Person looking at a graph.
Photo: Frank Busch via Unsplash

SaaS companies need new metrics

Protocol Enterprise

Hello and welcome to Protocol Enterprise! Today: what VCs miss about SaaS revenue, Intel’s new chip play and where enterprise tech execs are going next.

Spin up

Despite the shift to the cloud, there’s no shortage of interest in data infrastructure. Global spending on data center infrastructure is expected to grow by 10% over the next five years to reach $350 billion, according to a report from the Dell’Oro Group.

SaaS metrics need a reboot

In the SaaS industry, where it’s common for a small number of customers to account for the majority of a startup’s revenue, traditional metrics like average contract value (ACV) don’t always provide the full picture.

That’s why Lightspeed Venture Partners’ Nnamdi Iregbulem developed a new metric called weighted average contract value (WACV). Nnamdi argues that WACV can do what ACV can’t: tell a startup where most of the revenue is coming from, which customers are most important and where the most risk is.

In a conversation with Protocol, Iregbulem discussed his new metric, revenue concentration in SaaS and what VCs miss about SaaS revenue.

You came up with this metric called weighted average contract value that tries to capture a lot of what we're talking about, which is [that] just looking at the average doesn't necessarily account for some of these larger customers. Why is the standard ACV metric not as useful in comparing different companies with different customers?

The reason that the standard average calculation for contract value has limited usefulness across different companies is because if the underlying concentration in these different companies is very different, then the average is just not comparable. It's not telling you the same thing.

In the same way that, again to go back to statistics, taking the average of a normal distribution is telling you something different than when you take the average of a skewed distribution. So people are implicitly assuming when they say, “Oh, this company has an ACV of this and this company has an ACV of that,” that they have very similar revenue concentration. But if they don't, then you're actually making a real mistake. But that standard average is so easy to calculate, people default to it.

What kind of insights does looking at the weighted average give a startup that only looking at the average doesn’t?

The most interesting takeaway from it is that it tells you where the revenue in your business is most concentrated. In other words, it shows what kind of customer is most responsible for the majority of your revenue. That is very interesting because it tells you where the risk in the business is, it tells you where the growth in the business likely is, it tells you whose bread needs to be buttered, so to speak, and who you really need to be paying attention to.

The average really does not tell you that. It tells you what the typical customer looks like, but not what the typical dollar revenue looks like. And so I think it's actually a really useful reframing as one is thinking about, “OK, here's all our revenue, here's where it is. Where should we be spending time? Where should we be allocating resources?” It's actually totally fair to have a more customer-centric view, too. But if you're only thinking about unit economics, or ROI, that's where having this revenue-centric view becomes really valuable.

What are some advantages for investors? As an investor, if I could see their weighted average versus their average, how might that inform my understanding of the company?

It’s a very common mistake I find among investors where they'll meet a company, the company will have X number of customers and the standard ACV will be fairly small because most of their users are either free users or in some kind of lowest-tier version of the product. But they do have a couple of meaningful customers that are spending real revenue or paying the highest tier of a product or what have you. But because there's so many total customers, their average number ends up being kind of small. And if you as an investor don't dig into that a little bit more, you can be fooled into thinking, “Oh, these guys aren't selling to enterprise-style customers, they're only focused on small, lower-quality revenue.” When it turns out actually, most of the revenue is coming from pretty high-quality customers. Unless you double-click and go a layer deeper, you'll miss that. So I think investors should be really focused on this as a metric, and should be calculating it if they have the data.

Read the rest of the interview with Nnamdi Iregbulem here.

— Aisha Counts (email | twitter)

On the calendar

So you decided to go multicloud. Now what?

It’s never been easier to use multiple cloud providers for modern tech infrastructure needs, but should you use multiple cloud providers? A panel of experts will explain the arguments for and against multicloud computing and how businesses should think about their options as the market evolves. Join us at 10 a.m. PT March 2. RSVP here.


Dataiku is the only AI platform that connects data and doers, enabling anyone to transform data into real business results — from the mundane to the moonshot. Because AI can do so much, but there's no soul in the machine, only in front of it. Without you, it's just data.

Learn more

Intel’s revamped server chip lineup

Intel is pushing back the launch of its first server chips made with its most advanced processing technology to 2024, company executives said Thursday at its investor meeting in San Francisco.

The “Granite Rapids” Xeon server processors will now launch a year later using an improved form of extreme ultraviolet lithography tech called Intel 3. For 2023, Intel said it will now launch Emerald Rapids, which will use the company’s current non-EUV tech called Intel 7.

Intel CEO Pat Gelsinger said Intel moved the launch of its chips back a year to take advantage of the improvements the company plans for Intel 3. Intel 4, the company’s first version of EUV manufacturing, will now be used for PC and cryptocurrency mining chips.

At the same time Thursday, Intel also announced a new line of low power Xeon chips for data center processing that doesn’t require the most complex computing or highest energy consumption. Called “Sierra Forest,” the chips will launch in 2024 and use efficient processing cores that appear to position the new chips to compete against some of the Arm designs that have begun to make inroads in data centers.

— Max A. Cherney (email | twitter)

Enterprise moves

Steven Rodgers isleaving Intel as executive vice president and general counsel. Rodgers is retiring effective May 31 and will be succeeded by interim pick Susie Giordano.

Edward Bush was named COO at Dataiku. Bush was formerly the VP of Finance and previously worked at WeWork. His announcement comes alongside a string of internal promotions at Dataiku.

John Zissimos is Okta’s new CMO, replacing Kendall Collins, who is among a string of executives that have left the company in the past year. Zissimos was formerly chief digital officer.

Around the enterprise

Dropbox released earnings today, announcing revenue of $2.16 billion, up 12.7% year-over-year.

Upcoming versions of Firefox and Chromemight break websites, Mozilla told developers.


Dataiku is the only AI platform that connects data and doers, enabling anyone to transform data into real business results — from the mundane to the moonshot. Because AI can do so much, but there's no soul in the machine, only in front of it. Without you, it's just data.

Learn more

Thanks for reading — see you tomorrow!

Recent Issues