You can build many technology capabilities using low-code and no-code platforms, including applications, databases, workflows, integrations, Internet of Things (IoT) data streams, data visualisations, and more. They are versatile platforms to help replace spreadsheets, reduce technical debt, or experiment with machine learning.
Low-code platforms promise developer productivity, higher-quality experiences, and easier-to-maintain applications by providing abstraction layers compared to traditional coding.
No-code raises the promise by empowering non-developers to build and support all or parts of an application without needing devops practices or the ability to configure cloud infrastructure.
These abstractions can provide significant benefits. I’ve rolled out many low-code and no-code applications and seen other companies benefit from using them. Coding is still very important for many strategic applications, but organisations can launch and support many more capabilities when low-code and no-code development options are available.
However, there are trade-offs that may not be apparent until you’ve used these platforms a few times. The pandemic also created a boom for low-code and no-code because many IT departments had to rapidly modernise applications and build job-specific tools.
The increased demand brought on a new wave of low-code and no-code tools, opening the door for some that have overpromised and underdelivered needed capabilities.
What are the signs of trouble? I asked several experts to provide some hints. I'll also share some of my experiences.
1. Low-code projects don't meet user expectations
When your stakeholders picture a three-bedroom house and a garage, but all you can deliver with a low-code tool is a shack with a bathroom, you are likely missing everyone’s expectations.
Low-code platforms require training to use them effectively and discussions with stakeholders on trade-offs to achieve a business outcome. When developers can’t achieve the business objective or vision, it may require rethinking the platform choice and technical approach.
Tam Ayers, field CTO at Digibee, says, “A key indicator is when an organisation begins to adjust requirements or lowers their expectations of desired business outcomes due to the limitations of their low-code platforms. Any low-code platform should accelerate value delivery to the business, not the other way around."
2. Low code is inadequate for business requirements
Many low-code and no-code platforms allow developers to customise the implementation with custom code. But if you are adding too much pro code, being confined to a low-code platform may be constrictive. Alternatively, if business stakeholders are writing requirements and aren’t open to the solutions accelerated through low-code platforms, you might as well develop a custom solution.
David Brault, product marketing manager at Mendix, agrees, “A low-code solution that requires developers to leave the platform and revert to full-code development environments to make enhancements to an application is one that will consistently underdeliver.”
Guljeet Nagpaul, chief product officer at ACCELQ, adds, “One sign that your low-code platform is not working concerns customisations. If you find that your platform needs constant customisations, that suggests that the code is being written without the discipline of architecture and sound design.
"The maintenance of this customisation will quickly become unsustainable and ultimately drag down the return on investment.”
3. Platforms advertise no code but still require developers
A platform must live up to its category and promise. No-code platforms should be just that — platforms that non-technologists can use to develop and support a capability without the need for IT to develop, test, and deploy. No-code platforms are the tools for citizen developers, businesspeople who have the time, interest, and sufficient technical acumen to build capabilities with simplified tools.
But that doesn’t stop people from claiming that a platform or capability is no code.
Dinesh Varadharajan, chief product officer at Kissflow, says, “If business users struggle to create simple processes or apps on their own and continue to depend on IT, it means that the no-code platform isn’t offering an inclusive approach as promised.”
4. Low-code platforms claim you don’t need IT or developers
Low code is different from no code; the expectation is that some coding experience will be needed to develop an application, database, or integration. Low-code platforms aim to help developers build solutions faster, easier, and with less support than a pro-code solution.
Although they often have visual development capabilities, some coding experience or IT knowledge is often required in a low-code environment’s development life cycle.
It’s possible to hear a low-code platform say that IT isn’t needed to support the platform. Francis Carden, VP of intelligent automation and robotics at Pega, says that’s a red flag.
“When a low-code solution is promising that you don’t need your IT department’s involvement, there’s a disconnect,” he says.
“Yes, you might be able to build fast, but what happens when things go live? Who determines the viability and risk at that point, and who supports what you build when things need updating, fixing, or when compliance enforces critical changes?” This overpromising of being able to work without IT will result in complications down the line.”
Worth noting: Some platforms support both no-code and low-code paradigms with one set of tools for citizen developers and more advanced low-code capabilities for software developers. But even when deploying a no-code solution to citizen developers, saying you don’t need IT at all is an overpromise that can lead to technical debt, security issues, and other complications.
5. Low code leads to a motley of system integrations
I’ve created apps and workflows that connect several low-code platforms into one overall solution’s architecture. But Kevin Marcus, CTO and cofounder at Versium, asks whether the purchase, configuration, and integration of multiple SaaS and low-code solutions outweigh the benefits.
“Rigidity of low- and no-code systems often lures teams into requiring even more systems to handle cases that lay outside of the original system’s capability,” he says.
“Tragically, this leads to a smorgasbord of systems that need to be connected and integrated together, often taking even more time and resources to solve basic issues that would have been easier solved directly through IT or engineering using proper tools in the first place.”
This illustrates one reason low-code and no-code development requires IT architecture support. Perhaps a minimum-viable product is achieved by integrating low code with a software as a service. But, if several iterations later the solution mushrooms into many integrated tools, then IT might suggest refactoring to a more robust solution.
6. Low code needs access to multiple integrations and raises security risks
Alon Jackson, CEO and cofounder of Astrix Security, says to raise red flags when any platform requires opening too many ports and full-access integrations. He is concerned about “integrations requiring a high level of access to a large set of critical enterprise systems, undermining traditional security review processes and exposing companies to potential data leakage.”
These important concerns apply to any integration, whether done in SaaS, low-code, or custom code. Jackson suggests secure implementations and integrations require these key capabilities: visibility, threat detection, contextual mitigations, security policies, and enforcement guardrails.
7. “One platform fits all” can be a barrier
Brault adds a second concern about whether low-code and no-code technologies support flexible delivery and cloud hosting platforms. “Signs of a poor low-code solution include not supporting native mobile app development in addition to web and Progressive Web Apps support, or not being cloud native or multi-cloud supportive.”
I believe that not supporting or poorly supporting mobile development is a significant concern when building apps on any platform. The question is whether you can easily configure web versus mobile experiences based on the expected user personas and use cases.
8. Low-code development is hard to test
So low-code and no-code platforms make it easy to build and deploy applications, databases, and integrations, but how easy is testing them? That’s a concern raised by Cyril Otalora, director of solution engineering at Provar.
“A testing strategy is often an afterthought with low-code platforms,” he says. He highlights the risk: “The promise of quick deployment, lower costs, and higher security goes out the window if companies can’t keep up with the regression risk and resort to costly and tedious manual testing.”
My experience with low code and no code
I previously shared seven keys to selecting low-code platforms. I use many different low-code and no-code solutions myself. Here are my pet peeves on platforms that overpromise:
- Low-code platforms that release new versions that require you to rewrite or significantly rework your applications
- Platforms that don’t communicate outages, defects, or other problems that impact application performance
- Advanced platforms with poor technical support (I shouldn’t know more about the platform than the support reps.)
The key message is IT leaders have to do their homework. Low-code and no-code platforms offer significant benefits, but they require research and proofs of concepts to validate their capabilities.