Startup metrics, in plain English

The full revenue opportunity if every possible customer in the world bought your product.

TAM is the ceiling. It's not what you'll capture — it's the total dollars on the table. Use it to answer 'is this market big enough to matter?' Calculate top-down (industry reports) AND bottom-up (units × price) and compare; if they're 10× apart, your assumptions are off.

The slice of TAM you could actually reach with your product, language, and geography.

SAM forces honesty: filter TAM by who you can actually sell to today — right country, right language, right segment, right tech stack. If you sell English-only SaaS to SMBs in North America, exclude everyone else.

The realistic slice of SAM you can capture in 3–5 years given competition and resources.

SOM is what you'll actually fight for. Investors care about SOM more than TAM because it shows you understand competition, your sales motion, and the constraints of your team and capital.

The exact type of buyer most likely to need, afford, and stick with your product.

ICP isn't a persona — it's a filter. Apply it to your customer list: your top 10 happiest, longest-tenured customers should all match it. If they don't, your ICP is wrong.

What it costs you, on average, to win one paying customer (ads, sales time, onboarding).

Include EVERYTHING: ads, salaries of sales/marketing staff, tools, agency fees, content costs, free-trial infrastructure. A common mistake is reporting only ad spend, which makes CAC look artificially low.

The total revenue you expect from one customer before they churn.

LTV depends entirely on churn. A small change in churn dramatically changes LTV. Always use gross-margin LTV (after hosting/payment-processing costs), not raw revenue, or you'll overestimate.

How many dollars of lifetime revenue you get per dollar spent acquiring a customer. >3× is healthy, >5× is great.

The single most important unit-economics metric. If <1×, every customer loses you money. If >5× consistently, you might be leaving growth on the table — consider spending more on acquisition.

How much, on average, each active customer pays you per period (usually per month).

Rising ARPU = healthy upsells/expansion. Falling ARPU = customers downgrading or you're acquiring lower-tier users. Track ARPU per cohort, not just blended.

Predictable subscription revenue normalized to one year. Investors price SaaS off ARR.

ARR only counts recurring subscription revenue — exclude one-time setup fees, services, and usage overages. Investors value SaaS at 5–15× ARR depending on growth rate.

Predictable subscription revenue per month. ARR ÷ 12.

Decompose MRR into 4 movements: New (new customers), Expansion (upgrades), Contraction (downgrades), Churned (cancellations). Net New MRR = New + Expansion − Contraction − Churn.

Your concrete plan to reach the first customers — channels, pricing, message, sales motion.

Your GTM motion must match your ACV (annual contract value): <$1k = self-serve/PLG, $1k–$10k = inside sales, >$10k = field sales. Mismatches kill startups.

The smallest version of your product that real users will pay for and give feedback on.

MVP ≠ broken product. It should be small in scope but high in quality on that scope. Famous example: Dropbox's MVP was a 3-minute demo video, not code.

The single sharp use case that gets you in the door — even if your long-term vision is broader.

Best wedges are narrow and urgent. Superhuman wedged on 'fast email for power users' before becoming a productivity suite. Figma wedged on 'real-time design collab' before tackling all of design.