Beware of the integration!
Enterprises have to constantly decide, at every step in their digital journey, should they build or buy. This question often is posed as a critical, do or die decision and the answer varies on a case by case basis. Building can be expensive, take longer but offers future proofing and more dependability whereas buying offers a faster time to market, less risk and accountability forced through contractual terms.
However, a key point often overlooked is the cost of integration. Integration can be required at multiple levels.
Vendor applications typically require a two-way connection between the enterprise systems and the vendor application. The application requires incoming data and information from somewhere in the enterprise technology stack and an output stream of information back into the enterprise at one or more points in the stack or workflow.
Vendor provided platforms typically have similar integration requirements as Vendor applications requiring an incoming data & information connection and an outgoing information connection into the enterprise process, workflow, platform or product.
Application to Application integrations where an application needs to be connected to another application to either provide data or signals to enable the downstream application to create value can be seemingly deceptive. Application-To-Application integration costs can grow at O(n^2) as potentially, worst case, each application could be connected with every other application.
Enterprise Stack Fragmentation
The problem of integration is exacerbated by the fragmentation of the enterprise at the organization level. This problem is also known as “Shadow IT” is driven by superficially differing needs of multiple lines of businesses in an enterprise. Shadow IT typically leads to multiple instances of similar technology stacks that cause data, compute and information to be silo’d. Stack fragmentation and its removal become a massive, integration cost for the enterprise.
One can argue that the above two types of integration costs are to be expected and are a result of the tradeoff between agility, time to market and independence. This is true however it also leads to enormous amounts of tech debt that in the long run, negates the advantages that it offered in the short term.
However, the clearest and present danger of integration in the Digital Journey comes from technology fragmentation. At no point in time, has there been an amalgamation of such fundamentally disruptive technologies. The pressure to act, react and innovate to capture competitive advantages is forcing enterprises to knowingly get deeper into the impending integration nightmare. Technologies such as Artificial Intelligence, Machine Learning, ChatBots, APIs, Analytics, BlockChain, Knowledge Graphs, Big Data, Mobile etc are front and center and enterprises have to decide not only what to build vs. buy vs. leverage but they have to do so while dealing with enterprise stack fragmentation. Often, multiple lines of businesses could be investing in the same technology at the same time creating products that are quite similar but independent and disjointed.
There is no good solution to the problem. Experimentation and agility is key to dominance in the digital market. Integration costs are real and need to be carefully analyzed and incorporated in go to market plans. In fact, this exact problem led to the creation of LearnCloud where these technologies are fit together with minimal integration cost for the enterprise. Regardless of how integration costs are controlled and minimized, it is key that they are thought about upfront because they are real and only add up as time goes by.