Gannett Case Study

Gannett has a leading role in the publishing industry, operating the largest local-to-national media network in the country—the USA TODAY NETWORK—with its iconic USA TODAY and 107 local properties in 34 states and Guam. It also operates Newsquest in the U.K., which provides 150 local digital and print news brands. To stay competitive in the turbulent media market, Gannett turned to Amazon Web Services (AWS) and Chef, an AWS Advanced Technology Partner, for an automated computing strategy that lets it rapidly deploy new digital service offerings while maintaining its hybrid IT infrastructure.

A key to success for today’s media companies is rapidly producing, delivering, and continually updating digital content. Gannett reports that its focus on this strategy led to fourth-quarter 2015 results that included more than 100 million unique monthly visitors, a nearly 32 percent increase in its digital advertising, and 39 percent growth of digital-only subscriptions across its U.S. properties.

That growth has been supported by a critical restructuring of the way Gannett’s IT department runs its development and operations tasks, with help from AWS and Chef. It was a transition that allowed Gannett to gain the speed and agility it needed to compete in its industry.

Gannett's traditional digital product deployments were hobbled by multiple handoffs, manual tests, and siloed development and IT operations teams. Maintaining accurate, repeatable builds was difficult, with many build failures and tests that often ran in the wrong environments. Deployment and provisioning times could range from a few days to several weeks as developers went through multiple steps before handing off their code to a physically separate hosting team working out of two physical data centers.

“It was a situation that had to change,” says Erik Bursch, Gannett’s vice president for platform as a service. “Today, media content is increasingly consumed through digital and mobile channels, and we’re experiencing dramatic shifts in how consumers use technology and their expectations for what we can offer, such as more personalization of news content. Our task is to consistently provide them with the fastest and best digital experiences possible. Critical to that effort is our ability to react faster. We can’t underestimate the value of speed and consistency in getting our products to the market.”

Internally, there had been a steady push to use cloud technologies. On the development side, a kind of "shadow IT" evolved, with developers spinning up Amazon Elastic Compute Cloud (Amazon EC2) instances and using personal Heroku accounts, and then linking these to production DNS. However, there was no oversight over the costs of applications, and security had no way to audit the stacks.

Gannett developers had created a "shadow IT," leveraging the scalability of AWS tools and scalability to build applications. But as illustrated above, their efforts were disconnected from the Gannett data centers where deployments took place.

Bursch says there was an obvious need to integrate development and operations for greater efficiencies. “But we had to figure out if it was viable, because it would be a large investment and significant shift in our processes and in our culture,” he says. “Our only other choice was to continue doing what we were already doing, but it was extremely painful to deploy updates. With our desktop environment, we were lucky to get one update deployed per week, and there was a lot of internal frustration.”

Bursch says Gannett recognized the cost and agility benefits of the cloud. With AWS, developers were using standardized tools and resources while benefitting from scalability and the cost-efficient compute-on-demand model.

A turning point came when Frank Hanson, Gannett’s senior manager of platform as a service delivery, wanted to rebuild a hybrid architecture on AWS for a development environment that would mimic production. The goal was to provide a far more streamlined, automated solution for deployments. Hanson turned to Chef’s DevOps platform, which provides service configuration, deployment, and management tools that work on AWS. Other developers noticed, and soon discussions turned to the possibility of automating both Gannett’s development and operations processes using Chef and infrastructure as code principles.

The decision entailed a lot of variables and cross-enterprise consensus building. For example, the IT team decided early on to involve people from security and finance and began planning for training, including grassroots, side-by-side mini-training sessions.

"It was really the introduction of DevOps to our teams," says Hanson. "There was pushback at first from both development and operations teams over learning new processes. But this resistance is to be expected with such a drastic culture change, at least in a company that operated the way Gannett did for so long."

The IT teams started small, working first with the mobile API group, copying and pasting example code into Chef "cookbooks." Initial successful projects led to more teams using Chef and AWS for additional features and functionality, such as content management.

"We got real buy-in when people understood that Chef was code, the same as any other application. Before that, they thought it was just scripts to get into the environment," says Hanson. "As people became more experienced, they moved beyond copy and paste and started to consider what changes they could make to actually benefit their applications. Growing expertise and confidence led the teams to consider how to improve the process as a whole. There were a lot of different developers or different teams making advances with Chef that caught on and got picked up by other teams. It really started to spread like wildfire after grassroots efforts began to pay off."

The DevOps approach began filtering through different Gannett teams. For example, Bursch began including members of the QA team with developers to incorporate requirements into code at the outset instead of making modifications after applications were completed. Security teams learned to frame their needs in terms of Serverspec requirements. People from operations and architecture joined development teams to help design environments that supported the applications and to write cookbooks that would provision those environments.

Using Chef for automating application delivery on AWS enabled Gannett to create a consistent development and deployment workflow that automates the creation of infrastructure and application stacks while meeting the needs of finance and security. By early 2016, about 40 percent of Gannett’s IT teams were accessing the AWS cloud, with the company moving toward 100 percent adoption.

Using Chef, Gannett was able to link and streamline the efforts of developers working in the AWS cloud with deployment functions in the company’s physical data centers.

The benefits are already impressive. Different teams are working more closely together, there is greater visibility for tracking and auditing changes throughout the environment and, most importantly, applications share a common deployment methodology that can be customized to expedite application delivery.

"We’ve been able to reduce application deployment times from weeks using our former on-premises infrastructure to just minutes using AWS and Chef,"says Bursch. "Instead of a single desktop application deployment in a week, like we experienced in the past, we’re now deploying an average of 25 per week. That means more timely services for our customers by using the latest digital technologies to build and retain readership."

Chef
Chef, an Advanced Technology Partner in the AWS Partner Network (APN), provides DevOps automation tools. For more information about how Chef can help your company build, deploy, and manage your infrastructure, see the Chef listing in the AWS Partner Directory.