Test and Learn - When and How Should you Apply a More Experimental Approach to Process Improvement?

Add bookmark

Nigel Warren

Light map
In last month's column, we defined the term Low-code Development Platform with the help of Clay Richardson, Principal Analyst at Forrester Research, who I interviewed on the topic recently.
We compared Low-code Development Platforms to more IT-centric BPM Suite products, and deduced the following two takeaways:

The Key Takeaways Re: Low-code platforms...

They enable faster development via graphical configuration instead of programming.
They lower the learning curve and reduce the IT burden for set-up and solution development.
As Clay Richardson put it in the interview:
"What’s different now with Low-code is it’s about fast-ramp. It’s the ability to train up very quickly. People are saying we can… train people and have them productive in a matter of a week as opposed to 9 to 12 months."
This 9 to 12 months of investment in training and professional services is typical for IT-centric BPM Suites, and is a real barrier for adoption. Reducing the learning curve by adopting a Low-code approach for some of your BPM type projects has two main consequences:

#1 Lowering the Entry Bar

Most obviously, you can start projects faster with less up-front investment in training. This lowers the entry bar, enabling you to build a business case for a wider variety of projects, some which were previously considered too small to deserve investment.
If you look around your business and find people putting up with manual procedures, email instead of workflow, copying data between systems and spreadsheets etc. you have to ask the question: "Why do people put up with this?"
These kinds of chronic process problems are considered too low priority to get fixed, because people can work around the problem. Such inefficiencies persist, because the established in-house technology tools that can be used to resolve such issues are too technical and therefore in the hands of just a few highly trained IT practitioners.
In many organizations this results in a 6-9 months IT queue. Not only does this make it difficult for IT to prioritize improvements for the kinds of low-level problems that people can workaround, it breeds apathy on the part of staff and line managers who have often given up asking for improvements, as a 6-month delay makes the daily kludges and workarounds seem like a cost worth paying.
Having a Low-code alternative for process automation means that more of these kinds of chronic issues can get addressed.

#2 Empower Less Technically Skilled Developers

Who’s involved with BPM style automation at your company? Typically, it’s a small cadre of highly trained developers. You just need to look at the recruitment pages of leading BPM Suite vendors to see the problem.
"Junior Application Developer – Required experience: 2-3 years in Java-centric application development, experience of J2EE, Oracle, DB2 and SQL."
These requirements reveal that many BPM Suites are still highly technical products, meaning business stakeholders and IT specialists still struggle to collaborate throughout the process improvement lifecycle. Business departments, and even IT departments, end up outsourcing development on BPM Suites to system integrators. The learning curve is simply too long and too steep to achieve the required expertise in-house.
By comparison, Low-code Development Platforms lower the learning curve so that less expert developers can get involved. In some businesses and public sector bodies, that’s particularly important as they struggle to recruit and retain highly skilled developers. As Low-code platforms reduce or eliminate the need for coding, they can alleviate such resource constraints.

Where to Focus early Low-code Investment?

So, if Low-code Development Platforms offer faster development and "fast-ramp", where should you consider using them instead of incumbent BPM Suites?
In this regard, Gartner’s concept of Bi-Modal IT is helpful. This proposes that any organization should have two separate modes of IT delivery: (1) focused on stability; and (2) on agility. Mode 1 is traditional and sequential, emphasizing safety and accuracy. Mode 2 is exploratory and nonlinear, emphasizing agility and speed.
Clearly a Low-code approach is more appropriate to Mode 2, where businesses value speed and agility over stability and scalability. Moreover, as defined by Forrester in theirAge Of The Customer research, speed and agility are at a premium when you seek to improve the ways in which you "win, serve and retain customers" – what Forrester calls "Systems of Engagement".
Simply put, if your company is haemorrhaging customer goodwill due to poor service levels, you need to stem the flow far more quickly than you might care to fix an efficiency issue. Each lost customer may be far more damaging to your business than shaving a fraction off your operating cost.
Here are just a few examples of customer-centric process improvement that we’ve seen delivered in a matter of days or weeks using a Low-code approach:
  • A complaints case management system deployed to around 800 holiday reps in remote resorts – that has dramatically cut complaint handling time and associated costs.
  • Local council citizen service requests – rapid development of common and reusable processes that entail, citizens, places and services.
  • Grant submission and approval workflows in central government.
  • Bereavement case management in a financial services organization – that ensures a relation only needs to report the family member’s death once even if there were multiple accounts and products involved.
  • Phone call prevention – automated status updates that keep customers informed, so they don’t need to phone your contact center and sit in an IVR queue (You can apply this to a wide variety of situations, including: account opening, applicants and repairs).
  • Product launches, promotions and recalls – short-lived and urgent communications intensive processes that are often experimental in nature.


From Waterfall to "Test and Learn"

Having established where you might want to apply a Low-code approach, the next question is: "How?" Are the techniques and methodologies different when using a Low-code technology approach?
That’s another question I put to Clay Richardson in our interview. Here’s what he said:
"What we found when we did interviews with companies using these platforms is that they’re adopting new methods that actually focus on live trialing ideas. So you get an idea in, build a solution really quickly, test it out with customers and then validate whether it’s something customers want."
This aptly describes a "lean start-up" test and learn approach, where rapid experiments are the name of the game. The trick here is to enable experimentation that allows an organization to fail quickly and cheaply and therefore learn quickly from the experience.
The challenge for traditional IT approaches is that funding experimentation is hard to justify. As Forrester say in their recent report "The New Economics Of Experimentation":
"When it comes to experimentation, traditional technology management funding models aren’t applicable. These models tend to be risk- averse and struggle to put a price tag on money saved as a result of learning what doesn’t work."
This is where Low-code Development Platforms really come into their own compared to hand coding and more IT-centric BPM Suites, because you can produce working systems in days instead of months, you can take a more iterative, prototyping kind of approach. That, says Forrester, is crucial if organizations are going to "embrace an entrepreneurial mindset to better support the relentless pace of fast-moving and ambiguous business requirements".
A big part of mitigating the risks of failure, is to shorten the development cycle; something that is very hard to achieve if development is in the hands of a few highly trained specialists, who are in short supply, and therefore subject to an IT queue and risk-averse work allocation. As argued above, that problem is reduced with the lower complexity and reduced learning curve associated with Low-code platforms.
Of course the goal is not failure, but success. Low-code Development Platforms, such asMATS, are Cloud-based and have consumption-based license models. This means that any successful experiment can be scaled from low-cost pilot to enterprise deployment almost at a moment’s notice.
Agree? Disagree? Have your say.
Watch the full interview with Clay Richardson by clicking the image below:
Next month in Low-code Corner:
"Using Low-code techniques to improve customer experience and reduce costs at the same time."