No-Code Overload

Superpath doesn’t exist without no-code, but it’s straining under the weight of it these days. It’s the reason the business has scaled, and yet it often feels like a scourge on my day-to-day work. Maybe it’s hedonic adaptation—surely, this is better than it was before—or maybe I’ve traded in one set of problems (manual tasks) for another.

I’m using the term “no-code” to describe everything from our client portal (which runs on Airtable and Softr) to our freelance payroll integration (a combo of webhooks and Airtable) to the dozens of small automations (mostly in Zapier) that move subscribers from one place to another, send welcome messages to new members and so on.

I’ve got 69 Zaps processing 8,300+ tasks each month. It’s amazing how much happens when all this works properly. Writers get paid, Superpath gets paid, customers get notified, Airtable bases get updated, and data moves from one place to another.

But I also find myself troubleshooting these workflows constantly. The more sophisticated the workflows get, the more room there is for errors, and the more time I have to spend troubleshooting.

I love no-code. What went wrong?

I actually love no-code. My favorite things to do at work are write, get lost in a spreadsheet and build things with no-code. Each scratches the same itch to create. I imagine this is the same reason others love to design or write code—it’s incredibly satisfying to build things. So there’s a real tension between my sincere love of no-code and how much headache it’s caused me over the last few years.

Those headaches often feel like a steady stream of unrelated issues, but upon reflection, they fall into a few buckets:

1. Technical debt

I had no idea what technical debt was until recently. It’s not a term that had ever come up in my career as a marketer, but now it’s something I think about every day.

I’ve built all kinds of apps and automations for Superpath, but I’m not a developer or product manager. Most of this stuff got built in a hurry because the pain of the problem it was solving was significant enough that I needed to move quickly. There’s no product vision or documentation—just me making stuff with a “build it now, fix it later” mentality.

Source: Asana

In reading about technical debt, I realize that my version is lightweight. It can be a huge problem for real software companies. For Superpath, it’s an annoyance that I mostly just live with. Over the years, we’ve accumulated technical debt mostly because:

1. I prefer to build rather than buy in most cases, and
2. Developers are expensive, so I do most of the work myself.

Every now and then, I’ll spend a few hours fixing up a workflow that is causing issues and I hate it. The delight I get from building these workflows is replaced by misery. Because I’ve failed to document anything, I have to investigate my own work. Why did I create this filter? Why did I tell this workflow to run every week instead of every day? Is there another automation triggered based on the success of this one?

It’s a literal maze of apps, actions, triggers, steps, filters, delays and, more recently, AI. Whenever I try to chip away at our technical debt, I find myself wondering how much effort I’m actually saving by automating this stuff in the first place. Did I get carried away building stuff because it was fun to do? Or was the initial problem worth all this work? Most of the time, I honestly can’t remember what it used to be like, so I have to assume that fixing bad no-code is better than abandoning it altogether.

2. Pushing the limits of no-code

In some cases, my no-code workflows include extra steps (and therefore opportunities for problems) because I’m not working directly with an API. If I knew how to write even simple code, I could probably reduce complexity, remove extra steps and work with these tools directly rather creating a bunch of workarounds.

We have several high-volume automations that power our marketplace and freelance payroll that I think are pushing the limits of no-code. These workflows also cause the most headaches. I’m almost certain that working directly with APIs (Airtable and Wingspan in this case) would be a much more direct solution, but I’m intimidated by the idea. If I hire a developer to rebuild this stuff, won’t it cost a fortune? And then, won’t I lose the ability to troubleshoot on my own? I’m not that interested in handing over control at for $150/hour.

AI may end up changing this. I had never honestly considered “real” code as a possible replacement for no-code until my friend Andrew Tate (this Andrew Tate, not that Andrew Tate) showed me Vercel. You give it a prompt, it gives you code. I am amazed by it. I’m still overwhelmed by the process—databases, servers, GitHub, etc.—and am not yet ready to make this move. (Though, if I’ve learned anything recently, it’s that AI is changing so fast that this could be viable even a few weeks from now.)

3. Invisible automations

This, perhaps more than anything else, drives me crazy about no-code. Because most of the apps I use are connected by Zapier, automations run in the background. For example, if I change a status in Airtable, it triggers a workflow. It's invisible and it's consequential. I'm mostly able to remember this kind of thing so I don't accidentally trigger something I don't mean to, but all this remembering clogs my brain and I still end up triggering stuff by accident anyway.

I've also automated so many things that I forget they are even there. For example, when someone posts a freelance gig in our Slack channel, we ask them to fill out a form so we can standardize the data, then that posts in Slack, and then we send some follow-up emails to see how it went. It's complicated considering that it's only marginally better than just letting people post directly in the channel. And when it doesn't work, I don't even know. I recently found that no follow-up emails had been sent for almost a year. Did it matter? The emails are a nice touch, but certainly not a requirement.

Zapier has been adding features to ease this. You can now add notes to each step of your workflows, see a version history and setup error notifications. This is great, but there’s a larger problem here. No-code almost always entails connecting apps, and Zapier (or similar tools) are the glue that hold them all together. Even a simple workflow likely involves at least three tools: a data source, Zapier and a data destination.

So for the end user, you still have to remember that changing a status in Airtable might send an email to a customer or queue up a payment for a contractor. It’s incredibly difficult to track all this, and documentation only helps so much. Am I really going to reference documentation every time I change a status in Airtable? Of course not!

The state of no-code: we need change

I absolutely love building things, and no-code offers me, a non-technical founder, the opportunity to create customer experiences, tools, reports, etc. without code. This is revolutionary. I consider my skills as a Zapier and Airtable power user to be critical in my own career and the growth of Superpath.

I recently hired someone to help me with admin work, some of which is no-code troubleshooting. As I started explaining a few common errors, my brain hurt trying to remember not just how the workflows should work, but why they even exist at all.

Lately, the entire business feels duct-taped together by Zapier. What used to feel amazing—a one-person business scaling via automation!—now feels like a very disconnected experience both for me and our members/customers.

I haven't decided yet what to do about this. In some cases, we should evolve from no-code to code. In others, we should probably just eliminate unnecessary workflows. And still in others, this is just the small cost of automating work that would be far too painful to do manually.

But something has to change. Every SaaS tool now offers automation inside the app and also integrates with Zapier or webhooks so you can build things off the platform too. This was wonderful 10 years ago when it unlocked your data for the first time, but it’s turning into the software equivalent of “cutting the cord.” It used to feel amazing to stop paying a cable provider. And now, a single weekend of NFL games requires three different streaming services.

Like former Netscape CEO Jim Barksdale says, “There’s only two ways I know of to make money: bundling and unbundling.” No-code is currently in an unbundled phase that I believe is setting the stage for a bundled future (which will eventually be disrupted by unbundling).

For me, no-code emerged as the single best way to scale Superpath as a one-person business. It’s still very much at the core of that idea and is built into every part of our operations. Perhaps the challenge isn’t that no-code is a problem in itself, but rather that it doesn’t scale particularly well. Or that it replaces code but not product management or vision.

Either way, I’ve got a case of no-code fatigue and I know I’m not the only one.

‍