Profitable startups are back in vogue

Startups have always chased vanity metrics.

During the dot-com bubble, it was eyeballs: visitors and free users. After the crash, we collectively got smarter and started focusing on revenue, but during the zero interest rate period (ZIRP), startups chased revenue growth at the (literal) expense of everything else.

Today's AI startups do the same, stunning us with their sky-high monthly recurring revenue (MRR) or annual recurring revenue (ARR).

Revenue is better than just counting users: at least these are users who actually pay you. But revenue doesn't mean your startup won't fail.

Revenue quality

What's the quality of your revenue? Why are users paying you? Are they getting the value they expect, or will they quickly leave?

Churn is the rate at which customers stop paying you. The basic formula is:

Churn Rate = (Customers Lost in Period / Customers at Start of Period) × 100

If you started the month with 100 customers and lost 5, your monthly churn rate is 5%.

High churn kills growth. You can't keep pouring new customers into a leaky bucket. If users decide in two months that your product doesn't work or isn't for them, your growth will eventually stall.

Churn directly affects your customer lifetime value (LTV), which is how much total revenue you'll get from a customer before they leave:

LTV = Average Revenue Per Customer / Churn Rate

If your average customer pays $100/month and your monthly churn is 5%, the LTV is roughly $2,000 ($100 / 0.05). Cut that churn to 2% and your LTV jumps to $5,000. Same customer, same price, but you keep them longer, so they're worth more than twice as much.

Once you have sizable revenue, churn is probably your most important metric. You want high-quality revenue: customers who match your ideal customer persona (ICP), who are happy, and who pay you month over month or year over year.

The best companies have negative churn. They upsell services over time, so revenue per user increases even as some customers leave. Instead of fighting to maintain revenue, they grow it from the existing base.

Unit economics

What does it cost you to earn one dollar?

This is where COGS (Cost of Goods Sold) comes in. COGS is the direct cost of delivering your service to a customer. For AI startups, it can be the tokens they pay to OpenAI, Anthropic, or whoever. For a SaaS company, it might be hosting costs, payment processing fees, and customer support tied to that customer.

Your gross profit is what's left after you subtract COGS from revenue:

Gross Profit = Revenue - COGS

If you're selling AI services for $100 but spending $200 in tokens, your COGS is $200 and your gross profit is -$100. You're losing money on every sale. Your unit economics are broken.

You cannot lose money per unit and make it up in volume.

Sometimes you might do this temporarily. Maybe you implemented something quickly to get to market fast, and you can optimize later to bring costs down. If you have a clear plan to either reduce costs or increase prices, it might be okay for a short period. But if you don't have that plan, you're just selling a dollar for 50 cents. That doesn't work.

The ARR Illusion

There's another problem with MRR: people multiply trailing MRR by 12 to get ARR. This is wishful thinking.

This assumes your churn is low enough that the median customer will stay for more than a year. But if a customer leaves after two months, your ARR for that customer is two times the MRR, not 12 times the MRR.

You can't assume you'll collect that annual revenue if your customers aren't sticking around.

The Path to Profitability

More and more startups like Linear are focusing on being actually profitable. This means not only having growing MRR or ARR, but also ending each period with more money than you started with.

There's an important distinction here: gross profit (revenue minus COGS) versus net income (also called net profit).

Net Income = Gross Profit - Operating Expenses

Your operating expenses are everything else: salaries, rent, marketing, software subscriptions, legal fees, accountants. All the stuff that keeps your business running but isn't directly tied to delivering your product to a specific customer.

You can have positive gross profit (making money on each sale) but still have negative net profit (losing money overall) if you don't have a lot of customers, or your operating expenses are too high. This is common for startups. You make money on each customer, but you spend more on salaries and growth than you bring in.

When people talk about being "profitable," they usually mean positive net profit. For startups with runway from investor funding, this means you're not depleting your cash reserves. You're increasing them.

Early-stage startups often aren't aiming for profitability yet. They focus on growth or validating their solution. VC-backed startups expecially aren't in a rush to become profitable. Some haven't been profitable for years before becoming successful (like Google). But for bootstrapped startups who rely on founders' own funds or family and friends, profitability matters from day one

And in the wider world (mom and pop stores, lifestyle businesses, large and public businesses), profitability is the whole point.

Black is the new black

With the end of the zero interest rate period, more startups will need to focus on becoming profitable faster. AI startups may defy gravity for some time because of high expectations, but that won't last forever.

As Paul Graham says, aim for ramen profitability: profitable enough that you can pay your bare expenses and not depend on external funding. That gives you freedom. You don't depend on others to set your direction, goals, and approach.

Revenue can tell a great story, but profit is what actually keeps you alive.