Average revenue per user (ARPU) and net revenue retention (NRR) sound interchangeable – both have something to do with revenue per customer – but they answer two completely different questions.
ARPU tells you what each customer is worth right now. NRR tells you whether that number is going up or down. You need both. One without the other is half a story, and half a story is how SaaS companies talk themselves into bad strategy.
We’ve been digging through SaaS metrics for a while, and the pattern we keep seeing is that founders track one of these religiously and ignore the other. So we’ll cover the formulas, walk through worked examples with consistent numbers, lay out the benchmarks that actually matter, and finish with how to track both inside the CRM you’re already paying for. No finance-school detour.
ARPU and NRR: Two Sides of the Same Coin
ARPU is a snapshot. Take your revenue for the month, divide by your users, and you have one number that tells you what each customer is worth on average. It’s a still photo of monetisation.
NRR is a trend. It compares this month’s revenue from a group of existing customers to last month’s revenue from those same customers. It tells you whether that group is expanding, holding flat, or bleeding out.
Here’s why you need both. A company can show a flat ARPU and look stable – when in fact half their high-value customers are churning and being replaced by smaller ones. NRR catches that. Another company can show a 110% NRR and look like a rocket – but if the ARPU is $4 and they’re a B2B SaaS, the absolute revenue per customer is so low that none of it matters at scale.
ARPU is “how much is each customer worth.” NRR is “are we keeping and growing those customers.” Together they tell you whether your revenue engine is healthy.
For the rest of this guide we’ll use one fictional company – call them Picklebase, a 20-person B2B SaaS doing $100,000 in monthly recurring revenue from 250 paying customers. Every formula gets applied to the same business so you can see how the metrics interact.
What Is Average Revenue Per User (ARPU)?
Average revenue per user (ARPU) is your total revenue divided by your total number of users over a chosen period. It’s the most basic monetisation metric you can calculate – and despite the simplicity, plenty of companies get it wrong by not thinking carefully about what counts as “revenue” and what counts as “a user.”
The metric started in telecoms, where measuring revenue per mobile subscriber was the only useful way to compare carriers of different sizes. It moved into media (revenue per viewer), then SaaS (revenue per paying customer), then mobile apps (revenue per active user). Same formula, very different absolute numbers – which is why benchmarks only mean something inside a given industry.
The ARPU Formula
ARPU = Total Revenue ÷ Total Number of Users
(measured over the same period)
That’s the whole formula. The work is in defining the two inputs.
Total revenue – for a SaaS business with a single subscription product, this is straightforward: your MRR or ARR. For mixed businesses (subscription + transactional, or freemium + premium), pick one revenue stream at a time. Mixing ad revenue with subscription revenue produces a number that doesn’t tell you anything.
Total users – this is where most teams trip up. Are you counting paying users only? All registered users? Monthly active users? Each choice gives you a different metric, and they’re not interchangeable.
Worked Example: Calculating ARPU Step by Step
Picklebase at the start of the month:
| Input | Value |
|---|---|
| Total monthly recurring revenue | $100,000 |
| Total paying customers | 250 |
ARPU = $100,000 ÷ 250 = $400 per customer per month
That’s the headline number. Each customer generates $400 a month on average – or $4,800 per year if everyone stayed for twelve months (they won’t, which is why we need NRR).
Now, this is blended ARPU – all customers averaged together. If Picklebase has a starter tier at $99/month and an enterprise tier at $999/month, a $400 average tells you roughly how the mix splits, but it hides the fact that 80% of revenue might come from 20% of customers. Most SaaS founders find segmenting ARPU by plan tier or customer size more useful than the single blended number.
ARPU vs ARPPU (Average Revenue Per Paying User)
If your product has a free tier, the question of who counts as a “user” gets interesting fast.
ARPPU – average revenue per paying user – divides total revenue by only the customers who actually pay you. ARPU divides by everyone, including free users.
For Picklebase, those numbers are the same because there’s no freemium tier. But imagine a different company with 10,000 free users and 1,000 paying users at $50/month:
- Total revenue = $50,000
- Blended ARPU = $50,000 ÷ 11,000 = $4.55
- ARPPU = $50,000 ÷ 1,000 = $50.00
Which number is “right”? Both. ARPU tells you how well your overall monetisation is working – what each registered user is worth, including the dead weight. ARPPU tells you what a converted customer is worth. Track both, and the gap between them tells you how much upside you have if conversion improved.
Note: many companies use “ARPU” loosely to mean ARPPU. Always check what’s in the denominator before comparing your number to anyone else’s.
What Is a Good ARPU? Benchmarks by Business Model
Benchmarks vary by an order of magnitude across business types, so absolute numbers from outside your category aren’t useful. Some directional reference points:
| Business Model | Typical Monthly ARPU |
|---|---|
| Consumer streaming (Netflix, Spotify) | $5 – $15 |
| Mobile games (free-to-play, blended) | $0.50 – $5 |
| Consumer subscription apps | $5 – $20 |
| SMB SaaS (self-serve) | $30 – $150 |
| SMB SaaS (sales-assisted) | $200 – $1,000 |
| Mid-market SaaS | $1,000 – $5,000 |
| Enterprise SaaS | $5,000 – $20,000+ |
Spotify reported €4.62 per user per month in 2024 – and that’s considered solid for consumer audio. A $400 ARPU at Picklebase looks middling against mid-market SaaS but excellent for SMB SaaS. Your category is the only benchmark that matters.
What you’re really looking for is the trend: is ARPU growing month over month, holding flat, or shrinking? Growing usually means upsells are working, pricing power is intact, or the customer mix is shifting toward higher-tier plans. Shrinking is a warning – and that’s where NRR comes in to tell you exactly which.
What Is Net Revenue Retention (NRR)?
Net revenue retention measures how much recurring revenue you keep – and grow – from your existing customers over a period. It deliberately excludes any revenue from new customers. The point is to isolate the health of the customer base you already have.
If 100% of your customers stayed and paid exactly what they did last month, your NRR would be 100%. If some of them upgraded, your NRR exceeds 100%. If some downgraded or churned, NRR drops. Whether the number is above or below 100% tells you whether your existing customer base is expanding or contracting in revenue terms – independent of how many new logos you’re closing.
This is why investors care about it so much. New customer acquisition can mask retention problems for years; NRR can’t.
The NRR Formula
NRR = (Beginning MRR + Expansion − Downgrade − Churn) ÷ Beginning MRR × 100
The four inputs:
- Beginning MRR – what your existing customers were paying at the start of the period
- Expansion – upsells, cross-sells, seat additions, plan upgrades from those same customers
- Downgrade – same customers paying less than they were (smaller plan, fewer seats)
- Churn – same customers who cancelled entirely (see our deep-dive on customer churn rate for the full definition and calculation)
You compare the ending revenue from that same cohort back to the beginning revenue from that same cohort. New customers acquired during the period are not in the calculation. That’s the whole trick.
Worked Example: Calculating NRR Step by Step
Back to Picklebase. Same starting position: $100,000 MRR from 250 customers. During the month:
| Movement | Amount |
|---|---|
| Expansion revenue (upsells, more seats) | +$8,000 |
| Downgrade revenue (smaller plans) | −$2,000 |
| Churn (full cancellations) | −$3,000 |
| Net movement | +$3,000 |
Ending MRR from the same cohort = $100,000 + $8,000 − $2,000 − $3,000 = $103,000
NRR = $103,000 ÷ $100,000 × 100 = 103%
Picklebase grew its existing customer base by 3% in a single month – before any new customer revenue is counted. That’s healthy.
Those $3,000 churn losses don’t go away next month. They’re permanently out of the base. The 103% says “this month was good”; it doesn’t say “we’ve solved retention.” Look at NRR as a trailing-twelve-month figure to see the actual trajectory.
Is NRR the Same as Net Dollar Retention?
Yes. Net revenue retention and net dollar retention (NDR) are the same metric and the same formula. Net dollar retention is the term most common in US tech and SaaS investing – Snowflake, Datadog, and most publicly listed SaaS companies report NDR in their earnings. Net revenue retention is the term most common in customer success literature and outside the US.
Different acronyms, same number. Don’t get confused by a board member asking for “NDR this quarter” when you’ve been tracking “NRR” – they’re asking for the same thing.
What Is a Good NRR? Tier Breakdown and Benchmarks
This is the question every SaaS founder eventually asks. The honest answer involves tiers, not a single number:
| NRR Range | What It Means |
|---|---|
| 120%+ | World-class. Usually enterprise SaaS with strong expansion motions. Snowflake famously hit 158% pre-IPO. |
| 110–120% | Excellent. Top quartile for SaaS. Strong product-led expansion or land-and-expand sales. |
| 100–110% | Healthy. Your existing customer base is net-growing. Median for private SaaS sits in this range. |
| 90–100% | Watch closely. You’re holding steady at best – and you’re depending on new acquisition for all growth. |
| 80–90% | Leaking. Existing customers are net-shrinking faster than expansion can offset. |
| Below 80% | Serious retention problem. No amount of new acquisition fixes this long-term. |
For benchmarks, SaaS Capital’s annual survey of private SaaS companies consistently shows the median NRR around 102–106%. KeyBanc’s SaaS Survey reports similar numbers, with top-quartile companies above 115% and bottom-quartile below 95%. These are public, citable, and updated yearly – use them rather than vendor proprietary studies.
A note on growth stage. Early-stage SaaS (under $1M ARR) often shows lower NRR because they haven’t built expansion motions yet – and that’s fine. Late-stage SaaS without a strong NRR is a structural problem, because once new customer acquisition slows, there’s no second engine.
NRR vs GRR: What Gross Revenue Retention Adds
NRR is the headline metric, but it can hide problems. That’s where GRR comes in.
Gross revenue retention (GRR) uses the same formula as NRR – but it ignores expansion. It tells you only how much revenue you kept from existing customers after churn and downgrades. By definition, GRR can never exceed 100%. It’s the leak in your bucket, measured without the offsetting refill.
GRR Formula and Worked Example
GRR = (Beginning MRR − Downgrade − Churn) ÷ Beginning MRR × 100
Picklebase, same month:
GRR = ($100,000 − $2,000 − $3,000) ÷ $100,000 × 100 = 95%
So Picklebase has 103% NRR and 95% GRR. Both are true. The difference between them – 8 percentage points – is the expansion revenue masking a 5% revenue-churn hole.
That’s the value of tracking both together. NRR alone says “we’re growing.” GRR alone says “we’re losing 5% every month.” The pair says: “we’re growing because expansion is outpacing churn, but if expansion ever slows, we have a leak to fix.”
When to Watch GRR Instead of NRR
For most SaaS conversations, NRR is the headline. But there are two cases where GRR is the metric you actually need.
Early-stage businesses – when you have very few customers and one upsell can swing NRR by 20 points, GRR is a more honest read of retention. Strip out the noise.
Land-and-expand pricing models – where you deliberately under-price the initial sale to drive expansion later, NRR will look amazing because expansion is the whole pricing strategy. GRR tells you whether your initial customers are actually sticking around long enough to expand.
Investors who understand SaaS will ask for both. Get into the habit of reporting them side by side.
Renewal Rate: The Third Retention Metric
Renewal rate is the third leg of the retention stool, and the simplest of the three. It measures the percentage of customers (or contracts) up for renewal in a period who chose to renew.
Renewal Rate = Customers Who Renewed ÷ Customers Up for Renewal × 100
Note what’s missing: dollars. Renewal rate is counted in logos, not revenue. A 90% renewal rate means 9 out of 10 customers who had to actively renew did so, regardless of whether they renewed at the same plan, upgraded, or downgraded.
This is useful as a complement to NRR and GRR. NRR can be 110% even if you’re losing 30% of customers – as long as the remaining 70% expanded enough to overcome it. That’s a fragile business. Renewal rate catches it: a low renewal rate plus a high NRR means a small number of accounts are propping everything up.
For most subscription SaaS, healthy annual renewal rates are 85–95%. Below 80% is a warning sign. Enterprise SaaS with multi-year contracts typically reports higher renewal rates because the friction to cancel is higher – that doesn’t mean those customers are happier, just locked in longer.
How ARPU and NRR Interact (and What That Tells You)
This is the section nobody writes, and it’s the most useful one.
Pair ARPU and NRR together and you get a 2×2 read on the health of your business:
| ARPU growing | ARPU shrinking | |
|---|---|---|
| NRR > 100% | Healthiest case – existing customers paying more, and the customer mix is shifting up too. Keep doing what you’re doing. | Customer base is expanding but new customers come in cheaper than the existing ones. Check whether you’re discounting to chase volume. |
| NRR < 100% | Existing customers shrinking but average customer is worth more – usually means low-value churn (the right customers staying, wrong ones leaving). Sometimes intentional. | Worst case. Customers leaving and the ones who stay are paying less. Pricing or product problem. |
A useful pairing for unit economics is ARPU against customer acquisition cost (CAC) and against customer lifetime value (populate after publish). If monthly ARPU is $400 and your CAC is $1,200, you need 3 months of payback before each customer breaks even – and your NRR tells you the probability they’ll stick around that long.
The 2×2 above is a starting frame, not a verdict. Real decisions need the third metric (CAC, churn rate, or growth rate, depending on the question) – but reading ARPU and NRR together rules out about half the bad strategic moves a SaaS founder might make.
How to Track ARPU and NRR Inside Your CRM
You don’t need a separate billing analytics platform for this. If you’re running on a CRM with the deal and account data you already collect, you have most of what’s needed. (For the broader picture, our guide to CRM metrics and KPIs covers the wider set of numbers worth tracking alongside ARPU and NRR.)
The data points you need:
- Customer-level MRR or ARR – stored as a custom field on each Account or Deal record
- Account creation date and churn date – most CRMs track this automatically
- Expansion deals – flagged separately from new business (a custom field on Deal: “new business” vs “expansion” vs “renewal” vs “downgrade”)
- Active customer count – derived by filtering accounts with active subscriptions
In Pipedrive, you can build a custom report that segments deals by stage and pulls account MRR – Pipedrive’s reporting handles ARPU calculations natively if you set up the custom fields correctly. The harder part is tagging deals as expansion vs new business, which requires sales discipline.
In HubSpot, the Revenue Analytics tools (Sales Hub Enterprise) include built-in NRR cohort views – but smaller teams on Pro tier can reproduce the same number with a custom report joining Deal and Company objects, plus a deal type property.
In Monday CRM and other lightweight CRMs, you’ll likely need to export to a spreadsheet monthly and run the calculation there. That’s fine – what matters is consistency, not automation. The numbers tell the same story whether they come out of a SQL query or a Google Sheet.
For an honest read on which platform fits your stage, see our best CRM for small business breakdown – the right pick depends as much on your reporting needs as on pipeline management.
The discipline that makes this work isn’t the tool. It’s tagging every closed-won deal correctly so that “new business” stays separate from “expansion” stays separate from “renewal.” Without that distinction, you can calculate revenue, but you can’t calculate NRR. Most CRMs that fall short on retention reporting aren’t lacking features – they’re lacking the data entry discipline upstream.
How to Improve ARPU and NRR – 7 Practical Moves
The strategies that move these metrics aren’t glamorous. Most are about doing the basics better.
- Segment your pricing tiers and look at the gap. If 80% of customers are on your lowest tier, your pricing structure is doing the work of compressing ARPU. Either the middle tier isn’t compelling enough or the entry tier is too generous. Either way, that’s where ARPU comes from.
- Build an expansion playbook into customer success. NRR above 100% only happens when somebody is responsible for spotting upsell opportunities – usage spikes, new use cases, team growth at the customer. Make this a named role with a quota, not a “if we have time” task.
- Reduce contraction, not just churn. Downgrades hurt NRR less spectacularly than churn but accumulate just as fast. Watch for accounts moving from annual to monthly billing, dropping seats, or moving down a tier – these are warning signs that often precede full churn.
- Use renewal as an upsell window. Renewal conversations are the highest-leverage upsell opportunity in SaaS because the customer is already evaluating the relationship. Going into a renewal with a clear list of features they’re not using (or new ones they’d benefit from) converts renewals into expansion.
- Price by value, not by feature count. ARPU grows fastest when pricing is tied to a metric the customer cares about – number of users, volume processed, revenue managed – rather than a feature list. Value-based pricing scales with the customer’s growth instead of capping at a flat plan.
- Fix onboarding before paid acquisition. Customers who hit their “aha moment” in the first 30 days have dramatically better retention. If GRR is below 90%, look at first-month activation rates before you spend more on ads. The leak is upstream of the marketing budget.
- Track ARPU and NRR by acquisition channel. Different channels bring in different customer types. Customers from inbound demo requests often have higher ARPU and NRR than customers from cold outbound. Knowing which channels produce expansion-ready customers – and which produce churn-prone ones – is the highest-ROI use of these metrics combined.
FAQ: ARPU, NRR, and SaaS Retention Metrics
What is ARPU in simple terms?
ARPU (average revenue per user) is total revenue divided by total users over a period. It tells you, on average, how much money each user generates.
What is NRR in simple terms?
NRR (net revenue retention) measures how much recurring revenue you keep and grow from your existing customers month over month, including upsells and subtracting downgrades and churn. Above 100% means the customer base is net-expanding.
Is NRR the same as net dollar retention?
Yes. NRR and NDR are the same metric, same formula. Different terms used in different contexts.
Can NRR exceed 100%?
Yes – and that’s what you’re aiming for. NRR above 100% means upsells and expansions outweigh downgrades and churn, so existing customers generate more revenue this month than last.
What is the difference between NRR and GRR?
GRR excludes expansion revenue, so it can never exceed 100%. NRR includes expansion, so it can. GRR shows your retention floor; NRR shows your net retention performance.
What is a good NRR for SaaS?
For private SaaS, the median sits around 102–106% (SaaS Capital, KeyBanc benchmarks). Above 120% is world-class, 100–120% is healthy, 80–100% is leaking, and below 80% is a serious problem.
How is ARPU calculated for a freemium product?
Two ways. Blended ARPU divides total revenue by all users including free. ARPPU divides by paying users only. Most freemium businesses track both – blended shows monetisation rate, ARPPU shows what a converted user is worth.
