ARR
Annual Recurring Revenue (ARR) is the sum of recurring revenues generated over a year in a subscription-based business. Calculated by converting pure subscription revenue (excluding one-time fees) into an annual figure, it is the most critical metric for measuring the growth potential and financial stability of SaaS and AI service companies.
Detailed explanation
Why It Matters in Tool Selection
The size of an AI tool provider's ARR is directly linked to the 'sustainability' of its service. Since operating AI models incurs massive computing costs, startups with low ARR carry high risks of sudden service interruptions or delayed updates. When adopting tools in an enterprise environment, the company's ARR growth trend serves as a reliability indicator to verify its capability to provide technical support and execute its roadmap.
What to Check
- Is it inflated by including one-time setup fees in addition to pure subscription revenue?
- In the case of usage-based pricing, is the figure calculated by converting the average usage of the past 3 to 6 months into an annual rate?
- Is the Net Revenue Retention (NRR) disclosed alongside it to support the qualitative growth of ARR?
Examples
For an AI solution company with 50 enterprise customers paying a monthly subscription of 1 million KRW, the MRR is 50 million KRW and the ARR is 600 million KRW. If they secure a new annual contract of 200 million KRW and generate 50 million KRW in one-time consulting revenue, the new ARR is calculated as 800 million KRW, excluding the 50 million KRW consulting fee.
Confusing Terms
ARR vs. Revenue
Revenue represents historical performance and includes one-time income, whereas ARR is a forward-looking metric predicting the next 12 months based on current contracts.
ARR vs. ERR
ERR (Experimental Revenue) is temporary revenue paid by customers to test AI features, which is highly volatile compared to ARR.