Why Do We Need a Next Generation of Analysis and Design?
When the first Analysis and Design methodologies were invented in the 80s (SA/SD in the US, SSADM in the UK and Merise in France), they were intended to support the building of business applications which would last for a long time with minimum cost of maintenance. They were to ensure designs which would cope with the functional (the what) and non-functional (volume and performance) system requirements, which were supposed to stay as stable as possible. Data would be organized to represent required information according to well-defined data models which optimized data access and reduced required storage volumes. The famous relational data model imposed integrity constraints which were maintained on a permanent basis; and transactions were carefully controlled in order to avoid inconsistency in the event of hardware failure or performance bottleneck.
Analysis and Design delivered robust and maintainable applications where consistency and performance were built around invariants identified and defined by the analyst. Analysts expended significant time and effort to build information system that was resistant to change and for which success was measured by the length of their lifecycle, conformance to requirements, and delivery on time and on budget.
Analysts were creating digital objects in the great engineering tradition of building bridges, cars or aircrafts. In the world of engineers, investment in time delivered value is creating long-lasting assets. This mentality was replicated in the Business-to-Business world where business processes were stable and where companies pursued long term objectives that were to be managed through their information systems.
The digital era which started with the advent of internet brought a whole different perspective on time to the development, adoption and evolution of IT applications.
Business to Consumer applications emerged on PCs in the 2000s, and then on smartphones in the 2010s. Suddenly IT was not constrained to bringing progress through a well-defined set of objectives, as it had been previously for business systems. Instead it was offering successive waves of innovation to consumers, provided of course that consumers embraced their adoption.
The world of Digital is a Darwinian one where you need to start small, be adopted by consumers and grow fast, or else be rejected. Perhaps the most significant success stories are epitomized by the rapid success of the GAFA,built around the digital phenomenon ofdata platforms.Platforms embody one of the most significant business model disruptions of the Digital Era.
Google, Amazon, Uber, and Airbnb have all adopted platform principles to support their disruptive business modelsmodels which have led to a new type of market that shares common core characteristics with the payment card ecosystem.
With the advent of payment cards, a new market type emerged which has been termed multisided by economists. In the specific case of cards, we have two sides: (1) the consumer who is the cardholder and (2) the merchant which is offering payment services via a point of sale terminal capable of reading the card and capturing transaction details.
This economic model was formulated by French economist Jean Tirole, an invention for which he was awarded the Nobel prize for economy.
In a nutshell, a multi-sided market can only really take off once each side reaches a critical size; in the payment card example, no one wants to carry a particular card if it is not widely accepted, and no merchant would be willing to invest in a point of sale terminal if there were not many consumers who would use it.
This chicken and egg stalemate can be broken if a market player is able to offer a platform to both sides of the ecosystem with the commitment to pay very little, if anything, for the use of the platform until critical size is reached. It is the platform that is the tool which enables multisided markets.
In the example of Amazon, the platform is successful for customers because they have a single route to find just about everything they need; it is successful for merchants because they enjoy access to many more potential customers than they would without having their presence on Amazon. In the Airbnb model, the two sides are consumers and hotels; while with Uber, they are consumers and drivers.
It can be difficult to reach critical size on each side of the platform, but once achieved, multisided markets will naturally tend to grow to a monopoly position, driving the need for regulation to maintain competition.
In term of systems analysis and design, we need to distinguish between building the platform and building applications for the platform.
The unprecedented challenge of launching a platform is that of being able to effectively deliver services when its number of users is small on each side, and their trading volumes are low. The initial cost to users of participating in the platform needs to reflect low startup value, even though this means that the platform operator might have to run at a loss until user revenues build up. But probably even more challenging is the problem of being able to maintain the same architecture for the platform while the two sides are growing from zero to critical size.