Big Data Drives Boom For Microwork and Impact Sourcing

April 13, 2017

[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!