top of page

KPIs: How to know what's working

Not all revenue is created equal - prioritise the relevant parts and define metrics carefully

Revenue for B2B software companies comes in different shapes and sizes, including:


1. Structurally recurring revenue from software license income

2. Repeating (but not truly recurring) revenue streams e.g. from usage credits or ancillary embedded services

3. Non-recurring revenue, which can include professional services, consultancy or implementation fees


Investors generally attribute a higher value to recurring than repeating than non-recurring revenue because of better predictability. However many businesses will opportunistically accept non-recurring revenue when customers request it, particularly when the request is from a key client. A key demonstration of GTMF is knowing when to say no to custom, off-strategy work. It can be helpful to develop decision-making criteria for these ad-hoc client requests. GTM teams should work with product/technical leaders to systematically decide which revenue generating activities to focus on.


To muddy the waters further, there are also different flavours of revenue from an accounting perspective.

  • Bookings refer to the total value of a contract, often referred to on an annual basis (which can incorporate multi-year deals).

  • Billings refer to the actual cash invoice related to the revenue stream, which are often annual, quarterly or monthly.

  • Recognised Revenue under GAAP spreads a revenue stream proportionately over the life or duration of a contract.

Recognised revenue for a straightforward SaaS license is closely related to the key revenue metric for SaaS businesses: Annual Recurring Revenue (ARR). ARR is traditionally defined as the annualised value of the currently paying, deployed, and contracted customers. Accounting definitions of recognised revenue might differ slightly from this due to inter/intra-month variations and contract start dates.


One of the recent trends in the market is companies presenting Contracted ARR (CARR) as opposed to ARR, where CARR incorporates impending revenue from clients who have signed a contract but have not yet gone live. Consequently CARR exceeds ARR, with a lag as ARR catches up when clients go live. The gap between these two metrics needs careful tracking and presentation. This gap can be particularly contentious for SaaS businesses with long implementation or onboarding cycles, where ARR can be a low percentage of CARR.


It can be helpful to keep a detailed contract bank of licenses, with reconciliations between this document and your regular SaaS metrics reporting around CARR / ARR, and a bridge to traditional financial statements.



Anchor 2

Foundational unit economics cover the essential 'need to knows'

Over the last few years, there has been a significant increase in the range of SaaS metrics reported by private and public companies.


Foundational unit economics terms are based around lifetime value (LTV) and customer acquisition cost (CAC), with related metrics of LTV:CAC ratio, magic number and CAC payback period. These are now widely calculated for nearly all private SaaS businesses, with off-the-shelf benchmarks readily available for public businesses.


However, the devil is in the detail when it comes to definitions of these summary terms. Below we’ve provided a cheat sheet with precise definitions to ensure consistency and comparability.



ARPA = Average Revenue Per Account

GM = Gross Margin (see below)

Churn = Gross Churn (often on a volume or logo basis) over the same period of time as your ARPA


For example, a business with $40k ARPA and 75% GM with 10% annual  logo churn means clients on average have a 10y lifetime, and the LTV is $300k.


N.B.1 Using net churn, not gross, can result in infinite lifetimes. There are several ways to solve this. Companies can calculate a 3 or 5 year fixed period LTV. But more precise calculation is possible by using cohorted gross logo churn to accurately predict lifetime length and incorporate the impact of net upsell on ARPA over time.


N.B.2 Be careful here if your Average Selling Price (ASP) significantly differs from your ARPA, which might be the case if you are a PLG business rapidly moving upmarket. This can dramatically bias your LTV calculation.



S&M related expenses should include all S&M and an apportionment of operating costs used by the S&M team to be truly ‘fully loaded’ - though some companies exclude parts of senior resource or experimental/brand spend.


Most businesses calculate this on a 'new client only' basis.


N.B.1 Many businesses will lag their S&M expenses when calculating this figure by approximately the length of the sales cycle. This might be 1 month for SMB SaaS, but can be 1 or 2 quarters for enterprise SaaS. Aligning the spend to the sales cycle ensures the acquisition cost is lined up with the clients acquired through that expense, and therefore should tally with analysis conducted on pipeline conversion, or rep productivity etc.


N.B.2 S&M costs aren’t the relevant acquisition costs for all businesses. For example, PLG businesses will have artificially low CAC when considering only S&M.


Cost of Goods Sold (COGS) or COS aren’t the most clearly defined terms in the SaaS world. Whilst the general accounting for gross margin has improved at the later stages of private and public markets, there's a high level of variability in the earlier stages.


Our view is that a reasonable measurement of your COS should include:


  • Hosting/infrastructure

  • Data costs

  • Professional services

  • Part of customer support / success


N.B. The CS team is a challenge here. Costs associated with onboarding / servicing / gross retention should be included in COS, but any CS staff who are revenue generating sit more naturally in S&M opex. Our loose rule of thumb is any CS that have revenue-linked quota targets move below GM into S&M.


But off-the-shelf KPIs can be misleading, particularly for modern GTM approaches

There are three popular KPIs that are very commonly tracked. These three KPIs are worth calculating because they are pervasive and well benchmarked, but it’s important to understand their limitations. These limitations are particularly apparent when it comes to newer GTM approaches such as product-led or community-led growth.


This metric is the ROI-style ratio of lifetime value to the relevant acquisition cost. Generally, 3:1 is considered a good threshold to aim for, as a unit customer’s profitability should give sufficient margin to pay for other overheads and allow the business to be theoretically profitable at scale.


A key issue is that actually achieved lifetime figures often aren’t available until a SaaS startup is >3 years old. Reliable lifetime predictions can be elusive and unreliable, particularly in new emerging categories.


Another issue is businesses with negative net churn have implied infinite lifetimes. Using fixed-period ‘lifetimes’ can help, or using gross cohort churn rates to statistically estimate projected lifetimes.


It can also be confusing how to incorporate upsell. Our recommendation is to make sure your required CS cost is included either in GM or by adjusting CAC. Otherwise this results in highly spurious ratios, particularly for enterprise SaaS companies, or business with very high upsell rates.


This is such a common and challenging KPI 'problem' we have written a detailed ‘how-to’ article on this piece for SaaStock, available here.


New ARR / Total S&M

This simple 'rule of thumb' metric provides a measure of the ratio between sales and marketing investment with new revenue.


Different versions of 'new revenue' are used, so be careful when comparing benchmarks – some use only new revenue from new clients, others net new revenue incorporating upsell and downsell, and some benchmarks use ARR while others just differences in GAAP revenue.


The periods of new revenue vs S&M are sometimes lagged to account for sales cycles.


For businesses with a traditional sales-driven GTM strategy 0.7+ is considered healthy while 1.0+ is considered strong. Magic number calculations for PLG businesses are often extremely high, as new acquisition isn’t driven exclusively by S&M, so the denominator doesn’t represent the ‘true’ cost driving the revenue. Consequently this benchmark isn’t widely used for PLG driven businesses.


This statistic is broadly helpful, but it takes a very simplistic view of sales and marketing efficiency and has several issues. Nnamdi Iregbulem at Lightspeed wrote an excellent article on the shortcomings of this metric available in the Further Reading box below, which includes omitted/confounding variables, reverse causality and observation vs intervention issues.



This popular metric calculates the time to recoup upfront CAC investment with gross profits (i.e. marginal client contribution margin) from a new client. It’s consequently a key measure of capital efficiency and a critical determinant of SaaS burn.


SaaS/math aficionados will spot that this is the inverse of the magic number (defined above) scaled by GM.


Some investors will consider payment terms to compare “Cash CAC payback”, because annual or longer client payment terms can mean a 23-month theoretical payback period actually pays back on a cash basis in ~12 months.

Benchmarks here vary by size of client (e.g. enterprise is typically 12-24 months, SMB often <12 months) but, again, models such as PLG will see these as unusually low.


N.B. We often use current ASP rather than ARPA in this calculation, so the payback period more accurately reflects projected payback periods of clients being won today. This is something to be aware of if your business is rapidly moving upmarket, when ASP might be significantly higher than overall ARPA.

It's really important to be aware that these 'traditional’ metrics can give biased results depending on your GTM strategy, and that averages themselves are blunt and can obfuscate from what's really going on under the surface.

You might find these metrics less helpful if your GTM strategy is focused on PLG, or enterprise sales with high levels of upsell, or community driven adoption, for example.


Increasingly we're finding that more thorough analysis and customer segmentation is needed for accurate and reliable performance insights.



Cohort economics can shine light on changes over time and distinctiveness between GTM strategies

Creation of 'cohorts' facilitate analysis on the performance of the same group of clients over time. Cohorts are typically grouped according to some common attribute, for example a the quarter that clients were acquired to allow comparison over time, or client characteristics (such as SMB/mid market/enterprise, or self-serve/assisted sale, etc) to allow cross-sectional comparison between distinctive groups.


Below we present three of our favourite analyses which can be particularly helpful when assessed on a cohort basis.

Increasing ARPA can be a powerful driver of revenue growth, and is particularly crucial for PLG businesses moving upmarket.


It’s useful to see what’s driving ARPA growth. Looking at ASP vs churned ARPA (as well as ARPA of up and down-sold clients) is often illuminating.

In the illustrated example, overall ARPA has more than doubled from $5k to $10k over 4 years. This has been driven by relentless increases in ASP for newly won clients, as this increases from $10k to $40k, whilst the business consistently churns clients with smaller ARPA.

The Quick Ratio is measured as the total absolute MRR increase (new & upsell) as a proportion of MRR decreases (downsell & churn).


The ratio is a measure of MRR growth, and can be shown as a comparison to its components.

This business had a healthy quick ratio (black line, right axis) due to strong new MRR growth, which pared in Q5-6, as churn and downsell increased. Looking at this analysis for different groups of customers will give a clear view on the drivers of your MRR growth, and potential improvements.

Quick ratios of 3 are respectable, and 5+ is better. The metric is a helpful indicator of not only MRR growth momentum, but also MRR growth quality.

Net Revenue Retention (NRR) defined as Monthly Recurring Revenue (MRR) from a cohort in period Xn divided by MRR from the same cohort in period X.


NRR is arguably most VCs favourite SaaS metric! It encompasses all upsell/cross-sell, downsell and churn.

In the illustration shown, you can see 4 MRR cohorts (note: use MRR, not accounting revenue). This shows that while on average customer cohorts have healthy NRR, the direction of travel is deteriorating, with recent (shorter) cohorts underperforming.

The gold standard for cohort economics requires taking some of the average metrics you will be familiar with and applying them to individual cohorts.


This approach can add a lot of value for companies with more diverse GTM strategies and customer bases. Many SaaS businesses are increasingly pursuing dual-GTM strategies.


For example, a SaaS company could have high value enterprise customers with complex requirements who are supported with a more traditional sales organisation; meanwhile high velocity, lower value, user driven adoption would come through a highly automated, PLG approach. The cohort behaviour and economics of these customer groups will clearly differ significantly, as will the team structures to serve these different customer groups. Splitting your customer base will allow you to see important nuances in behaviour and economics between the different groups.

Calculating cumulative gross margin per cohort and comparing against that cohort’s acquisition costs allows you to be able to calculate actually realised CAC payback and lifetime values. This is the backwards-looking gold standard for how your business has performed historically. It can be hugely insightful for commercial planning decisions on how much to ramp up spend, as well as for financial planning and budgeting.​

Cohort economics can also show you how you’re actually tracking towards your theoretical LTV

BONUS: Download Oxx's SaaS metrics, unit economics and cohort analysis template

Through analysing the KPIs and metrics of (literally) thousands of SaaS businesses, we have developed a comprehensive template which we’re making available for anyone to use.

This template only requires 4 data inputs to work: your MRR per client data, sales and marketing costs (or customer acquisition costs), gross margins and monthly burn.


Inputting just these 4 datasets allows you to immediately and automatically calculate a plethora of useful visualisations of your metrics, using the standardised off-the-shelf definitions.

The core analysis includes overall MRR/ARR growth, movements in client numbers and ARPA, underlying drivers of ARPA, quick ratios, core unit economics including CAC payback and LTV/CAC, magic number, as well as retention stats (gross and net), burn metrics, and a full suite of cohort analysis, including cohorted gross and net retention, and cohorted gross profit vs CAC.

Critically, the template also allows you to automatically split the entire customer base into 2 cohorts at a time (eg self-serve vs assisted, SMB vs enterprise, Europe vs North America etc.) and allows you to see almost all the analysis described in this section that's not related to sales and marketing costs (as cohort attribution here is tricky to unpick).

Download Oxx's SaaS metrics, unit economics & cohort analysis template

Please check out this link for 2 videos on how to use & understand the template.

We hope you find this template helpful. If you have any questions or issues downloading or using the template, you can reach out to us on And if you want to discuss it further you can reach out to Phil, who created the template - he's always happy to chat!

Create an operational GTM dashboard and allocate owners to ensure consistent impact of performance metrics

Having reliable and timely visibility over a comprehensive suite of performance metrics is crucial to successfully working towards GTMF. Building out a dashboard of commercially relevant sales and operating metrics is therefore an essential step. Increasingly, businesses are also incorporating product usage and engagement related metrics in these dashboards.


Running Minimum Viable GTM tests and tracking the impact across metrics can feel like a game of whack-a-mole. When one indicator pops up, others disappear. The best way to understand the correlations and associated actions in detail is to incorporate comprehensive visibility, and get as close to real-time as you can. These make for great 'war room' dashboards on communal screens in the (physical or virtual) office!


This process is as much about culture as it is about having visible metrics in place. Build a culture where the entire team understands the importance of tracking and taking action off the back of key metrics. Team leads should feel a sense of ownership, which allows GTM teams to more rapidly course-correct when things are drifting.


Taken to its extreme, some ultra-responsive SaaS sales teams use true real-time indicators from prospective clients (eg browsing behaviour on-site, interactions with chatbots, etc) to prioritise the highest value leads, shortening the feedback loops and driving sales.

Section 2 - Dashboard.png

As PLG has become a more popular GTM strategy we've seen an explosion in the depth and range of tooling in the GTM stack. This reflects the need for more real-time integration between customer behaviour, product teams and commercial teams.

There’s now a wide variety of tooling for not only product and user behaviour measurement (such as Pendo or Segment), but across the product development/testing stack, community support, demand generation and customer support.

Often the fastest growing software companies we see are also the ones with the most comprehensive real-time reporting in place. This reporting flows into the relevant product and commercial teams, with agile feedback loops to iterate both product design and commercial outreach, responding to customer signals faster than ever.


bottom of page