Why the future of product analytics is codeless

Written by Adrian Nestico  | 

7 min

 

We’ve entered a new era in software. The much-heralded codeless movement is upon us. 

You might have also heard it discussed as “low-code/no-code,” but regardless of which term is used it’s important to understand that codeless isn’t so much a type of software as it is a new paradigm in how we use and build software. CapitalG’s Alex Nichols and Jesse Wedler recently observed that “no code is not a category itself…[but] in the same way that PCs democratized software usage, no code will usher in the next wave of enterprise innovation by democratizing technical skill sets.”  

Fundamentally, the low-code/no-code movement is a response to the problem that there simply aren’t enough software developers and engineers to meet current demand. This shortage of talent has accelerated the importance of developing codeless tools that can efficiently solve problems for business stakeholders without technical skill sets. Instead of continuing to rely on an ever-shrinking pool of engineering talent, low-code/no-code moves the job of developing and maintaining tools away from engineering teams and into the hands of the teams that actually use them, giving engineers more time to do what they do best: build the product. 

Pendo’s stance on codeless

At Pendo, we know a thing or two about the low-code/no-code movement. Since day one, Pendo was built on the (then radical) idea that product tools shouldn’t require a developer to maintain. 

That’s why Pendo decided to build Retroactive Analytics, a codeless approach to collecting and analyzing product data, often referred to as “auto-tracking.” Pendo’s founders noticed that the traditional way teams gathered product data was slow, inefficient, and engineering heavy. Traditionally, the product team would tell engineering to install tracking code on a feature and schedule the work for a future sprint before waiting weeks or even months for a meaningful amount of data to flow in. 

Pendo does things differently. Our Retroactive Analytics automatically collect every feature click, screen, and page load, all without tracking code. Then Pendo users–even those with no coding skills–can identify what to track by navigating through their product and tagging the most relevant features and pages with a simple point-and-click interface. 

It’s not an overstatement to say this has been revolutionary for Pendo customers. Before Pendo, less-technical teams were inevitably reliant on engineering to collect the data they needed (if they even went through the trouble of collecting it at all). With Pendo, these teams now have self-service access to not only the underlying data itself, but also the analysis tools to make use of it. 

Codeless enters the mainstream

While the low-code/no-code movement has garnered a lot of attention recently, the core ideas behind the movement aren’t new. In fact, tools like Microsoft Excel are cited as some of the earliest examples of low-code/no-code software. 

In a recent conversation with OpenView Partners, Howie Lui of Airtable explained that “the word low-code has existed for a long time. What’s really happened isn’t that low-code/no-code has appeared, but that it’s suddenly become mainstream […] it’s increasingly a thing.” In other words, new tech isn’t what’s behind the ascendance of the low-code/no-code movement. Instead, it’s the increasing rate at which organizations of all sizes are finally able to embrace codeless tools to help solve key business problems.

Despite the improvements to organizational velocity codeless solutions deliver, there are still detractors as they continue to go mainstream. Just like companies that resisted moving processes into Excel in the 1980s, today’s biggest critics will argue that codeless solutions don’t scale well or aren’t as secure as traditionally developed tools. 

What we hear from codeless skeptics

At Pendo, we still occasionally encounter these detractors when it comes to Retroactive Analytics. The loudest objections are typically the same criticisms levied at codeless solutions as a whole: Either that “low-code/no-code tools don’t scale” or that “codeless tools can’t capture the same level of detail as code written by people.” Or our personal favorite: “because auto-tracked solutions require less technical expertise, they must somehow be less secure.” Let’s break down each of these objections:

Myth #1: Codeless analytics are less secure

The argument that Retroactive Analytics must be less secure than developer instrumented tracking comes from the fear that without an engineer behind every keystroke, organizations will inadvertently, but inevitably, end up collecting personally identifiable information (PII) or other sensitive data about their customers. 

While security concerns are understandable–companies that have built codeless analytics solutions haven’t always matched Pendo’s level of execution–Pendo’s Retroactive Analytics were carefully designed to mitigate this risk. Specifically, Pendo never automatically captures user-entered text or information input into form fields, meaning that sensitive data like passwords can’t be accidentally collected. 

Myth #2: Low-code/no-code doesn’t scale

Similarly, the idea that low-code solutions result in less scalable software use doesn’t hold up to scrutiny. Critics of Retroactive Analytics often remark that even though engineering won’t need to add and maintain tracking code, someone still ends up having to organize the data, and Retroactive Analytics make that process messy. 

This argument tends not to stick because all it’s saying is that “by making product tracking easier to do, people won’t do it as well.” The truth is that data needs to be organized no matter how it’s collected. What we’ve seen at Pendo is that having two teams manage the process, a technical team to own the data collection and another team to own the analysis, inevitably introduces the biggest enemy of scalable software use: human error. 

Myth #3: You can’t capture as much detail with codeless tools

Specific to analytics, the idea that low-code/no-code solutions can’t capture the same level of detail as tracking code stems from the misperception that low-code/no-code analytics miss the metadata and properties that lend deeper context to how users are interacting with your product (things like which option was selected from a drop-down list). 

What critics seem to misunderstand is that an inability to capture property data isn’t an inherent limitation of codeless analytics solutions. While it isn’t something that Pendo had at launch, today, users can easily capture property data associated with a feature interaction. In fact, adding the relevant properties to a feature works almost exactly like tagging features themselves: simply select the relevant properties using Pendo’s target mode from inside your product. 

Most importantly, critics tend to misunderstand that low-code/no-code tools are not meant to replace coded workflows altogether. They introduce broad organizational efficiencies, but sometimes there’s still a need for traditional development work. Pendo understands this reality and offers the best of both worlds: Retroactive Analytics as well as coded track-event analytics. Retroactive Analytics help you save time and effort in 99% of scenarios, but we know sometimes you really do need a developer to instrument code to capture a crucial backend process. 

While low-code/no-code will continue to draw criticisms from the most sluggish among us, it’s evident to anyone plugged into the tech newscycle that codeless solutions are increasingly impacting how business is done. At Pendo, we try not to pay attention to the naysayers. We’re proud to be a part of the codeless movement.