Imagining a software vending machine for “pay as you go” infrequent tasks
When you find a task that needs to get done and gated by a subscription, finding a solution becomes complicated. Sign up and cancel? Vibe code it? Read: "Everything Starts Out Looking Like a Toy" #257
Hi, I’m Greg 👋! I write weekly product essays, including system “handshakes”, the expectations for workflow, and the jobs to be done for data. What is Data Operations? was the first post in the series.
This week’s toy: enjoy retro weather design with this charming web page that shows you the current weather like it’s the 1990s Weather Channel. It turns out that having a constraint like low resolution graphics made developers focus on creating a great information architecture.
Edition 257 of this newsletter is here - it’s June 30, 2025.
Thanks for reading! Let me know if there’s a topic you’d like me to cover.
The Big Idea
A short long-form essay about data things
⚙️ Imagining a software vending machine for “pay as you go” infrequent tasks
I’ve noticed a change happening in software buyer intent where people want something more like a vending machine than an always-on service that charges a recurring subscription. These buyers are responding to the current world where subscriptions are everywhere.
This might sound familiar…you buy paper towels through Amazon.com and the checkout prompts you to "Subscribe & Save" to get an extra 5% off. You buy a coffee and the barista suggests a monthly coffee delivery subscription. You decide to watch a hit show and need a new streaming service because it’s exclusively on that channel.
You don’t need another subscription.
The subscription economy is projected to reach $5 trillion by 2025, but consumers are beginning to act against the burden of monthly fees. Recent surveys show that U.S. consumers are spending significantly more on digital subscriptions than they realize.
And we’re still building software the same way we did in 2010.
Avoiding the subscription trap
The subscription model for software made perfect sense when it first emerged. The pricing to buy month to month lowered the barrier to entry and made expensive software accessible to more people. (It also hid the full annual cost.)
Recurring revenue is very tasty, so almost every software became a subscription. And not just the big, complex applications that justified ongoing development costs. We're talking about smaller things like e-signature, PDF conversion, simple image resizing, and text reformatting that got rolled in as features on popular tools.
A subscription product is now available for almost anything you can imagine. You have to remember to cancel one-time services. You're paying monthly fees for tools you might use quarterly, or yearly, or never again.
Streaming services are seeing declining spending as users aggressively prune excess services. People are hitting their limit, gravitating toward convenience and predictability rather than endless monthly commitments.
There’s a parallel for B2B buyers
While consumers are drowning in subscriptions, the B2B world is quietly moving in the opposite direction. Usage-based pricing is becoming the new norm in the SaaS world with more and more companies adopting consumption-based models.
AWS charges you for compute time, not monthly server licenses. Their "pay as you go" model means you only pay when you need them. Customer support platforms charge per resolution, not per seat. Zapier charges for task executions, not monthly access.
The B2B world figured out something important. Buyers will pay for value delivered, not just for adding a seat.
What about occasional tasks?
Here’s an example that doesn’t fit well with a subscription. When I need to do something specific where I don’t have a current subscription (like convert a PDF to markdown, or clean up some audio, or generate a set of app icons), I need to do a search to figure out how to solve it.
When I look for options, these are the most obvious:
Find a free, ad-supported tool that might send my data to who-knows-where
Install an open-source solution and spend 30 minutes getting it working
Ask an AI assistant, which usually responds with "let me show you how to do that" and gives me code instructions
Subscribe to a tool I'll use once and then forget to cancel
None of these options feel right. The free tools are sketchy, the open-source route takes too long, the AI gives me homework, and the subscription feels like overkill.
I still need to do the task, and I need to get it done today.
Imagine a software vending machine
What if you could walk up to a digital vending machine, drop in your file and a few dollars, and walk away with exactly what you need?
This isn't a new idea! We used to buy software on disks and cartridges, and you paid once for the software. But the modern version would be different.
Instead of owning the software to do the task whenever you want, you're buying the copy made by the copy machine. It’s a small convenience fee to have it done right now.
This isn’t unique, and real-world examples already exist. Some companies offer e-signatures for a few dollars per document with no plan required. Others charge small amounts per image for AI background removal in credit packs.
They’ve figured out that some consumers want outcomes without signing up for an ongoing relationship.
Why changing the conversation is important
The pay-as-you-go model changes the fundamental relationship between user and software. Instead of "subscribe and hope you use it enough to justify the cost," it becomes "pay for what you actually use."
This model creates a different kind of accountability. The software has to deliver immediate, measurable value. There's no hiding behind monthly fees or annual contracts. Each transaction is immediate feedback on whether the tool is worth the price.
It changes how we think about data.
In a subscription model, the company wants to keep your data forever to justify the ongoing relationship. In a pay-as-you-go model, the goal is to process your data and delete it quickly. The value is in the transformation, not the storage.
It’s not rocket science
Making a model like this work is not a technical risk, and the technology is well-proven. Serverless computing means you pay for the compute time you actually use. Cloud storage with automatic expiration means files disappear after they're processed. Payment processors can handle micro-transactions efficiently.
Micropayments have become viable for transactions under $5, eliminating the need for large upfront costs or commitments.
The pieces are all there. We just need to put them together differently.
This isn't just about pricing models. It's about how we think about software ownership and value. In a subscription world, you're renting access to someone else's tools. In a pay-as-you-go world, you're purchasing specific outcomes.
Each model has its place. But right now, we're forcing everything into the subscription model, even when it doesn't fit.
The future of occasional tasks
I suspect we're going to see more of this model emerge in the next few years. When you have a specific need that doesn't justify a subscription, you want a specific solution that doesn't require a subscription.
Analysts are already seeing lifetime deals and one-time payments resurging. The companies that figure this out will build the digital equivalent of the corner store. Walk in, get exactly what you need, and walk out without any ongoing commitment.
The beauty of this approach is its simplicity:
No complex pricing tiers
No feature comparisons
No "which plan is right for me" decision paralysis
Just a clear price for a clear outcome.
Playing the long game
When you pay for a specific outcome rather than ongoing access, the relationship becomes more transactional but also more honest. The company isn't trying to keep you engaged or increase your usage. They're just trying to deliver the value you paid for.
This creates a different kind of trust. You trust them to do the job you paid for, not to keep you hooked on their platform. The pay-as-you-go approach represents a middle ground.
You don't own the software, but you do own the specific outcome you paid for. You're not committing to an ongoing relationship, but you're also not dealing with the complexity of full ownership.
It's a model that respects both the user's need for flexibility and the developer's need for sustainable revenue. It's honest about what's being exchanged - money for specific value, not money for potential access.
The future is (predictably) unclear
I don't think subscriptions are going away. They make perfect sense for tools you use every day, for services that provide ongoing value, for platforms that become part of your workflow.
But I do think we're going to see more alternatives emerge. More companies will offer both models: subscribe if you use it regularly, pay-as-you-go if you don't. More tools will be built specifically for occasional use cases.
It's a small change that could make a big difference in how we think about software, value, and the relationship between users and the tools they use.
What’s the takeaway? There’s a middle ground between one-time purchase and saas subscription models, and it’s evident whenever you hit an infrequent but annoying task that needs to be done. It would be great to have a “vending machine” approach to solve it.
Links for Reading and Sharing
These are links that caught my 👀
1/ Build some toys - Joshua Baretto is a dev after my own heart and encourages you to build toy software. When you’re working on something that doesn’t matter, you might get really creative … or discover a new thing … or just get blocked and never finish. The point is that we should all have more weird projects that we work on just because!
2/ Vibe data - It’s probably too early to declare that all data tasks are “easy” but the models have gone from “look up the syntax for this weird function” to “build me a dbt model for that and design a fact table while you’re at it.” For real data engineers, this is probably annoying. For the rest of us, it’s a great way to make a first draft.
3/ From search engines to …? - Brian Balfour shares some insightful thoughts on the current change in distribution platforms. We’re reaching an inflection point where we will find winners in the AI game.
What to do next
Hit reply if you’ve got links to share, data stories, or want to say hello.
The next big thing always starts out being dismissed as a “toy.” - Chris Dixon