Gartner SOA summit – pragmatism and innovation

I spent a great couple of days last week at the Gartner SOA and Application Development and Integration Summit in London, emerging encouraged by the combination of pragmatism and innovation that I heard discussed.

Inevitably, much discussion centred on the economy and its implications. While technology –as Massimo Pezzini pointed out - might be seen as having helped the problem happen by “providing the infrastructure to enable the financial services sector to get into trouble faster than anyone expected”, I didn’t hear much discussion about what could (or should) be done to prevent that happening again! On the other hand there was lots of debate on how IT can help companies survive and reinvent themselves for the future. The (not surprising) prescription was to focus on strict cost control and alignment of IT projects with clear business cases, while keeping sufficient investment to foster innovation in key areas. In the short term for example, organizations need to use their resources to rapidly adapt sales and marketing patterns in response to changed customer behaviours. Longer term, much discussion centred on anticipated structural changes to industries (especially financial services) requiring better support for complex inter-company relationships and networks.

SOA adoption
The message I took on SOA – from both Paolo Malinverno’s opening keynote and much that followed - is that it has clearly taken root, especially in Europe and N. America. However there also seems no doubt that the recession is having a Darwinian effect, culling the projects without well-defined short term business benefits. Long term strategic business cases for SOA programmes are “out”, pragmatism in the form of prioritizing rapid payback from fixing business problems is “in”. As a result, lots of initiatives that have started bottom up with no focus on business problems (service enablement rather than SOA) risk being cut – one figure floated (and hotly disputed) was that this could kill up to 80% of the projects started thus far!

Enterprise mashups?
Given this background, where are enterprise mashups? Well there was certainly plenty of interest, with David Gootzit from Gartner talking about their having a key role to play as the “face of SOA” and becoming an important way to build composite applications. However, listening to talks and speaking with attendees confirmed to me that while a lot of vendors and users have focused on using enterprise mahups to build simple “dashboard” type single page applications, there is a huge need and interest in using them to address the new integration agenda of business focused pragmatism and innovation. The idea of using the mashup approach to rapidly build process based applications for different user groups and tasks, leveraging a wide range of enterprise IT assets is – to judge from the response we had – just the capability that many organizations need. It seems that there are many cases where business process improvements and flexibility are needed but custom application development is seen as too slow, expensive and hard to maintain, and full blown WS-* and BPM are not appropriate.

Conclusion? SOA isn’t dead, just growing up in a rather more austere environment than the one it was borne into. And enterprise mashups can make a massive contribution, so long as they focus on really changing the economics of building and using business applications.

What’s the call centre industry telling us?

Given that Corizon works with many customer service organizations to streamline and simplify the use of applications for agents and customers, we take a keen interest in what the industry is saying and how it is responding to the economic environment. One organization that collects and analyses this kind of information in the UK is the Customer Contact Association, and I thought some of the findings they presented at a briefing this week worth sharing more broadly, along with relevant insights from our own work with customers.

Impact of the recession

When asked about changes being made in response to the recession, the results from organizations are interesting. It seems that they are not rushing to outsource calls and contacts, but are more likely now to outsource back office processes. However, by far the biggest shift is in consolidation, accelerating moved from multiple smaller sites to fewer, large and more economic centres. Based on our experience, the former is perhaps an encouraging sign for the contact centre, evidence of recognition that they are not just cost centres, but can be the key to keeping customers happy and buying more – both essential success factors in the current climate. The latter clearly makes economic sense, but gaining even greater economies of scale will require an increase in mutliskilling, something that has been a problem in many of the call centres we visit due to the learning associated with different roles and the IT systems they require.

Role of technology

The demand for technology solutions seems to be running strong. 58% of those surveyed by CCA felt that performance was being hampered by out of date legacy systems, while 47% felt that new technology would be a better investment for improving performance than training and upskilling their teams; 57% said that they would invest in unified desktop solutions.

However, there is also a sting in the tail for technology vendors, with about half the respondents feeling they repeatedly had to fight off attempts to sell them unnecessary technology, and a significant number demanding that new applications are created with more user input than happens at the moment. While the first point probably needs no comment, the latter point resonates strongly with our experience. It seems obvious that adoption, productivity and process improvements all strongly correlate with making sure applications are actually fit for purpose and built around the users, but the evidence suggests it is too often forgotten.

It was also intriguing (not to mention counter to many stereotypes) to hear that the CCA found that women managers are more interested in adopting new technologies, and develop better relationships with technology vendors than their male counterparts.

Moving to multiple channels

Perhaps not surprisingly, the need to embrace and move to new channels is prominent in CCA’s research. 88% of respondents saw increasing self-service on the web as a key enabler of cost reduction in response to the downturn. In principle, this frees agents for more complex, “value added” tasks, but the reality at the moment seems to be that the volume of inbound calls is not reducing. Why this should be is not clear, but there is increased awareness that integration across channels is essential if the result is not to be increased, rather than reduced frustration.

Finally, although there is obviously awareness that social networking and social media represent an important shift in the environment and are impacting service providers, it seems that the full implications are still being digested before conclusions are drawn.

Interesting times

Conclusion? It seems that if anything the recession is crystallising the debate on the role and value of the contact centre, self-service and the technologies that underpin them, and is driving change in interesting directions.

Are mashups lipstick on a pig?

One of the debates currently going on is regarding the lifespan and role of mashups (cpettey/Gartner). Are mashups just tactical, throw away applications for short lived micro-requirements that come and go in the business (aka situational requirements) with the associated business case challenges? Are they just good to temporarily paper over a gap while your systems catch up with user requirements? In other words, are enterprise mashups short lived either for the simplest of “applications” or simply as lipstick on a pig as suggested to me recently by the CIO of a very large telco?

One of the great values of enterprise mashups is that they can deliver in very quick time frames by leveraging exiting systems and applications to streamline processes for people. This can be to address a “situational requirement” such as a dashboard, combining RSS feeds for data reporting etc. They can also help mop up legacy applications before retirement i.e. provide lipstick on your old architecture while you deliver the new one. They can prepare for the deployment of packaged application such as CRM or ERP. This provides the value of quickly testing in real time, with real users the required “user process” (i.e. the set of steps a user need to go through to complete a task). This can even be implemented in the look and feel of the enterprise application. Users will then not be disrupted by the deployment of the enterprise app while deployment will be speeded up by “learning” from the tested user process implemented in the mashup. Once the enterprise application catches up, then the mashup can evolve to do less, be retired itself or remain to provide the flexible, process driven UI that would continuously require updating. This will depend on architectural decision.

It is in the flexible delivery of process driven UI that enterprise mashup become a key part of modern architectures and are destined to power solutions that will remain in use for a long time. Why? Because modern architectures need lipstick as much as the old ones! With WOA, SOA or any component based approach, user-centric applications are still required when a business process need surfacing to a user and those end up being built bespoke for each user group. This is because user interfaces are not service enabled, not turned into re-usable components. This is where enterprise mashups play a key part: with the creation of visual, mashable components that can easily and quickly be mashed up into user centric solutions, they provide a “face” to web services and deliver the flexibility and process adaptation required at the user level to deliver flexible architectures from top to bottom.

The answer to the lifespan and role is that enterprise mashups provide real value when they address real process pain point, have a role to play in legacy and modern architectures and can endure for considerable periods of time while the content and function continually evolve.

What do you think? Do you see the same trends?

Enterprise Mashups and Cloud Applications

Corizon’s recent launch of mashup extension packs for Siebel and eBusiness suite has prompted a number of conversations with journalists on the extent to which this value proposition transferred to enterprises adopting SAAS applications. I was surprised by the initial perception that enterprise mashups were required because of the difficulty of deploying applications on premise. Conversely there seem to be preconceptions that the need would be much less with on demand applications.

Enterprise mashup adoption is driven by the need to transform the economics of integrating applications to make people more efficient. The deployment of new applications, whether they are on-premise or in the cloud,increases the requirements for integration. The more new SAAS business applications are adopted, the more such integration will be required. This is exactly the experience reported by GE CIO Gary Reiner: implementation costs for applications:

“...they're largely around interfacing with existing systems, process changes and data cleansing, those three costs exist regardless of whether GE hosts that application or whether the supplier hosts that application.”

As a matter of fact, cloud computing is even more adapted to enterprise mashups. The key reason for this is that they promote the concept of keeping each functional application as vanilla as possible to reduce costs. Why re-invent a CRM package and not just adopt best industry practice? Differentiation is not gained by modifying each application but by integrating them to streamline your desired processes. For example superior end to end customer experience delivery is more likely to be achieved by seamlessly integrating the different applications required to support a particular customer journey (trouble ticketing, service diagnostics etc.) rather than customizing each application. This is where enterprise mashups play a critical role: by providing a quick and non intrusive way to create mashable components from existing applications and easily combining them to mashup a fit for purpose solutions that guide users to complete a task, they are perfectly suited to integrate your cloud with your on premise applications.

We are big proponent of service based applications at Corizon – this is where the company started! We also happen to believe they will be one of the key drivers to accelerate adoption of enterprise mashups. As enterprises adopt SAAS, they will be looking for new approaches to quickly, flexibly and cost effectively integrate their new cloud applications to allow people to do their jobs better. In the same way that enterprise mashups transform the economics of owning on premise enterprise applications, they will transform the economics of integrating cloud applications.

Funnel shaped mashups

An interesting conversation on freedom vs control at the desktop in call centres, in conjunction with much recent speculation as to the "shape" of the economic downturn and recovery, made me wonder what “shape” a mashup solution needs to be, and what that would tell us about its design and delivery.

In a situation like a call centre you give the agent a job to do because you want him or her to do two types of things within a call which can appear quite inconsistent. First there is the need to analyse a situation based on “conversational” interactions with the customer and the systems on the desktop. This involves relatively unstructured interactions with the applications and customer, drilling into different information sources and exploring options with the customer. However, as quickly as possibly, a path needs to be chosen. The agent then must switch mode and move into a much more structured, regular approach to ensure no mistakes are made, all details are captured, and frequently that regulatory compliance requirements are met.

So it seems to me that for these types of activities, mashups are funnel shaped! The cone represents the user working in an unstructured, conversational way before moving into the spout and being constrained and guided through a defined process. So, if enterprise mashups are to support this type of interaction, then they need to support both modes of working in a single solution, and to get through the spout as quickly and efficiently as possible.

Taking the analogy further, not all funnels need to be the same shape. In fact many factors can determine the shape of funnel needed. For example

• experienced agents need to be able to work in a less constrained way compared to rookies, and will find constraints demotivating and frustrating
• new products are often trialled and launched with very broad cones while the organization learns using highly skilled agents. As the products mature and the agent base grows, processes then need to become more constrained and more “spout heavy”.
• knowledge management solutions can be used to narrow the cone sooner by reducing the up-front analysis work for service desktops
• sophisticated propensity-to-buy and segmentation based intelligence can help to move to the spout quicker and optimize outcomes

One thing is clear though, whatever funnel shape you start with, being able to change it and experiment is essential. The shape of the funnel is a key “lever” for call centre managers as they struggle to find the optimum balance between allowing initiative and empowerment and enforcing process repeatability. That’s where the flexibility and control provided by enterprise mashups are invaluable.

Integration without customization for Oracle E-Business Suite

I repeatedly hear from our customers about the challenges they face in creating integrated user experiences around their investments in CRM and ERP applications.  They find themselves constrained by long release cycles, full workstacks and the expense and scarcity of skilled resources.  As a result, users have to carry on “swiveling” between applications on the desktop – bad news for the reputation of the applications and for efficiency, accuracy and costs.

Fixing this problem is a great use case for enterprise mashups, replacing complex customization projects with the much more lightweight and portable mashup process.  We call it “integration without customization”.  A while ago we announced a toolkit to be able to easily embed our mashups in the user workflow of Siebel applications; this week we have followed this up with our “extension pack” for Oracle E-Business Suite.  The pack overcomes the challenge of embedding web UI in the EBS forms UI.  By providing everything an EBS developer would need to be able to get a mashup deployed and launched within the EBS application, it cuts a task that could take many days into hours.   As a result EBS owners can extend and integrate user desktops with minimal cost and effort.

We believe integration without customisation is a killer “mashup pattern”, especially in a economy where the focus is on getting more from existing investments like major enterprise applications.  Through the rest of 2009 we will be working with our customers and partners to understand where else pre-built capabilities of this sort can smooth the process, so let us know where you see the clearest need.

Contact centres are ripe for enterprise mashups

Spending time recently in a contact centre reminded me of a quip I heard from Lord Saatchi. He was talking to a senior director of a well know consumer brand and saying: “do you know the old saying about half of your marketing budget does not deliver but the issue is which half? Well for your company, I have worked it out: neither!”

So which half of your IT budget works in the call centre? Neither. You just have to look at how many applications call centre agents have to deal with during a call! Research shows on average 6 but in our experience the number is usually much higher. This is despite years of investment in technology in call centres. The reason for the failure? Each time a new requirement for the agent emerged, a new application would be built or an existing one would be modified. This lead to a proliferation of applications on the desktop, making the job of maintaining them (and integrating them) harder and harder for IT. No wonder release cycle of enterprise applications on the desktop can now be measure in months or years! This results in inefficient processes with dramatic economic consequences: high cost to serve customers while delivering, in a majority of cases, a poor customer experience. No wonder so many companies decided to give up on call centres and outsourced them despite the fact they are one of the most important (and inescapable) windows to customers.

This is why call centre is such a ripe environment for the deployment of enterprise mashups. By changing the economics of building and maintaining integrated solutions to call centre agents, they address the application proliferation issue, reduce dramatically the cost to serve customers by making agents more efficient and deliver clearly measurable ROI in short time frames. This is where we think the real value of enterprise mashup can be realized: in delivering real composite applications to address immediate business needs but in a way that is repeatable and complements modern architectures. A theme I will come back to regularly in this blog.