In every age, we like to think we have unique views on innovation and technology adoption. Yet each era's outlook seems to build on what's come before. In 1991, during the emergence of startup culture, Geoffrey Moore wrote an important book called Crossing the Chasm, which tried to describe the high-tech purchasing behavior of companies. He specifically identified the difficulty immature technologies (startups) have transitioning to mainstream applicability. His goal was to provide a method to market and sell technologies to these companies in a differentiated way.
This work built on some early 1960s research by Everett Rogers. His 1962 book Diffusion of Innovations built upon the general diffusion theory of the late 19th century.
The Diffusion of Innovations curve illustrates how enterprises adopt technology and provides a greater understanding of enterprise architecture's functional and non-functional requirements.
One of the least recognized aspects of the Diffusion of Innovation curve is that it represents both a company's requirements and the technology it chooses to adopt. When a technology exists in a phase of the curve that matches the company's strategy, then the chance of successful adoption is highest.
Companies do not exist only at one point on Rogers' curve. One part of an industry or a company might exist as a Late Majority or Laggard, while another part of the industry or company could easily be an Innovator or Early Adopter. This trend drives much of the disruption of industries and the process of transformation within large enterprises. The finance industry is one of the most notable examples in recent years. The creation of electronic banking systems, trading systems, international currency transfer, online banking, credit transactions, and cryptocurrencies are a series of innovations that have occurred through a combination of startup innovation and innovative business units within existing companies. I think the healthcare industry is the most prominent industry dealing with this today.
As an enterprise architect, it's important to know what level of architecture discipline is effective for which business goal. Applying traditional waterfall architectural principles to an "innovation" business strategy results in frustration and delays, or worse, a solution that does not meet the market's needs or pressure.
A master of any method or discipline is one who knows when and how much of it to use in each situation. In some cases, no single enterprise architecture methodology is the best approach. Suppose we map the Diffusion of Innovation curve categories with the adoption approach each generally reflects (as outlined below). In this case, we can translate that to the likely technology solutions they require and the type of architecture method they should use.
According to Rogers, Innovators are risk-takers and have the most contact with scientific sources and other Innovators.
- Highest risk-takers
- Don't need proof of value
- Intrinsic to business strategy
- Wish to be close to science and engineering technology
- Strive to keep capability internal
- More likely to use innovation to invest in people over product (Marten Mickos of MySQL fame coined the phrase "more time than money or more money than time" to represent companies' will to invest the time to get a capability versus buying a solution.)
- Upstream open source
- Startup companies with engineering resources primarily in early-stage funding
- Engineering partnership relationships are often not commercial
- In-house developed solutions
- Custom production deployment
- Unlikely to be a formal architecture methodology
- Agile technology adoption
- Evaluate and implement in parallel (feedback loop)
- Business benefits are more intangible than strategic
- Isolate risk (technical and financial) through small-start isolation
Rogers says Early Adopters have more financial liquidity, education, social status, and opinion leadership, but they are more discreet in their adoption choices.
- Similar knowledge and capabilities as Innovators
- More discerning in choices
- Measure risk against some quantification of benefits—likely to try and reduce risk through acquisition strategy
- Often seen as teams in Late Majority or Laggard companies
- Success is what often causes a company to transition from Innovator to Early Adopter. This rapid growth is primarily responsible for the "bozo explosion," which is the inability to keep hiring A-players and to trust execution.
- Startup products in late-stage funding rounds
- Commercial technology relationship
- Enable self-service technical capability
- Packaged technology
- Shadow IT facilitates innovation
- Enterprise architecture dispersed
- Central influence (add requirements) but not control
- Small implementations to prove concepts
- Document requirements and architecture before production implementation
The Early Majority take longer to adopt innovations and have average social status and contact with Early Adopters, says Rogers.
- Opinion leadership is less important
- Also called fast followers
- Proven technologies and methods for adoption
- Companies that fit in the Early Majority category often fit into Gartner's bimodal definition and have aspects of Innovators and Late Majority
- Business justifications focus on qualitative or strategic value
- Less concern for cost-benefit and ROI
- Proven distribution or productization
- Large vendor support
- Second generation
- Reference architecture
- Customer-referenceable proof of value
- Equal focus on non-functional requirements
- Transforming enterprise architecture
- More agile (recursive) and adapts to business needs
- May still have a central process but very focused on business relationships
- More "request for information" than "request for proposals" or formal proof of concept—still use minimum-viable-product-based requirements for tech selection
- Process needs a tighter feedback loop for requirements
The Late Majority are skeptical of innovations and adopt them after the majority of society, says Rogers. They also have less financial liquidity, social status, or contact with others.
- Crosses the median point of a standard distribution
- Company or group is skeptical
- Moving to cover the risk
- Need a solid business case and proof of value
- ROI calculations are a feature of the business justification
- Third major version/generation
- Service-enabled deployments
- Structured training and education
- Traditional enterprise architecture processes
- Methodology is the primary driver of technology adoption
- Multi-phase technology evaluation, technical security, financial, etc.
- Formal request of offer and contractual implementation
The Laggards adopt an innovation last. They're more traditional, adverse to change, and older, and have little social status or contact with people outside their family and close associates, says Rogers.
- Driven by process and tradition
- Technology becomes commoditized
- Many alternative solution providers exist
- Acquisitions are often driven by purchasing or financial metrics (low-cost provider)
- Mid-lifecycle technology
- Simple install/configuration
- Established best practices for non-functional requirements
- Integration with legacy operations capabilities
- Multivendor distribution and support
- Traditional enterprise architecture
- Strong central decisions
- Paper analysis with the business case and implementation strategy precedes funding
[ A free guide from Red Hat: The automation architect's handbook. ]
The accelerating need to change architecture methods
In the 1960s, technology development might have progressed in a timeline similar to the Diffusion of Innovations curve. As innovation moved from hardware to software at the beginning of the 21st century, innovation no longer progressed in a linear fashion. Gartner created the Hype cycle, which drove its research and analysis across a number of technology spaces and industry groups. This concept, illustrated below, has stood the test of time and is complementary to the Diffusion of Innovations curve.
Innovators and Early Adopters deploy primarily in the Technology Trigger phase. As market awareness peaks, the Early Majority start to include adoption in their short-term plans. This continues through the Trough of Disillusionment (or the Chasm) because the Early Majority requires a much larger focus on non-functional requirements ("ilities"). The inclusion of non-functional capabilities is a large engineering effort that takes away from the focus on innovation. Late Majority and Laggards begin to adopt during the Slope of Enlightenment and Plateau of Productivity phases.
How open source affects the curves
As Marc Andressen said in 2011, "software is eating the world." Much of the innovation that occurs today is in software, and much of that software development now happens in open source. A mature open source model serves to accelerate the movement of software through the Technology Trigger phase to the Slope of Enlightenment. It facilitates a mechanism for the Innovators to have a direct relationship with engineering, while in parallel allowing Early Adopters and Early Majority access to a more tested distribution. A single open source project can fulfill the needs of customers across the phases of Diffusion of Innovation and can exist in multiple places on the Hype Cycle.
The appropriate use of architecture methodology is critical to a successful implementation that supports business for the short and long term. The entry point to adopting technology is no longer a single "architecture method." Modern architects must adapt their methods to meet the desired adoption approach.
As more development occurs in open source software, and with infrastructure easily available on the hybrid cloud, an Innovators approach is appropriate for every enterprise in every market. The Diffusion of Innovations curve gives us a framework to decide how to adapt as architects.