Gross vs Net Revenue Retention: The Month-by-Month Method We Use at ORM

Gross vs Net Revenue Retention, Read Month by Month

Gross vs net revenue retention is not two flavors of the same metric. It is two different questions about the same base, and if you only report one of them you are hiding the answer to the other. Gross retention asks how much of what you started with survived. Net retention asks whether your existing customers are worth more today than they were a month ago. Teams that blend the two into a single quarterly headline lose the one signal that matters most: where the money is actually moving.

I have built forecast models for B2B SaaS companies, and the retention numbers I trust are never the ones pulled once a quarter from a board slide. They come from a monthly waterfall that reconciles to the dollar. This post covers the exact method we use at ORM: the components, the formulas, and why the distinction between gross and net changes how you read the base.

The Two Questions the Numbers Answer

Gross revenue retention measures how much of your beginning ARR you keep after every loss. Customers leave, they drop a product, they downgrade a seat count. Gross retention counts only the damage, and it can never rise above 100%. It is the purest measure of how leaky the bucket is. Net revenue retention starts from the same losses, then adds back the expansion your existing customers generated. When upsell and cross-sell outrun churn, net retention climbs past 100% and the base grows without a single new logo. That is why investors fixate on it, and it is also why it is so easy to abuse.

Here is the trap. A company can post 118% net retention and still be rotting. If three enterprise accounts triple their spend while a quarter of the mid-market walks out the door, the net number looks triumphant and the churn rate underneath it is a fire. Net retention without gross retention next to it is a vanity metric wearing a suit.

The Components: ORM's Monthly Waterfall

We call it the Monthly Reconciling Waterfall, and it is the same eight lines every month for every customer we model. The structure is deliberately boring, because boring reconciles.

The rule that makes the whole thing work: beginning ARR for any month equals ending ARR from the prior month. Nothing falls between the cracks. Every dollar that moved is on a line, and the lines sum to the ending number. That is the property most retention reporting lacks, because it lives in a quarterly snapshot where a chunk of contraction can quietly net against a chunk of expansion and disappear.

Notice the line that trips people up. New Customer ARR sits in the waterfall because your full ARR bridge needs it, but it feeds neither retention number. Retention is measured against the base you started with, not the logos you added. Fold new customers into net retention and you have built an acquisition metric and mislabeled it retention.

The Formulas, and the Line Everyone Gets Wrong

Group the middle six lines into two buckets. Contraction ARR is churned customers plus churned products plus product decreases. Expansion revenue is new products sold to existing customers plus increased product usage from that same base.

Gross revenue retention:

GRR = (Beginning ARR minus Contraction) / Beginning ARR

Net revenue retention:

NRR = (Beginning ARR minus Contraction plus Expansion) / Beginning ARR

The line everyone gets wrong is the expansion line. Teams reach for total expansion, including the ARR from brand-new customers, because it makes the number bigger. Strip it out. Expansion in the retention formula means expansion from the base only. If a new logo bought three products in its first month, that is 100% new-customer ARR and zero retention expansion. Mixing the two is the single most common way a net retention number gets inflated by accident.

A Worked Example (illustrative, not a benchmark)

Take a fictional analytics vendor, Marlowe Systems. On the first of the month its existing base is worth $10,000,000 in ARR. Over the month:

- Churned Customer ARR: $400,000 - Churned Product ARR: $150,000 - Product Decrease ARR: $250,000 - New Product ARR (existing customers): $500,000 - Increased Product ARR: $700,000

Contraction totals $800,000. Expansion from the base totals $1,200,000.

Gross retention: (10,000,000 minus 800,000) / 10,000,000 = 92%.

Net retention: (10,000,000 minus 800,000 plus 1,200,000) / 10,000,000 = 104%.

Read those together and the story is honest. Marlowe is losing 8% of its base to churn and contraction, and its existing customers are expanding enough to more than cover it. The 104% is real growth from the base. The 92% is the warning label. If you only saw the 104%, you would never fund the churn work that 92% says you need. If you only saw the 92%, you would miss that expansion is already winning. That is why the distinction changes how you read the base: gross tells you the size of the leak, net tells you the size of the pump, and you cannot run the business without both gauges.

Why Monthly Beats Quarterly, and Why Gross Comes First

Two contrarian positions, both of which I will defend.

First, read gross before net. The market trains everyone to lead with net retention because it is the number that clears 100% and looks like growth. Lead with gross instead. Gross retention is the metric expansion cannot hide behind. When gross erodes, you have a product or a segment problem that a few expanding whales are papering over, and net retention will keep smiling right up until the whales stop expanding. The health of the base shows up in gross first, every time.

Second, compute both monthly, not quarterly. A quarterly retention number is an average of three months of movement, and averages bury the trend that would have told you to act. In our data the earliest churn signal is not a cancellation, it is silence: a customer with zero support cases in a year is at real risk, the same as one with seven or more, while three to five moderate tickets usually marks a healthy, engaged account. A customer health score built on signals like that only earns its keep if you are reading retention on a monthly cadence fast enough to respond. Quarterly reporting finds the leak after the water is already on the floor.

The monthly waterfall is what makes both of those disciplines possible. It reconciles, so you can trust it. It decomposes, so you can see whether a soft month came from customer churn, product churn, or downgrades. And because beginning ARR chains cleanly from the prior ending ARR, you can run it forward as a forecast instead of only backward as a report. A retention number you can only look at in the rearview mirror is a number that arrives too late to change anything. At ORM we build the models that turn that monthly waterfall into a forward view, so you see the shape of your retention while you can still bend it.