solutions, to remain competitive. The challenge lies in vast amounts of data. Big data is providing the opportunity for […]

The hyper-competitiveness of the e-commerce industry has turned it into one of the biggest drivers of technology innovation. Today, the need for differentiation and the pressure to discount is driving e-commerce companies to look more closely at big data innovations such as predictive and prescriptive analytics, as well as artificial intelligence solutions, to remain competitive.

The challenge lies in vast amounts of data. Big data is providing the opportunity for innovation, but its abundance and unstructured nature are inhibiting progress. To be data-driven, you first need to make sense of the data. For machines to read data, it first needs to be in a structured format. Human computing is unlocking opportunities hidden within a fog of data.

Big Data as a Driving Force in Innovation

Some of the most important innovations that have revolutionized the e-commerce industry have come from big data. eBay, a global leader in personalization, is a great example. Big data serves as a tool to drive innovation in customer experience with its ability to recommend similar items to a shopping guest. Also, big data powers eBay’s predictive machine learning algorithms that detect fraud based on purchase history and take action based on learnings within the data.

The eCommerce industry has access to substantially more data than many other industries. From CRM data and purchase history to social, website traffic and cookie data, eCommerce has an abundance of data to help them plan their next strategic move.

What Big Data Can Do For e-Commerce

The abundance of data that we now have access to gives retailers the opportunity to provide guests with a personalized experience and improve business operation efficiencies. According to Jeff Bezos, big data is powering the core functions of e-commerce businesses today.

“Machine learning drives our algorithms for demand forecasting, product search ranking, product and deals recommendations and much more. Though less visible, much of the impact of machine learning will be core of this type – quietly but meaningfully improving core operations.” – Jeff Bezos, Amazon CEO

For example, when TeeSpring, a custom apparel company, was looking to make it easier for their guests to find the right t-shirt design for them, they knew that with better data, they could offer their customers a better experience. For a guest to find exactly what they are looking for (or the perfect things they didn’t know they needed), TeeSpring knew that each graphic would need to be tagged with descriptive terms to ensure it showed up in related searches. Enhancing the search functionality with better, more accurate data helps TeeSpring offer a better customer experience.

Let’s look at a few other examples of big data-driven innovations:

Personalized User Experience with Customer Profiling

By combining incoming data to build customer profiles, you can better understand what motivates a customer to purchase. Compiling multiple touch points from different channels in real-time allows you to create a holistic profile. After purchase, you can build a more robust profile with basket analysis and leverage data mining to understand more about who your shoppers are, including their age, income range, buying habits, and preferences. Customer profiles allow e-commerce companies to customize the shopping experience.

Dynamic Pricing

As we mentioned, the e-commerce market is hyper-competitive. Reduced customer loyalty means that shoppers will hunt for the best price. e-Commerce companies need dynamic pricing to remain competitive. Data needs to be gathered from multiple sources including competitor pricing, sales, regional preferences, and customer actions.

Customer Service

Exceeding expectations in customer service will build loyalty and generates word of mouth. With robust customer profiles that connect data sources, when a customer calls to complain, the customer services agent can see that what the customer has already said on contact forms and social media. This data equips the agent with more information to satisfy the guest.

Predictive Analytics

Know what is going to happen before it does. Big data powers machine learning algorithms that predict events. This feeds into supply chain visibility by understanding sales patterns to ensure product availability and avoid out-of-stock items.

Big data can also be a Big problem

All of this data and the need to remain competitive is driving technology innovation in e-commerce. While it is the data that enables innovation, the data also hampers progress. Leveraging big data is a big challenge. E-commerce companies are drowning in data. The amount of data, the variety of data, and the speed at which it is piling up are the biggest hurdles for e-commerce companies to overcome when looking to leverage their data.

Amount: Insights and product teams have access to so much data that they are overwhelmed.

Variety: Data comes from multiple channels and in different formats. It’s impossible for machines to make sense of this abundance of unstructured data.

Velocity: The heap of unstructured data is growing rapidly every single day.

e-Commerce companies can’t be data-driven without first understanding what the data is saying. For technology innovations to be leveraged, this abundance of data must first be structured. Machines aren’t currently capable of doing this; humans must intervene. From the product side, data labeling takes up the majority of a data scientist’s time. This time would be better spent focusing on core projects that will provide e-commerce companies with the competitive advantages needed today.

Human-in-the-loop computing unlocks the value hidden within data at scale

Product teams are hindered by the amount of data that requires labeling. Recent innovations such as increased bandwidth and cloud computing have created a newly digitized economy across the globe providing opportunity in places where they previously did not exist. At iMerit, we constantly encounter unexpected use cases where customers are exploring their data and call upon us to unlock parts of the trove to improve the customer experience.

By handing off data labeling requirements to data specialists, product teams can focus on the innovations that will provide the competitive edge required today. Our scalable, on-demand workforce provides e-commerce companies with data services that promote:

  • Increased search relevance providing a better customer experience
  • Conversion rate optimization through up-selling and dynamic pricing opportunities
  • Enhanced customer service with converged data

To remain competitive, e-commerce companies need to be focused on what their customers want. On-demand workforces allow you to outsource the data work while remaining focused on what’s next.

Do you have data needs? Launch a custom workforce to handle your data work today.

0

Blog, train ai

Today is the day! The iMerit team is very excited to be attending CrowdFlower’s Train AI conference again this year as a premiere NDA partner and Silver Sponsor. iMerit partners with CrowdFlower to offer you a secure option for your sensitive and critical data projects through iMerit’s curated team of over 800 in-house data experts. There are so many great Artificial Intelligence conferences to attend every year. While many conferences […]

Today is the day! The iMerit team is very excited to be attending CrowdFlower’s Train AI conference again this year as a premiere NDA partner and Silver Sponsor. iMerit partners with CrowdFlower to offer you a secure option for your sensitive and critical data projects through iMerit’s curated team of over 800 in-house data experts.

There are so many great Artificial Intelligence conferences to attend every year. While many conferences will focus on the algorithms, Train AI takes a different approach focusing on the other 90% of the work, the data preparation, training data, and feature deployment.

What to look forward to at Train AI

There are so many great speakers at this year’s event. Here are a few that we are really looking forward to hearing from.

Creating Software with Machine Learning: Challenges and Promise

Peter Norvig, the Director of Research at Google will be how the traditional programmer’s role is being replaced by a trainer. The trainer’s role is to show the computer examples until it learns to complete the task. Peter will explore the exciting opportunities that this shift presents.

Computer Vision: helping machines see and understand the world

Lukas Biewald, the founder of CrowdFlower, will be going over all of the changes that CrowdFlower has experienced over the past year from deep learning and computer vision applications. He’ll be talking about use case from a variety of industries including healthcare, self-driving cars, drones, and animal conservation using tools that support their customer’s labeling and human-in-the-loop workflows.

Amplifying machine learning with human intelligence

Toby Segaran, the Director of Engineering at Reddit will explore ways that human intelligence can be used during the process of optimization, topic modeling, and entity resolution to improve algorithmic results as they’re happening. This stems from the fact that the most common approaches to using machine learning often uses humans to initially label a dataset for training and quality checks.

Reaching unparalleled quality with human-in-the-loop

Maran Nelson is the Co-Founder & CEO at Clara Labs. She will be discussing how human-in-the-loop processes are increasingly the best pathway to reach a high-achieving system. Maran will discuss the implementation of this hybrid structure that takes what works best in AI and humans and creates outperforming products.

Tackling The Limits of Deep Learning

Rochard Socher, the Chief Scientist at Salesforce, will be discussing the problems and limitations of Deep Learning. In this talk, Richard will be going over solutions to some of these including: How to predict previously unseen words at test time, how to have a single input and output encoding for words, how to grow a single model for many tasks, and how to use a single end-to-end trainable architecture for question answering.

Are you Attending Train AI?

If you’re at the event today, be sure to stop by the iMerit booth, say hi to our awesome team, and learn about how iMerit partners with CrowdFlower to offer NDA Channel services. If you have data labeling projects that are sensitive or require a high degree of security, come chat with us!

0

Blog, self-driving cars

The New York Times reported on Sunday that Waymo, the self-driving unit of Google’s parent company Alphabet, and the ride-sharing startup, Lyft are teaming up to bring self-driving technology to the mainstream. “We’re looking forward to working with Lyft to explore new self-driving products that will make our roads safer and transportation more accessible. Lyft’s vision and commitment to improving the way cities move will help Waymo’s self-driving technology reach […]

The New York Times reported on Sunday  that Waymo, the self-driving unit of Google’s parent company Alphabet, and the ride-sharing startup, Lyft are teaming up to bring self-driving technology to the mainstream.

“We’re looking forward to working with Lyft to explore new self-driving products that will make our roads safer and transportation more accessible. Lyft’s vision and commitment to improving the way cities move will help Waymo’s self-driving technology reach more people in more places,” Waymo said in a statement to Recode.

Together, these companies have partnerships with the majority of major auto manufacturers. Lyft announced a partnership earlier this year with General Motors to test the Chevy Bolt with the general public within the next few years. Waymo has deals with Fiat Chrysler and Honda testing their technology on the road.

What does this mean for Uber? They have poured hundreds of millions of dollars into the development of self-driving cars to catch up to Google and view the technology as crucial to their future. Uber has had a rough year to date; this may be a considerable setback for them.

The Latest Self-Driving Technology Updates

Self-driving cars are one of the hottest things in tech right now. It feels like just yesterday we were saying “can you imagine!?” Here we are in 2017 on the cusp of having autonomous cars pick us up at our front door. To get an update on where we are, here are a few more updates on what is going on in the world of autonomous cars.

Training Datasets Released For All to Leverage

car1

Self-driving cars use advanced Artificial Intelligence algorithms to make thousands of decisions. To know what decisions to make, the algorithms are trained using datasets of various scenarios.

Training datasets are usually very expensive to create because it takes a lot of time to annotate the images. Annotating a single image (or a single frame from a video) can take between seconds and hours depending on complexity or, how much you are looking to teach an algorithm.

Luckily for technology startups, according to TechCrunch, Mapillary is releasing a free dataset of 25,000 street-level images from 190 countries, with pixel-level annotations that can be used to train automotive AI systems. Mapillary is a crowdsourcing company that uses computer vision to read images uploaded to their database by people around the world using smartphones to identify locations in 3D and recognizes the order of objects within them.

The release of this dataset opens new opportunity for tech startups to advance machine learning algorithms used in self-driving cars. It’s no surprise that this dataset release was sponsored by big auto manufacturers Lyft, Toyota, and Daimler.

Humans may be what is slowing down self-driving cars

car2

The benefits to self-driving cars are many: safer roads, less traffic, lower fuel consumption, and don’t forget enhanced human productivity – no more lost hours driving, you can now be productive on your commute. With all of these benefits for humans, it turns out that we may be the problem holding the technology back.

Driving takes a certain amount of assertiveness, according to John C. Dvorak, Columnist at PCMag.com, self-driving cars are too polite. In ‘right of way’ situations like 4-way stops, human drivers will assert their intentions to go; the autonomous car may sit until the intersection clear. If a cyclist is hogging the road, it will slow down and drive behind until the path is clear.

John Adams, a professor at University College London, says “Driving in cities would be unacceptably slow if autonomously-operating cars were required to assume that every pedestrian might jump into traffic as fast as humanly possible. But if pedestrians came to learn that cars would always avoid them then they would likely act in much less controlled ways on streets and pavements.”

Will the algorithms become more advanced to handle these situations? Or will humans have to adapt to allow for these polite road warriors?

No more fighting over parking spots

car3

Once a self-driving car has dropped you off, it needs to find a place to park. As a human driver, we all know how difficult and annoying this can be. A hackathon team that came out of the TechCrunch Disrupt NY event, Val.ai created a way for autonomous vehicles to bid for parking spaces in an auction.

The tech-twist here is that these cars aren’t looking for an empty parking spot, they are negotiating with other autonomous cars which are currently parked and will be leaving soon. The model was based on public parking spots which bring up concerns about using public space for private use, a term TechCrunch calls “#JerkTech.” But, there is still lots of opportunity for private parking lots.

There you have it, the latest in self-driving cars. Do you work on self-driving technology? We would love to hear from you to discuss how iMerit’s dataset services may be of use to you. Get in touch!

0

Blog, computer vision

On Monday, the iMerit team will be arriving in Santa Clara, California to join the many innovators, technologists, and engineers from the computer vision industry at this year’s Embedded Vision Summit. The Embedded Vision Summit brings product and application developers, business leaders, investors, and entrepreneurs who are all focused on embedded vision together to see the latest in practical technology. Attendees get to see the latest in practical technology and […]

On Monday, the iMerit team will be arriving in Santa Clara, California to join the many innovators, technologists, and engineers from the computer vision industry at this year’s Embedded Vision Summit.

The Embedded Vision Summit brings product and application developers, business leaders, investors, and entrepreneurs who are all focused on embedded vision together to see the latest in practical technology. Attendees get to see the latest in practical technology and dataset services to bring visual intelligence into cloud applications, embedded systems, mobile apps, wearables, and PCs.

Along with showcases of the latest computer vision products and services, attendees will get to hear from industry experts in five in-depths tracks on technical insights, business, enabling technologies, and fundamentals of visual intelligence. Steven Cadavid, President and Founder of Kinatrax, will be presenting on their markerless motion capture system that computes kinematic data of an in-game baseball pitch.

Markerless Motion Capture System Captures Kinematic Data of MLB Pitchers

KinaTrax’s Markerless Motion Capture System consists of a variety of imaging devices that are mounted throughout the baseball stadium to capture the detailed movement of pitchers. They develop 3D kinematic models that are used by teams to monitor pitchers’ performance and to prevent injury. Currently, these devices are mounted at the home stadiums of the Cubs and the Tampa Bay Rays, along with another undisclosed stadium.

Computer Vision and Machine Learning algorithms are used to capture the biomechanics of a pitcher at over 300 frames per second. The video is recovered in 3D and reconstructed frame by frame, producing an image for every motion within the pitch sequence which are then annotated at 20 joint centers. KinaTrax leverages iMerit’s team of computer vision data experts to provide on-demand and scalable annotation resources from end-to-end through KinaTrax’s data analysis workflow.

kinatrax-pitcher-image

To create the 3D models, “iMerit teams process in-game footage and prepare it for analysis. We then annotate images and videos to create 3D models of the players. iMerit integrates directly with KinaTrax systems, facilitating a seamless handoff of data. In the past year, iMerit has built a precise 3D model for approximately 300 pitchers across all 30 MLB teams, covering most pitching personnel,” said Jai Natarajan, VP of Technology at iMerit.

Computer Vision Algorithms and Humans Work Together to Create Baseball’s New ‘Moneyball’

The idea of leveraging in-game data to enhance pitcher performance “will make a profound difference in major league baseball, it will change the game,” said Cadavid. Billy Beane revolutionized baseball with his analytical, evidence-based approach to selecting players dubbed ‘Moneyball’. KinaTrax’s offering is supplying team management with the data they need to make the best decisions about a pitcher’s health.

The 3D Kinematic models can be used to generate comprehensive and customizable biomechanic reports. According to Steven Cadavid, “The primary uses for it are evaluating the mechanics over time, the performance enhancement and injury prevention component. From an injury prevention, the datasets we’re collecting are really unprecedented.”

KinaTrax is taking a different approach to motion capture technology by leveraging video annotation to gain the data required to build 3D kinematic models. This means that the subjects, in this case, the pitchers, don’t need to wear markers to capture the data. This key element to KinaTrax’s technology allows them to not only capture training data but in-game data as well. Combining this with iMerit’s on-demand dataset service offering, humans in combination with technology are revolutionizing the game.  

Want to Learn More about Data-Driven Baseball?

If you are attending the Embedded Vision Summit May 1-3, be sure to check out Cadavid’s talk on Using Markerless Motion Capture to Win Baseball GamesAlso be sure to come by the iMerit booth to grab some Philz coffee. If you are unable to attend, check out this video.

Want to learn more about how iMerit can help your Computer Vision technology? Get in touch with our team today.

0

Blog, Dataset Creation

A recent study released by development consulting firm Banyan Global sheds light on Microwork and Impact Sourcing highlighting the different sides of the industry by looking at what clients want out of service providers and how to ensure that workers are being offered steady work with future opportunity. Microwork is a series of small tasks which together constitute a larger project. Impact Sourcing, also known as socially responsible outsourcing, refers […]

[Research] Big Data Drives Boom For Microwork and Impact Sourcing

A recent study released by development consulting firm Banyan Global sheds light on Microwork and Impact Sourcing highlighting the different sides of the industry by looking at what clients want out of service providers and how to ensure that workers are being offered steady work with future opportunity.

Microwork is a series of small tasks which together constitute a larger project. Impact Sourcing, also known as socially responsible outsourcing, refers to the creation of employment for high potential but disadvantaged people in the services sector via contract work.

Key findings in the study include:

  • Microwork industry is expecting 5x growth by 2020.
  • Data science, algorithm-based IT approaches and technology advances have expanded the range of work performed by microwork service providers.
  • Microwork service providers are structured in three different models offering varied benefits and drawbacks.
  • Clients selecting outsourcing partners look to business factors such as cost, quality, and timeliness as central decision factors.
  • Impact sourcing initiatives must be well structured and well-managed to produce cost-effective, high-quality products on time.
  • Concerns around microwork include security, implementation complexity, and potentially low-quality products when workers are not sufficiently skilled or managed properly.

Big Data + Data Science Drives Boom for Microwork

In today’s fast-paced digital world, change is the only guarantee. Organizations are struggling to keep up with innovation, having to spend the majority of their budgets and human resources maintaining their current offerings. This creates a gap between current offering and innovation. Over time, this gap threatens to open up as lean startups jump ahead and the competitive landscape evolves.

Technology breakthroughs and innovations have opened new opportunities for business. Today, advancements are commonly driven by leveraging the ever-increasing abundance of available data. The need to remain as lean as possible while extracting insights and innovation from data is the key driver in the evolution and growth of microwork and impact sourcing. This increase in data work has given rise to the microwork industry, changing the type of work available from very basic data work to work that requires domain-specific skilling, including image and video annotation to power machine learning algorithms or UGC content moderation to enhance user safety and experience. According to the study, it is projected that the microwork industry will grow from a $400 million industry today to $25 billion by 2020.

To remain competitive, companies need to be using data to their advantage; to better target customers, to provide personalized insights, and to enhance the customer experience. While everyone has access to data, the challenge is in gaining useful insights. To do this, the data needs to be in a structured format. Algorithms that are used in artificial intelligence, machine learning and computer vision rely on accurate data. The most advanced digital companies are heavily powered by human data services in the background. Cleansing data is a crucial task that needs to be done with extremely high accuracy to minimize false outputs and increase accuracy. Structuring data typically requires highly repetitive tasks completed accurately. High-growth companies are looking outside of their human resource pool to complete data tasks while their core team remains focused on core business objectives.

Technology innovations, such as cloud computing, increased bandwidth, and expanded access to reliable internet, are fueling employment opportunity for people all over the world. As the microwork industry grows, so does the opportunity for individuals in developing countries to join the “digitized economy,” offering skilled employment where there previously wasn’t any stable options, increasing their purchasing power.

The type of work that microwork providers can handle varies greatly but is centered in dataset skills. The study highlights work case studies from Ancestry.comGetty Images, and eBay. As an example, Getty Images, the world’s largest provider of digital media content, processes over 40,000 still images per day. To increase search accuracy for their customers, they outsourced image categorization to microwork service providers, including iMerit. Adding high-quality metadata to each image enables customers to find what they are looking for, providing a better customer experience.

Different Types of Impact Sourcing Models

The study has classified microwork service providers in three different categories, each one varying in advantages and disadvantages to both the client and the workers.

The Micro Distribution Model















This model is commonly referred to as ‘crowdsourcing’ – a company creates a platform that acts as an intermediary between client and workers. This model is highly scalable and low cost due to the lack of infrastructure as workers are self-employed and dispersed (working from their home or local cafes). This structure also has the capability for broad impact as it’s accessible to anyone with reliable Internet and basic literacy and numeracy skills.

The micro distribution model can cause challenges for workers due to fluctuating demand for services. Because of the dispersed nature of the structure, there is less opportunity for skill development and can create quality control difficulties due to lack of direct management.

The Direct Model



In the Direct Model, service providers build and operate delivery centers and employ and train local workers to complete work in those centers. Services providers will usually have a United States-based headquarters with delivery centers in developing and emerging countries.

In this model, workers are offered education, up-skilling opportunities, and a management team to offer guidance. While this model represents the highest level of investment, fixed capacity, and operational complexity for the service provider due to infrastructure and training, according to the study, it also has the highest potential for performing high volumes of work at a high level of quality.

The Indirect Model



This model adds a layer to the direct model between the client and the workers in the form of delivery partner, service intermediaries, or subcontractors, who find contracts to bring to their partners. While this is a very scalable model, quality control can be very complex and is affected by all collaborating firms.

iMerit is an example of the Direct Model with a headquarters in California and delivery centers in India. Employees are offered stable employment with continuous on-the-job training and advancement opportunities enabling them to join the digitized economy. A stable work environment, skills and opportunity advancement, and a management team on site leads to below 3% attrition rates and projects that are completed with above 95% accuracy on time, and on budget.  Knowing that the team you partner with today is the team that will be here tomorrow creates a sense that the iMerit workforce is truly a part of the client’s workforce. This helps build valuable long-term client relationships, encourages investment in knowledge transfer and technology, and adds to the quality of work.

Microwork increases production, scalability, and lowers costs without risk

With direct and indirect models offering the skilling required to complete more advanced tasks, more and more companies are looking to outsource their data work saving them money through lower salaries for workers, lower training costs, and lower attrition rates. According to the study, impact sourcing can lower client costs by up to 40% and replace the need for staff augmentation, saving addition costs, as service providers take on the responsibility of acting as the employer, or that workers are hired as freelancers.

Not only does enlisting dataset service providers reduce costs, but it can also help clients scale their work at a reduced risk. According to the study, the “large-scale digitization of service production and the unbundling of service value chains have enabled firms to view individuals and locations as calculable, marginal and substitutable in the performance of this work.” As an example, companies who are launching a new product will utilize outsourced workforces to help manage increasing amounts of data and to refine their product before launch. The study mentions that Microsoft did this when testing the algorithms used in its search engine, Bing.

Challenges and Concerns

There are concerns or challenges that arise when hiring an outside team to complete important work. According to the study, clients interviewed are concerned about security, implementation complexity, and potentially low-quality output. To reduce these concerns, initiatives must be well structured and well-managed to produce cost-effective, high-quality products on time.

In the case of the Direct Model, teams are managed in delivery centers. This creates an opportunity for high security as data is not leaving the centers. Through project managers, upskilling, and low attrition rates, teams can take on more complex projects. Because management is right there with them, quality control is a key component, ensuring the client receives high-quality work.

While I have focused mainly on the client-related aspects and opportunities that microwork and impact sourcing provide, the study details the incredible opportunity that it offers marginalized individuals all over the world. For more information and to read the entire study, you can check it out here.

For more information on the on-demand data services iMerit provides, get in touch today!

0