breme.

Research · Retention and loyalty

Is software churn decided in the first two weeks? What product usage data shows

By BREME · · 5 minute read

In short

Large product usage benchmarks agree on one uncomfortable pattern: most new users of a software product stop showing up within days, and the products that keep people in week one tend to be the ones still keeping them at three months. Amplitude reports that for half of all products, over 98% of new users are no longer active at two weeks. The data points to onboarding as the moment that matters most, but it shows a pattern, not proof of cause.

Who this is for: Customer success, onboarding and product leaders at software, SaaS and subscription companies deciding where to put effort in the first 90 days.

The question

Customer success teams often hear that churn is decided early, long before the renewal conversation. Plenty of blogs put a number on it. Many of those numbers trace back to each other rather than to a study we could find and read, so we left them out.

Instead we looked at what the large product analytics companies publish from their own usage data, plus one recent survey of customer success leaders. These sources are vendors, and their data covers their own customers, so read them as strong signals rather than settled science. Nothing here is Australian specific; all of it is global data.

What Amplitude's 2025 benchmarks show

Amplitude's 2025 Product Benchmark Report draws on anonymised data from over 2,600 companies across industries, regions and company sizes, covering September 2023 to September 2024. Three findings matter for onboarding.

First, early drop off is steep. Amplitude reports that for half of all products, more than 98% of new users are no longer active two weeks after their first action. Even top performers at the 90th percentile see day one activation of around 21%, falling to roughly 12% by day seven and about 9% by day fourteen.

Second, the bar for a good first week is low because most products fall short of it. Getting just 7% of a new cohort to return on day seven puts a product in the top 25% for activation. In enterprise products, the top 10% reach 12.4% day seven retention against a median of 2.1%.

Third, early and later performance travel together. Amplitude reports that 69% of products with strong early activation were also strong three month retention performers. At three months, top B2B technology products retain 15.6% of users while median B2B products retain 2.5%.

What Pendo and Mixpanel add

Pendo publishes benchmarks from aggregated, anonymised data across more than 6,800 applications used by 2,500 of its customers. Its best in class products, the 90th percentile, show three month user retention of 57% across monthly cohorts and a time to value of 0.2 days. The gap between that and the Amplitude median is a reminder that benchmarks depend heavily on how each vendor defines a user, an action and a product.

Pendo also reports that best in class products have a feature adoption rate of 15.6%, measured as the share of features generating 80% of click volume. In plain terms, even in the strongest products most usage sits on a small set of features. Onboarding that tours every feature is working against how people actually use software.

Mixpanel's benchmark work, drawing on its 2019 Product Benchmarks Report, found that average eight week retention is below 20% for most industries, and that over 35% retention at eight weeks is considered elite for SaaS and e-commerce.

What customer success leaders report

OnRamp's 2025 report, The First 90 Days, surveyed 161 customer success and onboarding leaders across SaaS and was released on 7 May 2025. Its key findings, as published, include:

  • 48% of customers abandon onboarding if they do not see value quickly.
  • 57% of companies that cut onboarding investment saw churn increase within six months.
  • 65% of companies using digital onboarding reduced time to value by at least 25 percent.
  • 74% of SaaS companies have adopted digital or pooled customer success models.

What the data does and does not show

It does show that the first one to two weeks are where most new users are lost, across thousands of products and more than one data set. It also shows that strong early activation and strong three month retention tend to appear in the same products.

It does not show that fixing onboarding causes better retention. A product with a clear, fast first win may simply be a better product. Amplitude's own figures are a share of products that performed well at both points, not a measured effect of onboarding. OnRamp's figures are what leaders report about their own companies, not measured outcomes.

It also measures users, not accounts. In business software, one person going quiet is not the same as a company cancelling. A renewal can survive weak individual usage for a while, especially on an annual contract. That is exactly why it is dangerous: the account looks fine on paper while the habit never forms.

Amplitude does share customer stories where onboarding changes were followed by better retention. One is Australian: it reports that SafetyCulture saw an 18% improvement in 30 day retention, alongside a 20% rise in publishing rates, after investing in AI powered conversion features. These are vendor case studies, so treat them as examples, not benchmarks.

What we would do with this

In our experience the useful move is to stop treating onboarding as a project that ends at go live and start treating the first fortnight as its own stage of the journey, with an owner and a measure.

Practically, that looks like this:

  • Name the first real win for each customer type in one sentence, and agree it with sales, product and support.
  • Measure the share of new accounts that reach it by day seven and day fourteen, not just whether setup was completed.
  • Check in on day three and day ten with something useful, not a survey.
  • Flag any account where nobody has returned in week two, and give it a named person and a date.
  • Report early activation next to renewal forecasts, so the two conversations meet.

The bottom line

The evidence does not say churn is sealed in two weeks. It says that is where most people quietly leave, and that products good at keeping people early tend to keep them later. For a customer success team that is enough to move effort forward, from the renewal back to the first fortnight.

If you are working on the first 90 days of your own customer journey, our customer success page sets out how we approach onboarding, check ins and health signals: breme.com.au/customer-success/.

Sources

See your journey from the customer's side

We map the end-to-end journey with you, find the moments that matter most, and build a strategy your teams and systems can deliver.

Start a conversation More insights