What does net revenue retention have to be?

It is the number a Series A investor asks about second, after growth, and the one founders are least likely to have calculated. It is also the number most likely to be quoted with a benchmark that does not apply.

The mechanics are simple enough to do on paper. What makes it worth a whole piece is that the obvious interpretation is frequently wrong, and that the benchmarks in circulation come from a different market with different definitions.

What it actually measures

Take the customers you had twelve months ago and ignore everyone who has joined since. Some have left. Some have downgraded. Some are paying more than they were. Net revenue retention is what that group is worth today as a percentage of what it was worth then.

Above 100% means the group is worth more than it was, without a single new customer. Below 100% means the base shrinks and every rupee of growth has to be bought.

The exclusion that matters is new business. A company can grow 60% with net retention of 80%, and frequently does. The growth rate describes what the sales team achieved; retention describes whether it will still be there next year.

Growth tells you how fast you are running. Net revenue retention tells you whether the treadmill is moving backwards.

Always report it with gross retention

Gross revenue retention is the same calculation with expansion removed. It counts only churn and downgrades, so it can never exceed 100%.

The pair is what makes either useful, and the reason is a specific failure mode. A company can have gross retention of 78%, meaning it is losing a fifth of its base every year, and net retention of 104%, because three large accounts expanded heavily. The net figure looks strong. The business has a serious retention problem that is being masked, and it will surface the moment one of those three accounts stops growing.

That pattern is common enough that experienced investors ask for gross retention first. A founder who volunteers both, and explains the gap, is answering a question that was coming anyway.

The arithmetic link is exact: net retention minus gross retention equals expansion as a percentage of the opening base. If your two numbers are close together, expansion is contributing nothing, whatever anyone reports about upsells.

What the published figures say, and why to be careful

There is no India-specific benchmark for net revenue retention. Searching for one returns American and global data, and that is what the Indian ecosystem quotes for want of anything better.

As of September 2026, surveys of private B2B software companies put median net retention around or a little above 100%, having drifted down from roughly 105% in the stronger funding years, with median gross retention in the high eighties. Public software companies at scale average considerably higher, in the 120% region, and those figures set the expectation that filters into private market conversations.

Three reasons to treat all of that as orientation rather than as a target.

• ·It varies enormously with contract size. Reported medians put enterprise accounts well above the average and small business accounts below 100%. A company selling ₹40,000-a-year subscriptions and one selling ₹40 lakh contracts are not comparable, and averaging them describes neither.
• ·Definitions differ between surveys in ways that move the number by several points: whether logos or revenue, whether downgrades count separately, how mid-period changes are treated, and what happens to customers who left and returned.
• ·The Indian market has a different shape. More companies sell into price-sensitive domestic segments, more revenue is in annual contracts rather than monthly, and the mix of enterprise to small business differs from the American surveys.

What is safe to say: gross retention consistently below 90% is treated as a signal of a product or fit problem rather than a sales problem, and net retention below 100% means an investor is underwriting your ability to keep buying growth. Those two statements survive the benchmark caveats.

A worked example

Illustrative figures, over twelve months. Opening ARR of ₹640 lakh from the customers in place at the start.

• ·Churn: customers worth ₹72 lakh left altogether.
• ·Contraction: customers who stayed reduced by ₹38 lakh.
• ·Expansion: customers who stayed added ₹96 lakh.
• ·So gross retention is (640 less 72 less 38) divided by 640, which is 82.8%.
• ·And net retention is (640 plus 96 less 72 less 38) divided by 640, which is 97.8%.

New business of ₹220 lakh took closing ARR to ₹846 lakh, a 32% growth year. A good year by most standards, on a base that shrank by 2.2% before any of that new business arrived.

The consequence is best expressed as a cost. At the same loss rate, this company needs ₹145 lakh of new and expansion ARR next year purely to stand still at ₹846 lakh. The year before, standing still cost ₹110 lakh. As the base grows, the same percentage becomes a bigger absolute number every year, and eventually the sales team spends most of its capacity replacing.

The four patterns, and what each one means

• ·Gross high, net high. Customers stay and spend more. Acquisition spend does its work once and keeps paying. This is what a strong software business looks like and it is rarer at seed than pitch decks suggest.
• ·Gross low, net high. A leaky base carried by expansion, usually concentrated in a few accounts. The most dangerous pattern, because the headline number reassures everyone while the product problem compounds.
• ·Gross high, net low. Customers stay and never spend more. Usually a packaging problem rather than a product one, and among the most fixable findings there is.
• ·Gross low, net low. Growth is rented. The business works while acquisition is cheap and stops working when it is not.

How to move it

Retention responds to different levers depending on which half of the number is the problem, and treating it as one problem is why efforts to improve it often fail.

If gross retention is the problem

Find out where it is concentrated before doing anything. Split it by plan, by contract value, by acquisition channel and by cohort. Churn is almost never evenly spread, and the usual finding is that one segment, often the cheapest and fastest-growing one, is responsible for most of it.

Then the question is uncomfortable and simple: should that segment exist at this price, with this support entitlement, or at all. A segment that churns in fourteen months and takes sixteen months to repay its acquisition cost is not a growth engine.

Beyond segmentation, the reliable interventions are onboarding, time to first value, and noticing declining usage before the renewal rather than at it.

If expansion is the problem

This is usually pricing rather than effort. A flat annual licence gives a growing customer nothing to buy, so no amount of account management produces expansion.

The structural fix is to tie price to something that grows with the customer: seats, usage, locations, volume processed, revenue handled. Companies that do this get expansion without a sales conversation, which is the mechanism behind almost every net retention figure above 110%.

Failing that, a genuine second product or a higher tier with something worth having in it. Adding a tier nobody upgrades to does not move the number.

Calculating it when your systems do not help

Most companies below Series A do not have a billing system that reports this, and the analysis is still an afternoon rather than a project.

1. 01Export every invoice or subscription charge for the last twenty-four months, with customer, date and amount.
2. 02Take the list of customers active twelve months ago. That set is fixed and nothing that joined since belongs in it.
3. 03Work out what that set was paying, annualised, at that date. That is your denominator and it never changes.
4. 04Work out what the same set is paying now, annualised, counting zero for anyone who has left.
5. 05Divide the second by the first for net retention. For gross retention, cap each customer at what they were paying at the start, so nobody can contribute more than 100% of themselves.

That last step is the one people get wrong. Gross retention is not net retention with a different formula; it is the same calculation with each customer's contribution capped at their starting value, which is what stops expansion leaking into it.

Two practical warnings. Use the date of first payment rather than contract signature, so the base is customers who actually paid. And decide in advance how you treat a customer who left and came back, because the answer changes the number and there is no convention that settles it.

When this is the wrong thing to optimise

Net retention can be improved in ways that make the business worse, and it is worth naming them.

Moving up market raises retention, because larger customers churn less. It also lengthens the sales cycle, raises acquisition cost and changes the product roadmap. That may be the right strategic move and it should be made for strategic reasons rather than to improve a metric.

Locking customers into multi-year contracts raises measured retention without changing whether they wanted to stay. It borrows from future renewals, and the churn arrives later in a larger lump.

And for a company genuinely early in its market, some churn is the cost of finding out who the product is for. A seed-stage company with 80% gross retention that has learned exactly which segment retains is in better shape than one at 95% that has never tested a second segment.

The number is a diagnosis rather than a goal. What matters is understanding why it is where it is.

How to present it to investors

1. 01Give both figures, net and gross, with the period and the definition stated.
2. 02Show it on a rolling twelve-month basis, quarterly, so renewal clustering does not distort it.
3. 03Split it by segment if you have the volume, and lead with the split rather than the blend.
4. 04Say what the trend is and why, including when the answer is unflattering. A founder who says gross retention fell four points because they moved down market and have since stopped is far more credible than one presenting a flat number.
5. 05Do not quote an American benchmark as though it were your target. Say what your own figure has done over the last four quarters.

Expect the number to be rebuilt from your raw billing data in diligence, so the definition you use should be one you can defend line by line. The ARR bridge calculator on this site produces both figures from five inputs, and the cohort retention piece covers where to look once you know which direction they are moving.

About Simplify

Simplify is a finance clarity and investment readiness practice working with founders across India, built on six years inside startups. We write about the questions founders bring before a decision, not after it.