- What's the formula?
- Monthly compounding. Each month, the ending ARR = starting ARR × (1 + growth_rate − churn_rate), where both rates are monthly. Equivalently: new ARR added = starting × growth_rate, churned ARR = starting × churn_rate, ending = starting + new − churned. The next month's starting ARR is this month's ending. Over N months: ARR_N = ARR_0 × (1 + growth − churn)^N. Compounding monthly (not annually) matters — naïve annual math overstates the ending number because it applies churn against a constant base, when in reality churn compounds against a shrinking base while growth compounds against a growing one.
- Why use monthly compounding instead of annual growth × annual churn?
- Because real SaaS revenue doesn't update once a year — it updates every month, every renewal, every cancellation. If you grow new business at 5%/month and lose 2%/month to churn, the math isn't (60% − 24%) at year end — that's a 36-point gap. The honest answer is 1.03^12 − 1 ≈ 42.6% net growth, because growth compounds on growth and churn compounds on a shrinking base. The gap between the back-of-envelope number and the real number widens fast as growth rates rise. A model that uses annual math will overstate ending ARR by 10-30% on realistic inputs.
- What is T2D3 and is it a realistic benchmark?
- T2D3 = "triple, triple, double, double, double" — the trajectory Bessemer Venture Partners formalized in 2015 after watching the early SaaS leaders (Salesforce, Workday, NetSuite, ServiceNow, HubSpot) compound from \$1-2M ARR to \$100M+ in five years. Year 1: triple to 3×. Year 2: triple again to 9×. Years 3-5: double, double, double to 18×, 36×, 72×. It's a benchmark for venture-scale, US-centric, horizontal-SaaS companies — the kind that need to justify large rounds and an IPO path. Most SaaS companies don't hit T2D3, and that's fine for most SaaS companies. The benchmark is meaningful as a check on whether your plan COULD support a venture-backed outcome, not as a verdict on whether your business is good.
- Should I use revenue churn or logo churn here?
- Revenue churn (also called gross dollar churn). The model is in dollars, not customer counts — logos that downgrade contribute partial churn, expansion within retained accounts offsets churn. The cleanest input is gross monthly revenue churn: total ARR lost from cancellations + downgrades, divided by starting ARR. If your accounting only tracks net dollar retention (NDR = 1 − churn + expansion), back out the gross churn separately or use (1 − NDR) as a proxy with a note that you're conflating the two. Logo churn (% of customers who left) is informative for retention diagnostics, but it overstates revenue impact when small customers leave and understates when whales leave.
- What's a realistic monthly growth rate?
- Depends on stage and segment. Pre-product-market-fit: noisy, can be 0% or 20% month-over-month. Early-stage SaaS with PMF: 8–15%/month is hypergrowth, sustainable for a year or two. Past \$10M ARR: 3–7%/month is excellent. Past \$50M ARR: 2–4%/month is excellent (because the base is so large). Past \$100M ARR: 1–3%/month is the norm for venture-scale companies. If you're plugging in 20%/month and projecting forward 5 years, the model will give you a fictional answer — sustained 20% monthly compounds to 800× in 5 years, which has happened approximately never for a SaaS company past \$1M ARR.
- How does net negative churn (expansion > contraction) fit in?
- Model it via the growth input. Net negative churn means your retained accounts are expanding faster than other accounts are leaving — companies like Snowflake and Datadog have done this with NDR figures above 130%. The model has separate fields for growth and churn, so put expansion ARR INTO the growth input (alongside new logos) and put gross contraction INTO the churn input. The net effect is the same. The reason to model them separately rather than as a single "net growth" number is that the math compounds differently when growth and churn are far apart — and most importantly, gross churn is the leading indicator. Watching net growth alone hides a deteriorating retention story until it's too late to fix.
- Does this work for non-SaaS recurring revenue (subscriptions, memberships)?
- Yes — anything with monthly recurring revenue and a churn rate fits the model. Newsletter subscriptions, gym memberships, consumer subscription apps, agency retainers, even Costco-style annual memberships (treat annual churn as 1 − (1 − monthly_churn)^12, or input the monthly equivalent directly). The T2D3 trajectory is specifically a venture-SaaS benchmark — for a newsletter or a memberships business, the relevant benchmark is something like 90%+ annual retention rather than a 72× five-year curve.
- Is starting ARR what I made last year, or my current run rate?
- Current run rate — today's MRR × 12, or today's contracted ARR if you sell annual contracts. ARR is a forward-looking number: "if everything I have right now keeps renewing, this is what I'll book over the next 12 months." Last year's revenue (TTM revenue, GAAP revenue) is a backward-looking number and includes churn that already happened. The two often differ by 10–20% in a high-growth SaaS company. The calculator's starting point is the run rate.
