Overview
About
Through a combination of data analytics, marketing, sales, and advertising products, Demandbase aligns GTM teams around common account intelligence, focusing brands’ time and money on accounts most likely to buy.
Challenge
Rapid growth, highlighted by M&A activity, had caused Demandbase’s cloud spend to grow at an unexpectedly rapid pace. Eyeing a new round of funding, Demandbase had to get spending under control, strengthen their unit economics, and empower engineers to autonomously manage cloud spend.
Solution
Implemented CloudZero to boost spend allocation, identify savings opportunities, and give engineers relevant, timely data about the cost of their cloud infrastructure.
Outcomes
- Reduced annual cloud spend by 36%
- Justified $175M in funding
- Fostered an engineering culture of cost autonomy
- Developed a single source of truth for all cloud stakeholders
The Challenge
B2B brands have troves of data at their fingertips — data that distinguishes interested prospects from window-shoppers, data that reveals customers’ greatest pains and deepest needs, data that instructs brands on how and when to engage.
But without a tool to separate signal from noise, all that data is little more than static.
Demandbase was founded to be that tool. Through a combination of data analytics, marketing, sales, and advertising products, Demandbase aligns GTM teams around common account intelligence, focusing brands’ time and money on accounts most likely to buy.
In mid-2020, just a few months into the pandemic, Demandbase acquired another account-based marketing (ABM) company, Engagio. “We wanted to build a product that integrated with their tech,” said Micheal Waltz, principal DevOps engineer at Demandbase, in an interview.
“The instruction to the engineers was: ‘Just get it working,’ so we invested as much as we could, getting the new product and features to market so our customers could quickly start using and obtaining value from it.”
The Demandbase/Engagio tale rings familiar in the SaaS landscape. Demandbase’s primary post-acquisition goals were: 1. Perfecting the product, and 2. Rapidly increasing market share — not keeping as close of an eye on cloud costs.“If we could position ourselves in the market, and get ourselves established, the idea was the costs would be justified,” Waltz said.
Right on cue, Demandbase’s cloud spend spiked, “like 10x, right away,” Waltz said. “We were aware of it, but getting the new product and features to market for our customers was our top priority.”
Then Waltz also started hearing, “We need to focus on our cost of goods sold (COGS). We need to know how much a tenant’s [customer’s] resource usage costs in AWS and GCP. We need to understand how much revenue they’re bringing in compared to resource usage cost so we can price more accurately.”
Making matters more complicated was the org-wide shift to Engagio’s Java-based tech stack, which leveraged in-memory analysis that greatly improved application response time.
However, in-memory analysis also requires cloud instances with enormous amounts of memory, so “we started running very large instance types, we’re talking r5.24xlarge, which run at multiple dollars per hour — very expensive,” Waltz said. (r5.24xlarge instances cost $6.048 per hour, making them some of the most expensive cloud resources.)
Moreover, all of Demandbase’s compute was based on Kubernetes, the famously flexible and infamously hard-to-manage container orchestration tool. “We ran into major scaling issues with Kubernetes,” Waltz said. “We went from 20-node clusters to 1,000-node clusters, which would also temporarily double in size during releases.
Due to the additional product features and improved application performance using Engagio’s tech, our total infrastructure size was significantly expanded and further increased cloud costs.”
By the time Waltz was tapped to lead Demandbase’s cloud cost optimization project, Demandbase’s cloud spend had surged. As the leader of Demandbase’s Cloud Center of Excellence (CCoE), Waltz had to find a way to get it under control.
“We obviously wanted to lower it, to get OpEx, COGS, and cost per customer/tenant under control,” Waltz said. “But we needed to do it without sacrificing performance or customer value.”
The Solution
Waltz took a two-pronged approach to the cost optimization project:
- Get the right people in the same room together
- Get them the data they needed to make intelligent optimization decisions
To Waltz, “the right people” meant both technical and financial representatives. “We wanted to look at the issue not just from an engineering perspective, but from a business perspective,” he said.
“I knew that whatever changes we made needed business context — they had to be informed by and measured against business goals.”
Demandbase had been CloudZero customers since 2019. “But until optimization became an organizational priority, we hadn’t devoted enough effort to making the most of the platform,” Waltz said.
But amid pressing cloud- and COGS-related challenges, that changed; CloudZero’s customer success (CS) team was integral to Demandbase’s optimization project.
Waltz worked with CloudZero’s CS team, which is made up of 100% FinOps practitioners, on several key optimization objectives:
- Boost their allocation: One of the CCoE’s key OKRs was raising their percentage of spend allocation. CloudZero worked with Demandbase to consolidate convoluted allocation logic, unify allocation frameworks from different organizations, and attribute tricky Kubernetes spend, quickly getting Demandbase over 90% allocation.
- Enter the metrics: CloudZero instituted systems for giving Demandbase the single unit metric most indicative of business health: cost per tenant.
- Empower their engineers: CloudZero developed custom dashboards around Demandbase’s key organizational metrics. The CS team then worked with Demandbase’s engineers, showing them best practices for navigating the CloudZero platform, identifying spend irregularities, and remediating them.
CloudZero brought Demandbase from spend obscurity to spend visibility, added business context to their spend data, and gave their engineers the tools for cloud cost autonomy.
Then, it was just a matter of finding savings.
The Results
36% reduction in cloud spend
The savings started with low-hanging fruit: Exploring their cloud spend environment in CloudZero Explorer, identifying resources that were still on but no longer in use, investigating why Amazon RDS spend was higher than seemed necessary.
“It’s kind of fun spelunking in your cloud spend,” Waltz said of the investigation process. “There were all these views we hadn’t seen before, and it was easy to get some quick wins.”
As Demandbase shifted out of fire drill mode and into sustain-and-monitor mode, savings shifted from reactive to proactive.
Waltz instituted biweekly meetings for members of the CCoE to review cloud spend and find savings opportunities, using CloudZero data as the basis for strategic conversations.
“We started using CloudZero data to anticipate how much we were going to spend on cloud resources in the future,” Waltz said.
“For example, we looked at ElastiCache cost trends to anticipate next year’s spend, and used estimates to figure out how much we could afford to spend on reserved instances (RIs).”
At the time of writing (May 2023), Demandbase has reduced their annual cloud spend by 36% — titanic savings, with no compromised functionality.
Helped secure $175 million in financing
In February 2023, Demandbase secured $175 million in financing from Vista Credit Partners. The savings they generated using CloudZero played a major role in securing the investment.
“It was huge for our company,” Waltz said. “And it happened because we were able to lower our cloud spend, to show that we were disciplined enough to keep our engineering expenses under control.
“Financiers want to see solid business fundamentals. They’re concerned with business health,” Waltz said. “A 36% cloud bill reduction and steady unit costs were integral to proving Demandbase’s clean bill of business health.”
Engineering culture of cost autonomy
Across repeated meetings and reviews that centered on cloud cost data, Demandbase saw their engineers take cost control into their own hands. “Our engineers are much more cost-aware than they used to be,” Waltz said.
“It’s a matter of giving the right engineers the right data, having conversations in a positive manner, and being able to quantify the big-picture impact on our business.”
Single source of truth
“We used to get a lot of questions about cloud cost,” Waltz said of the days before his cost optimization project. “Now, our leaders trust that we know what we’re doing, and that our cost data is accurate and healthy.”
Demandbase’s senior leaders use CloudZero data to inform strategic business conversations. CloudZero has played a key role in strengthening the relationship between Demandbase’s engineering and leadership teams by getting them aligned on trustworthy, controlled cloud cost data.
Going forward, Waltz expects CloudZero to continue playing a central role in Demandbase’s business. “It’s our go-to tool to understand cloud costs,” Waltz said.
“CloudZero bridges AWS, GCP, and Kubernetes, giving us a single pane of glass. It’s the first place we go to see how much something costs, and it’s a single source of truth for our finance, product, and engineering teams.”
