Automating operations

Avoid the workflow to nowhere by creating a new framework for digital content.
Author:
Updated:
Original:

The last few years have brought an unprecedented growth in the amount of content most broadcast, post-production, and other content creation and distribution entities have to manage. This trend will only continue in the years ahead as new content distribution channels emerge, and leveraging automated workflows can be a significant tool to successfully meet these challenges.

Unfortunately, budgets and headcount have not kept pace with the ever-growing volume of content in today’s modern media organization. Given the economic pressures forecast for the next few years, budgets and human resources are going to become even more precious. Many companies have already started the transition from physical assets and human-based work processes to file- and automation-based workflows. For others, the pressures of expanding content and flat or shrinking resources may finally provide the tipping point to make the leap to automation-driven work processes in 2009.

In many ways, this transition is inevitable as studies have shown automated file-based workflows can reduce the time or resources required by up to 80 percent for many content manipulation work streams. With this in mind, the goal for many organizations will be to use human talent only for operations that do not lend themselves to automation and for exception handling.

Decide what to automate

Faced with the task of deciding what processes can be automated and which automation and workflow technologies best fit a particular organization, many managers don’t even know how to start the process. As with any large task, it makes sense to break it down into smaller pieces and tackle them one at a time. Deciding what to automate requires analysis of the current processes and then ranking them against the company’s decision criteria. Each company has its own unique set of characteristics, so whatever framework is selected for gathering and designing the workflows should be as flexible and malleable as possible.

Many organizations have to be selective about which work processes are initially targeted for re-engineering and automation. The likely early candidates tend to fall into two broad categories. The first category is work streams that easily lend themselves to automation with little effort and cost, thus providing “quick wins” and building enthusiasm and momentum within the organization for further re-engineering efforts. The second category consists of work streams that may be costly or complex yet provide high return on investment or significant competitive advantages over competitor organizations.

Typically, work streams that don’t fall into these two categories get pushed to later phases in the initiative. Selecting the as-is work processes and designing and scoping the to-be automated workflows can be complex and time-consuming, so resources should initially focus on high value processes.

The 10 phases of workflow

Experience has shown that the process can be simplified by mapping existing and future work effort and processes to a series of phases, along with the triggering events and dependencies that move the workflow between phases. (See Figure 1) Many processes will only use a subset of the following 10 phases:

• Planning — the upfront knowledge about the work to be performed and the data entry within the various systems to prepare the workflows for activation. Resources, assets and time frames are first determined in this phase. The planning phase is typically driven by business artifacts such as program schedules, purchase orders or work orders.

• Receiving — the arrival of assets, metadata and business data into the organization from third parties or external departments within an organization. These arrivals are typically the triggers that start the workflow processes.

• Ingesting — the effort to enter the assets and data from the receiving phase into the organization’s internal systems. This can be tasks such as creating work orders, checking physical and digital assets into the content management system, transforming external metadata into internal standards, and unpackaging content from container formats. In addition, this phase can represent the transformation of the raw received assets into the proper working formats for the organization’s workflows via such tasks as dubs, conversions, or file encodes.

• Analyzing — the tasks regarding completing the gathering and input of additional data required for the workflow and verifying that the source assets are defect-free and meet the specifications required by the workflow. This includes tasks such as human or automated quality control checks, logging of assets and adding business or metadata to control systems.

• Preparing — the source content preparation work required as part of the workflow and any associated business processes. Typical tasks include editing, trimming and black removal. • Finishing — the tasks related to the final creation of the target assets and any associated business processes. These include element conformance, standards conversion, dubbing and transcoding.

• Packaging — tasks related to assembling and packaging of content into any required container formats for delivery. Typical tasks include creation of MXF files, tar/zip archives, digital cinema packages, CableLabs’ VOD wrappers, etc.

• Testing — tasks related to validation of assets, packages and other deliverables created by the workflow and the business processes related to such quality control activities.

• Delivering — the physical or electronic delivery of the finished products of the workflow. Typical delivery tasks include broadcast transmission, file transfer, streaming and physical shipment.

• Disposing — the final disposition of assets and other deliverables from the workflow and closeout of open business processes related to the workflow. These include vaulting or archiving of assets, returning assets to content owners and destroying assets.

Avoiding workflow islands

Once the automated work processes, resources, sources, triggers and deliverables are codified, it’s time to select a technology to implement the workflows. Many organizations have existing automation and workflow tools provided as part of their operations management, broadcast, encoding or editorial systems that can be leveraged to provide some or all of the required workflow functionality. While many of these vendors have made great strides in opening up their architectures, they still tend to focus on managing the automation of their specific segment of the end-to-end work stream. This leads to islands of automation within the enterprise workflow, with manual human tasks required to bridge the gap between environments.

In order to create an end-to-end workflow system, organizations must look beyond point solutions and strive to build a technology platform capable of managing the numerous heterogeneous systems present in the modern media environment. An enterprise workflow management system will naturally contain many complex features and interrelated components; however, there are five key attributes vital to an enterprise automation and workflow system.

Data management

Data is the fuel that powers the workflow engine as all automated decisions and events are driven by the data entered or created as part of the process. Without accurate and consistent data, the automated workflows cannot function and either stall with an exception or complete with potentially incorrect deliverables. Integrated, consistent, open and accurate business, asset and event data repositories are the cornerstone of any successful automation environment.

Unified physical and file-based media management

Many organizations maintain multiple disparate content stores. This is a natural outgrowth of the workflow island issue as many systems maintain their own databases of media and metadata. This duplication of data in several disparate systems adds rework and risk of error across the organization as personnel typically need to re-enter the same data into the various systems.

By consolidating the organization’s physical and digital assets into a unified repository, or implementing a unified metadata repository linking the disparate asset stores, workflows can better manage the movement, manipulation and status of all content required by the various automated processes in the enterprise. In addition, they can reduce the complexity and human effort required to manage the metadata entry, security and other media management tasks within the organization.

Comprehensive resource management

Centralized maintenance of the various attributes for people, devices and locations is a fundamental benefit of a resource management system. Each resource entity can have a series of related attributes that define its capacity, availability and capabilities. The workflow system can take advantage of these attributes to optimize automated resource allocation across the various workflows in the enterprise. Without a database of resource availability, capacity and capability to drive the automated workflow, many nontrivial workflow events will require human intervention to complete.

Central queue

In order to effectively provide for end-to-end process control and heterogeneous queuing across multiple technology vendors, organizations need to provide for a centralized control point for all tasks involved in the various automation workflows running in the enterprise. The centralized queues provide the “queue of queues” that allow processes, events and messages to traverse multiple proprietary vendor systems. The central queue can also be leveraged as the one place to manage and apply business rules to messages and processes as they move through the enterprise workflow system. Additionally, it allows for consistent and manageable application of business rules. The central queue also provides a messaging backbone and gateway and allows for the processing, transformation, and validation of real and non real-time messages and events throughout the various workflows.

Open integration environment

The final requirement for any enterprise workflow system is an open and extensible integration environment. The media management and automation market has a history of closed and proprietary systems. During the last few years, this has begun to change, and most of the major vendors in the broadcast, post-production and media management space have started to embrace standards-based messaging and data storage architectures. Standards bodies such as SMPTE and EBU are active in this area as well and have pushed technologies like XML, Web services, SOAP and TCP as the foundation for many of their standards efforts. The new BXF standard is a recent example of this trend.

Careful selection of a robust enterprise workflow management system built upon a technology platform with strong data management and open standards can provide significant cost, resource and time savings.

Conclusion

Many organizations are facing the daunting task of moving from manual- to automation-driven work processes in 2009. This can be made less difficult by developing a framework for the analysis, design and selection of these new workflows. It helps to look at automation technologies from an end-to-end perspective in order to prevent islands of automation connected by manual human tasks and thus provide the maximum benefit from implementing automation within the organization.

Tony Lockard is senior vice president of business development at Xytech Systems.