Annual Contract Value is one of the core SaaS metrics that every good operator measures and obsessively optimizes. But little is known about what is “good” and what is “bad” when it comes to ACV. If you Google “average saas acv” the first result is Quora, and the answer leaves much to be desired: ballpark guesses from an industry expert with no hard data.

This is very typical within the SaaS community— “benchmarks” get traded around between people on the inside, but are often based on very small and very skewed samples. This anecdotal data is better than no data at all, but what we really need is a comprehensive study of SaaS performance.

What we’ve heard about ACV

The Pacific Crest SaaS Survey is one of the most well-respected and popular studies conducted about the SaaS industry and the metrics that indicate success. Based on the responses of 306 SaaS companies, the report highlighted a median annual contract value of $21K, with 30% of participants below $5K and 20% above $100K.

While this study is immensely valuable in a number of areas, the sample is skewed towards very successful SaaS businesses that have either been sought after to complete the survey, or have volunteered their data. These companies aren’t necessarily representative of SaaS as a whole.

I personally love Tom Tunguz’s blog—his posts are wonderful, and he’s covered ACV previously. (article 1, 2, 3, 4) But, as he recognizes, his data comes from companies that have gone public; the unicorns of the SaaS world.

Changing the sample

At RJMetrics we are able to research industry trends using the aggregated data from our customer base. As it happens, we have more than a few SaaS customers who collectively have millions of SaaS subscriptions, so we set out to benchmark SaaS ACV with that data.

Our first cut at the data wasn’t particularly interesting: for the entire sample, the mean ACV was $391. That feels low based on data we have from other industry sources, and it’s not particularly meaningful without going deeper.

To add depth, we split the sample into customers of B2B and B2C companies. We know that B2B companies anecdotally charge higher prices, but we wanted to find out just how big this difference was in practice.

We found that the average B2C company had an ACV of $100, while the average B2B company had an ACV of $1,080; a 10x difference. The B2B vs B2C distinction is clearly important when looking at SaaS ACV. We decided to go one step further by looking at companies by quartile, and the results were dramatic:

Top quartile B2B companies show almost 100x higher ACVs than their B2C counterparts. So, while B2B SaaS companies certainly charge more than B2C SaaS companies, the most expensive ones charge a lot more.

ACV from SaaS IPOs

We decided to marry our dataset with IPO data for a sanity check. Out of 132 companies that have gone public in 2015, 15 have been SaaS. Eight of those companies are strictly B2B, 6 target both business and consumers, and only 1 is a purely B2C business. When analyzing their S-1 public financial information, we see a median ACV of $8,400. We split the companies into quartiles to get a better idea of the spread here.

The top quartile had an average ACV of $1,437,230, while the bottom quartile showed an average ACV of $94. This matches the trend we found in our data set, but shows an even larger difference than in our sample.

What ACV quartile a company is in doesn’t necessarily correlate with its revenue. For instance, both GoDaddy and Inovalon went public this year. Both companies are SaaS, but Inovalon reports that they have 100 customers, while GoDaddy reports 13 million. This gives GoDaddy an ACV of $106 and Inovalon an ACV of $2.7 million. Inovalon, however, reported 2014 revenues of $272 million while GoDaddy reported $1.4 billion during the same period. You can see this play out through the rest of the SaaS IPOs:

So what does ACV really tell us?

While it’s common for individual SaaS companies to make local optimizations to ACV, having a higher ACV isn’t necessarily “better”. What’s important is to architect the product, marketing, sales, and customer support experiences to align with your ACV.

We look forward to running more analyses on our SaaS customer dataset. SaaS practitioners shouldn’t need to operate with only anecdotes or poorly constructed samples to guide their sense of what is “normal”. If you want to stay in the loop on our SaaS research, you can sign up below.