E37: SAAS Economics with Michael Alt

‍

In this episode of the SaaS Operators Podcast we sat down with Michael Alt, founder of Upstack Data. We went deep on something most people underestimate until it hurts: pricing and the real cost of running software. 

‍

Usage pricing, infra bills, tool stacks, copycats, AI, gross margin. All the boring words that quietly decide who survives.

‍

Usage Pricing Feels Great Until It Doesn’t

‍

We started with the “November high” of usage based pricing.

‍

If you have a usage based product, November can give you a 30 percent bump in revenue without changing anything. More orders, more usage, more billings. Feels incredible for about a month.

‍

Then the support tickets hit.

‍

People do not anticipate the spike, their invoices triple, and suddenly your CS team is explaining math instead of the product. Usage pricing can work, but it also creates “unexpected bill” moments that erode trust if you are not careful. 

‍

Jeremiah talked about one of the first things he did at Stamped: he killed overage charges. They still use order volume as a lever, but if you blow past your plan on Black Friday, they just process it and talk about upgrading later. The philosophy is simple:

‍

I do not want my customers thinking about the bill. I want them happy enough that when they do look, it just makes sense. 

‍

That is the opposite of “surprise, you owe us 3x this month.”

‍

The nuance here is important. Usage and volume can work for infra type products. They make less sense for something like reviews, where the value feels more fixed in the customer’s head.

‍

What Is Bad Pricing Really

‍

We circled a simple idea: bad pricing is when your ideal customer does not want to pay it. That is it.

‍

Not “bad” as in Twitter outrage. Bad as in it pushes the right people away.

‍

If you bolt on overages or usage fees that your ICP hates, that is bad pricing, even if the spreadsheet says it is optimal. If your model pushes the customer to constantly re-evaluate you, you are inviting churn.

‍

On the flip side, we all admitted there are tools we would happily pay much more for. Slack. Notion. Google Workspace. Whisperflow for Zach. Products that disappear into the workflow and feel cheap relative to the leverage they create. 

‍

That is the bar: you want to land in the “I would pay 3–10x more and still not switch” bucket, not the “I am quietly planning my migration” bucket.

‍

Software Cost Hurts SaaS More Than Ecom

‍

One thing Jeremiah pointed out that most ecom founders do not see: SaaS founders actually feel the pain of software spend more than they do.

‍

In ecom, COGS is plastic, fabrication, and logistics. Software is an operating line item, but it is not the bulk of the unit economics.

‍

In SaaS, software shows up in both COGS and OPEX. Infra plus tools. AWS, GCP, Cloudflare, Postgres, logging, analytics, CDNs, queues, AI models. You can easily end up with a P&L where software spend as a percentage of revenue is higher than most ecom brands will ever see. 

‍

So when a SaaS founder complains about your pricing, it is not theory. They are staring at their own stack and feeling the same thing from the other side.

‍

Tools VS Infrastructure Are Not The Same Problem

‍

A big theme in the conversation was separating “tools” from “infrastructure.”

‍

Tools are the HubSpots, Notions, CRMs, point solutions. Most of us are pretty relaxed there. If a tool saves an hour per person per month, 200 dollars a seat barely matters at scale. The opportunity cost of over-optimizing that spend is high.

‍

Infrastructure is different. That is where small decisions compound into real money.

‍

We talked through real examples:

‍

• Jeremiah digging into a bloated S3 bill and finding 12k a month he could cut with about three days of engineering work.
• Rishabh looking at an infra vendor that was about to cost 500k a year and hiring contractors for a few months to rework it so the long term run rate dropped. 

‍

The rule of thumb was simple:

Only optimize when the savings are clearly bigger than the cost of the person who needs to fix it, over a reasonable time window.

‍

The nice part about infra is that the work usually improves performance too. Lower infra cost often equals faster product, less bloat, better UX. High infra cost is usually a symptom of architectural debt, not just “cloud is expensive.” 

‍

Capital Structure Changes How You Think About Margin

‍

Rishabh made a good point about why different companies think so differently about cost and gross margin.

‍

If you are sitting on a big balance sheet, funded by venture, you can afford to ignore gross margin for a while. Your investors are playing a different game. They want breakout winners, not tidy 30 percent margin businesses. So you hear this idea circulating:

‍

“Do not worry about gross margin right now.”

‍

That creates a whole class of companies with terrible unit economics that are still “fine” as long as there is another checkpoint ahead: another round, a sale, an IPO. The goal is to exit before the math catches up. 

‍

If you are bootstrapped or focused on profitability, that advice would get you fired. You do not have the luxury of ignoring gross margin. Infra, headcount, tooling, everything has to tie back to a real business model.

‍

Same market, completely different operating rules.

‍

Pricing As A Growth Unlock, Not A Haircut

‍

Zach spent a chunk of time literally printing bank statements and looking at contracts. Not from a “cut everything” scarcity mindset, but to renegotiate infrastructure deals, commit where it made sense, and bring marginal costs down.

‍

If you can lower your cost to serve by 30 percent on something meaningful, you just created room to spend 30 percent more to acquire a customer at the same contribution margin. That is not cost cutting, that is growth capacity. 

‍

Pricing on your side (how you charge customers) and pricing on the vendor side (how others charge you) both roll up into the same question:

How much can we afford to pay to win and keep the right customer.

‍

Copycats, Competitors, And Why They Still Lose

‍

Later in the conversation we shifted into copycats.

‍

Zach has more than his fair share of Foreplay clones. Same category, similar UI, sometimes even the same product names and icons. It is annoying. Not because of “competition,” but because bad copies make the entire category look worse.

‍

There is a big difference between:

‍

• A real competitor with their own thesis, data, and angle
• A copycat that just rebuilds what they see on your screen

‍

The first group makes the pie bigger. They convince customers you could not have convinced yourself. You end up learning from them.

‍

The second group muddies the water. They give customers a broken version of “how this should work” and make everyone’s life harder.

‍

We all agreed on one thing: the only companies you should actually worry about are the ones who are not copying you. They found a different secret. Those are worth studying. The ones who cannot even copy properly are not. 

‍

AI Makes Building Easy, But Running Is Still Hard

‍

AI came up a few times as the new copycat multiplier. You can now spin up a passable 80 percent product very fast, especially in low infra categories.

‍

That does not mean it is a business.

‍

Jeremiah told the story of selling little WooCommerce “widgets” early in his career. Tiny code snippets wrapped in settings. That kind of thing is dead. Now you just ask AI for the snippet and paste it in yourself. No point paying a yearly fee.

‍

So what still matters?

‍

• Real infrastructure
• Real data moats
• Real network effects
• A category insight that is hard to copy

‍

AI makes the starting line cheaper. It does not remove the hard part, which is running a performant, affordable, defensible product at scale.

‍

What actually matters for operators

‍

1. Pricing is not a spreadsheet game, it is a relationship game
2. Infra is not just a cost center, it is one of your biggest operational levers
3. Capital structure quietly decides how honest you need to be about gross margin
4. Copycats and AI will keep increasing surface area, but not everyone can run a real business at scale

‍

As a founder or operator, your job is to pick where you want to live on that map.

‍

Get clear on:

‍

1. Who you serve and how they want to be billed
2. What it truly costs you to serve them
3. Where you are happy to pay for leverage
4. Where you need to be ruthless about infra and waste

‍

Everything else is just noise that shows up on Twitter before it shows up on your P&L.

‍

‍