This is the first part of a three-part series looking at API maturity and API lifecycle management. Part one discusses business consultancy Accenture’s API maturity model. Part two will review how businesses can scale their APIs to manage growth, while part three looks at how API maturity trends are influencing the field of more traditional application lifecycle management.
The recent release of Accenture’s Industrialized API Model aims to encourage businesses to take a more strategic approach to the development and deployment of APIs in the enterprise. Industry sources say the API maturity timeline is three times the speed of previous SOA maturity life-cycles. But moving this fast is creating additional costs and complexity. Taking an ‘industrialized API’ strategic approach can help businesses build a flexible, composable enterprise, according to leading API industry stakeholders.
Enterprise adoption of APIs
With businesses showing greater confidence in the security of mobile, applications are being created at a rapid pace. The latest Good Technology Mobility Index — covering Q1 2014 — shows that the growth of business mobile applications continues to surpass previous quarterly growth rates. Q1 2014 saw a growth rate of 57% in new mobile applications built by Good Technology’s enterprise customers, above the 54% and 43% growth rates recorded for the last two quarters of 2013.
ABOVE: New application enterprises by industry sector, Q1 2014, from Good Technology Mobility Index Report.
Many of these use APIs to connect with a business’ datasets. While this growth shows the growing dominance of API usage in business, the actual use of APIs by enterprise is even higher. APIs are fueling the integration of web apps, Software-as-a-Service business tools, and connected devices: all growing at a similar rate to what Good Technology is mapping with mobile.
In April this year, SOA Software Senior Vice President Brent Carlson told ProgrammableWeb that API maturity cycles are double or triple the speed of SOA Maturity approaches. According to Carlson’s colleague, Laura Heritage, the typical API adoption pattern in the enterprise is becoming: enterprises develop internal APIs; then quickly try to move them over the firewall so they can be used with business partners; then in mobile apps; and finally, more broadly with third-party developers.
“We have moved into the ‘post mobile’ world,” declares Mark O’Neill, VP of Innovation at Axway. “Until recently, people thought of APIs always in the context of mobile apps. But the steep growth trajectory of APIs has now eclipsed even that of mobile. You can now use the touchscreen console in the back seat of New York City taxis to pay for a ride with your credit card points – through an API. You can now perform a company lookup to the D&B API right in your CRM – through an API. So, rather than starting with the app, start with the API and focus on usability. This then enables the maximum amount of clients for your API – including, but not limited to, mobile apps. This is the key to ‘API First’ development.”
Application Use Cases Are Initiating API Deployments
“Starting with APIs makes sense as they sit at the intersection between business objectives and IT enablement,” says Dr Teresa Tung, Senior Manager at Accenture Technology Labs and leader of Accenture’s API Industrialization Initiative. “Our API maturity model proposes a common industrialized API program to ensure that effort is not duplicated and that common best practices are applied to meet business objectives.”
ABOVE: Accenture API Industrialization Model from Accenture Connecting the Digital Ecosystem.
“Understanding what APIs are desirable first offers an opportunity to work backwards to figure out how to enable them and if it is worthwhile. For example, the creation of APIs may not always be feasible due to legacy issues, but is there a cost effective substitute that meets the business objectives?
“Leading with apps and their use cases alone is not advisable and can result in a large number of silo-ed APIs and even API programs where each evolved separately. The result is added expense and complexity of redundant but disparate APIs, API management products, and governance processes —- this situation can be avoided by laying the appropriate groundwork in the initial stages introduced within our API maturity model.”
“Although individual apps may evolve separately through their own ALM processes, a coupled approach between apps and APIs is critical. Ultimately it’s a cycle where apps provide the use cases and success metrics that justify the API program. And an industrialized API program provides apps (both external and internal) access to an enterprise’s data and services with the proper design, architecture, and developer support.”
Building an ‘API Backplane'
O’Neill agrees. He sees the best way for businesses to start an industrialized API strategy is to focus on the low-hanging fruit: the datasets that can be opened via API and immediately used by partners. O’Neill says:
“We see that enterprise APIs often begin by enabling discrete useful capability to partners such as inventory lookup, or shipment tracking. The best practice is to create a common ‘API Backplane’ which services internal consumers, partners, and customers. At the API layer, policies applied in the API Gateway decide what access to the API is delivered to each type of client.”
Tung understands how it may feel more natural for businesses to start with their use cases and work from there to enable an API, but cautions against it.
“Many programs started with a technology-first approach to APIs focusing on IT implementation. This approach works when use cases are well-known and is likely to get an organization through the initial release, but is not sustainable. It does not provide a method to evaluate success and accommodate new use cases.
“The turning point where APIs become a C-level topic occurs as they are recognized as business-first artifacts needed for an organization to compete as a digital business – as the connectors that enable digital interactions with consumers, partners, and even internally.”
The Accenture Industrialized API program assesses the maturity stage of an API by categorizing a business’ API strategy in one of five stages:
- Ad-hoc: When APIs are created to support individual application use cases, with no connection with business-wide strategy or governance processes.
- Organize: When a business creates APIs to focus on enabling some popular use cases in order to generate organizational-wide buy-in for an API strategy.
- Tactical: As the API strategy succeeds, the step sees businesses accept APIs as digital products that strengthen the business’ “foothold in the digital ecosystem”.
- Critical: APIs are now so accepted in a business that the are the first go-to tool when creating new business integrations, including “mission critical services”.
- Industrial: Accenture sees this as the stage where APIs are “the fabric of business operations” and used to create a leadership position for the business in the digital ecosystem.
To identify how a business can develop along the stages from ad-hoc to industrial, Accenture argues that businesses need to manage five technological and business dimensions including strategy and governance; architecture; development process; developer community; and optimization.
Within each of these dimensions are key business tasks that will help an API strategy scale and progress along the maturity stages.
For example, in the dimension of architecture, a business starts to inventorize its data assets as it moves out of an ad-hoc stage, creates identity and access management rules as it reaches the tactical stage, hones in on caching and traffic management at the critical stage, and is able to offer highly dynamic, personalized SLAs at the industrial stage.
Tung encourages business to use the Accenture Industrialized API model to map out their strategy. “Strategy and optimization stages often receive the least-focus, but are ironically the most important stages for creating an industrialized API program. These stages lay the foundation for what APIs to create and their design and support, and evolution. These stages determine the complexity and cost needed, and how the API program defines success,” she says.
In part two of our look at API maturity, we examine techniques that can help businesses scale their APIs. We survey techniques that can help API providers move beyond ad hoc application activation and into using APIs at the tactical and critical stages of their business. Finally in part three, we look at how the API maturity model is fundamentally changing the application lifecycle management processes within enterprise.