The SaaS Revolution
The wide availability of pay-as-you-grow SaaS applications have revolutionized the way IT is delivered and has lowered the barriers to entry for entrepreneurs all over the world, and that is a great thing! Gone are the days of setting up your own data center, purchasing expensive servers to host your company’s applications and then hiring a large IT Infrastructure team to support it.
In today’s word you can provision a SaaS application for your company in minutes and at extremely low or no cost to start with. As your company grows the solution grows with you. All of this is incredibly good news for businesses of all sizes, but the age-old IT dilemma remains true … how do we make all this stuff work together?
The Integration Dilemma
The marketing hype will tell you that you need not worry about integration because these new breed of applications come with out-of-the-box API-based connectors for integration and we all know API’s are magic … right?
Well, the answer isn’t that simple. Yes, API’s absolutely make connecting your portfolio of SaaS applications easier. However, this is like creating a new way of quickly and easily connecting the plumbing in your new building. It is a great improvement because it reduces the cost of the plumbing but here is the real problem. If your water is dirty, you have just made it very fast and easy to disseminate your dirty water to every part of your building. That is when “SaaS Happens”.
Better methods of IT plumbing do not provide you with better data, good Information Architecture does.
Cloud-Based ESBs Help but Don’t Solve the Fundamental Issue
One of the great things about today’s economy is how fast new solutions can be brought to market when there is a need. Many companies have seen the need to orchestrate the integration of a heterogeneous, best-of-breed portfolio of SaaS applications. These new cloud-based Enterprise Service Bus (ESB) solutions take API’s to the next level and when implemented properly do get companies closer to the “Plug-N-Play” panacea that everyone dreams of.
But again, this is just improvements on the plumbing. There is value in having that kind of capability, but it too does not solve the fundamental dilemma of disseminating bad data, good Information Architecture does.
Information Architecture is Still Critical
I remember a large presentation I was part of where a new Marketing Dashboard was being demonstrated. It was a thing of beauty and the various marketing leaders across the company were drooling in anticipation of putting it to use. Then they were asked to focus on the roll-up of the data by product offering and it was clearly not good at all. The root cause was easy to identify, bad information architecture. There was no standard in place for the product taxonomy that was used consistently across the systems of engagement. So, even though the IT plumbing had been put in place to pull the data together and a shiny new dashboard had been built, it only served to make it fast and easy for everyone to see how bad the data was.
No matter how much the technology improves, the fundamental need for good Information Architecture will always remain. The challenge is often that talking about data standards, common taxonomies and the enterprise governance to enforce them is not an exciting conversation. However, in the absence of those things your “exciting” conversations will be explaining why the data looks wrong and is not trustworthy.
The good news is that developing a good Information Architecture is not expensive. Further, enforcing it at the start of technology deployment is not expensive and can actually reduce costs if there are reusable assets for managing and consuming approved taxonomies. Unfortunately, cleaning up the mess after a large technology deployment done without good Information Architecture can be very expensive.
Need Strong Data Ops at the Edge
Finally, even after getting good integration plumbing in place and strong information architecture there is still room to improve the quality of your data by adding sophisticated Data Operations monitoring. The items we have discussed so far will ensure that your data is properly connected and using common taxonomies so it can be aggregated, but what happens when something goes awry? No one likes getting the call “I haven’t seen any new leads for product X in the last 3 weeks”.
It is also important that you put intelligent monitoring in place that can detect large fluctuations in transaction volumes flowing across your integration architecture. Think of it like fraud alerts you get from your credit card company. They alert you when they see a transaction that deviates from your normal purchasing pattern. Similarly, if you normally see 50 leads per day for product X and it goes to zero for several days, it is probably worth someone investigating.
To Sum Up
Yes, the IT world is a better place today with the ability to easily stand up pay-as-you-grow SaaS solutions that integrate much easier with other applications, but these improvements only lower the cost of the IT “plumbing” and do not eliminate the need for strong Information Architecture.