In a short video, Geoffrey Moore describes the focus evolution from Systems of Records to Systems of Engagement. At the end he highlights the fact systems of engagement probably require a very different type of IT than systems of records.
Systems of record host the key processes and data elements of the enterprise. Most often they have been implemented prior to the year 2000 to ensure the enterprise could survive in the new millennium. Great efforts and vast amount of money went into implementing them and adapt the enterprise to the software and/or the software to the enterprise.
Since they run reliably and support enterprise operations.
But a couple things have happened. Users, getting acquainted having information at their fingertips through smart phones and other mobile devices, are asking to be given the same access capabilities to the systems of records. New interaction mechanisms, such as social media, new sources of information increasingly allow a better understanding of market needs, customer demand and the environment in which the enterprise operates.
In parallel, time has been shrinking. What we accepted 10 or 15 years ago to take several hours or days now needs to be done in a matter of minutes, pushing traditional IT systems to their limit. Hence the new style of IT as Meg Whitman calls it, required to respond to user needs. Cloud is definitely part of this transformation of IT as it provides enterprises with the responsiveness and agility required to address today’s ever changing business environment.
As enterprises decide to move to the cloud, the question where to start typically comes up. In a previous blog entry I spoke about five use cases companies could envisage to start on their cloud journey. It ultimately turns around deciding to consume services from a cloud environment (being it private, managed or public). So, the question which application to move to the cloud first should be raised? Should we start with a systems of record or a systems of engagement one?
Should we start with Systems of Record?
As stated by Geoffrey Moore, most Systems of Record have been in place for the last 10/15 years. They are transaction based and focus on facts, dates and commitments. They often represent a single source of the truth. Indeed, they typically have been built around a single database, containing mostly structured information. Every time information is changed, the event is logged, so one can quickly find out who did what. Data is kept in the systems for extensive periods of time to ensure compliance, and access is regulated and contained. They are the core support systems for the operations of the enterprise.
A couple companies focused on the development of such systems, the most known being SAP and Oracle for their financial and manufacturing systems. Other enterprises may have written their own application, being left with a small team of knowledgeable resources to maintain it. These systems are considered business critical as the company cannot operate without them any longer.
Some enterprises may require increased flexibility in the use of such systems; they may want increased agility and responsiveness in case of a merger or divestiture. But are these the systems we should migrate to cloud first?
What would be the benefit? Well we could probably run them cheaper; we may be able to give our users that additional level of responsiveness, agility and flexibility they are looking for. But on the other hand, we would have to modernize an environment that runs well and supports the business of the enterprise on a daily basis. Or we should rebuild a brand new system of record based on the latest version of the software. This may be the only option, but to me, that sounds risky.
I’ve seen a couple companies doing it, but this was mostly in case of a merger, a divestiture, a consolidation of systems or a move to a new datacenter or IT delivery mechanism. And in most of those cases, it turned out to be the installation of the application in a cloud provisioned IT environment, not fully taking into account the automation, the flex-up/flex-down and self-service capabilities available with cloud.
Now, employees may want to be able to access the systems of record through mobile devices, they may want a more friendly user interface, they may want to combine functionality that is separate in the original system. This is a whole different ballgame.
Using web services, we could encapsulate the system of record and give the user what they want without having to disrupt the original environment. Over time we could consider updating/transform some of the functionality, shut them down in the original package and replace them with cloud based functionality. This reduces the risk and shields the end-user from the actual package, making it easier, when time is ripe, to transform the system of record without disrupting the end-user.
What is different with Systems of Engagement?
Systems of engagements have been developed later, so they involve newer technologies. In particular, many of them are built around SOA principles, making them more suitable to take full advantage of cloud technology. Their objectives are interactions & collaboration. It’s all about sharing insights, ideas, nuances. They are used within the frame of business opportunities and projects, so the relationships are transient in nature, requiring responsiveness and agility to be set-up quickly when required. Access is ad-hoc and in many companies may require the interactions with partners and/ or customers. Most often, information is unstructured, making search more difficult.
Obviously, Systems of Engagement are important, but they do not maintain the critical information to run the company.
They are important as a mechanism to share information, gain consensus and make decision, however they do not maintain the single source of the truth. That makes them more suitable for being experimented with. Their nature, their needs and the technologies used to build them, makes them better candidates to be migrated to cloud. So, I would suggest this is where we should start. Obviously, we don’t want our end-users to be left in the cold if something happens during the migration, but worse case, the telephone still allows people to exchange information if the system were down for some time.
The importance of data
Tony Byrne argues that Geoffrey Moore simplifies things by creating two clearly different categories. He points out the reality is probably messier. On the one hand, people are discussing important business decisions in collaboration systems, thereby creating records, while others may want to engage with their colleagues directly from the systems of record. He puts it in simple words: “your colleagues are creating records while they engage, and seeking to engage while they manage formal documents and participate in structured processes. Ditto for your interactions with customers and other partners beyond your firewall.”
Now, we have been able to trigger functionality from within applications for quite a while, so that’s not the issue. And the use of web services described earlier makes this reasonably easy to implement.
What is at the center of Tony’s discussion here is the data and the capability to move data between the systems of record and the systems of engagement. Right from the start, you should think about your data sources and think about information management. Again, technology exists today to access data within and outside a cloud environment. What is important is to figure out what data should be used when and where, and ensure it is properly managed all along the way. If you access and change data in a systems of record, do it in such a way that all the checking, security and logging functionality is respected. But this should be nothing new. Companies have been integrating external functionality within their systems of record for years.
When companies look at migrating to cloud, the question where to begin is often debated. In my mind, it’s important to show end-users the benefits of the cloud early on. That lends me to lean towards starting with systems of engagements, either by transforming existing ones or build new ones that will surprise the users positively. This will get their buy-in and give IT more “cloud” to transform the remainder of the IT environment. The whole discussion is how far you need to go. Because not everything has to be in the cloud at the end of the day, only what makes sense.
Links that may be of interest: