How NinjaCat Turned Cost Intelligence For AWS And Snowflake Into Fuel For Business Decisions

The Challenge

NinjaCat offers marketing performance management solutions to agencies, brands, and media companies across a range of industries. Through dashboards, reports, presentations, and more, NinjaCat’s tools allow their customers to unify and leverage a broad swath of marketing data. Customers can select from more than 100 integrations (e.g., Facebook, Google, Bing, The Trade Desk, and customer calls), gleaning detailed insights on the effectiveness of their campaigns.

NinjaCat’s customers range dramatically in size. Their smallest customers use only one integration, whereas their largest customers use dozens. Because part of NinjaCat’s pricing model includes number of integrations and volume of data, that size variability gets reflected in NinjaCat’s pricing.

“We have a broad range of customers, which means we have a broad range of how much customers are paying us,” said NinjaCat’s IT Manager, CJ*. “We knew with more visibility into what each customer was costing us, we know we can make better decisions to align our cost models and our pricing.”

*Note: For online privacy reasons, the subject of this case study has asked that we not use his real name. CJ is a pseudonym.

From a COGS (cost of goods sold) perspective, CJ said, the variation in cost comes down to data. NinjaCat builds on AWS and Snowflake to bring value to their customers. The heaviest users process significantly more data — exponentially more than NinjaCat’s lightest user.

Like many companies new to cloud cost management, NinjaCat got bills from AWS and Snowflake at the end of each month, and whatever the bottom line, the company attributed it to the cost of doing business. Without more granular visibility, NinjaCat had no way to understand the business context behind changes in cost — especially to Snowflake, which, CJ said, fluctuated much more unpredictably than AWS.

The lack of context made it challenging to have productive conversations between engineering and finance. “Engineers are driven to get high quality products out the door quickly — and they want to use whatever resources they need to build the best product,” CJ said. “But the CFO and the Board understandably have a different perspective on cost. We just need to make sure we’re marrying those two perspectives, so we’re making the right trade-offs for the best technical and business decisions.”

As NinjaCat kept adding customers, their cloud cost steadily mounted. They decided they needed to: 

• Gain visibility into what was driving increased cloud cost, while tying together AWS and Snowflake spend

• Develop methods for giving engineering the financial context of their decisions, and giving finance clear explanations for cloud cost trends, without slowing them down

• Understand their cost per individual customer, so they could ensure they were maximizing the value of each customer segment

The Solution

CJ had previously worked for another cloud cost vendor, so he came into NinjaCat understanding the importance (and potential cost savings) of cloud cost management.

But his boldness paid off. CJ and his team implemented CloudZero, a cloud cost intelligence platform that delivers relevant cost data to engineering teams about the products and features they’re building.

“I’m a very simplicity-minded person,” CJ said. “I like direct, simple, intuitive. I don’t like to have to hunt for things. From the initial demo of CloudZero, the first thing I was struck by was that I wasn’t overwhelmed. I could easily come into the product not knowing anything about it, and as long as accounts were linked, I could immediately start clicking around, finding trends, and gaining cost accountability insights.”

The Results

Achieved lower costs as they scale

Immediately, NinjaCat used CloudZero to identify anomalies and spend leaks, the “low-hanging fruit” that companies can quickly clear out — once they finally see it. For NinjaCat, that meant an immediate 6% reduction in their cloud spend, just by identifying a single anomaly.

“For four years, we’d been running an EC2 snapshotting product to do all the backup of our EC2 infrastructure,” CJ said. “But a year ago, something broke, and none of the snapshots were being pruned. CloudZero revealed that the cost for snapshots was steadily increasing. It started at a couple bucks a day, and over time, it rose — all of a sudden, we were spending thousands per month on our EBS cost.”

By redoing their snapshot lifecycle policies, NinjaCat was able to get that cost down to about 80% less per month. Through a series of similar initiatives, NinjaCat reduced their overall cloud cost by nearly 40%.

As they continue to scale, more efficient infrastructure means more autonomy over their cloud cost trends.

Gained context to drive business outcomes 

“When you’re experiencing growth like us and there’s a sudden influx of customers,” CJ said, “you need to make adjustments to infrastructure to keep up — and it’s a common practice to just throw resources at the problem. But changes made to handle the scale in the moment, that were meant to be temporary, end up becoming permanent.”

CloudZero enabled CJ to attribute specific costs to individual product families and investigate where they could “trim the fat” on existing infrastructure, as well as enable better decisions moving forward.”

“Now that my Snowflake and AWS spend is organized according to our business, I can look at our different products and features and explore usage history and see what specifically is driving cost trends. I also get anomaly alerts, which gives us guardrails to move fast without overspending.”

Rather than relying on a lump-sum bill for the previous month of cloud usage, CJ could get a real-time look at ongoing cost trends, bring them to the pertinent engineering team, and see if they were necessary. “CloudZero ended up facilitating better conversations between different stakeholders to determine whether increases are healthy. Some are, some aren’t.”

CloudZero enabled an important first step toward building a common language for finance and engineering. By showing engineering teams the costs of various initiatives (as well as unnecessary instances that were driving up monthly expenses), NinjaCat could facilitate a culture of accountability, in which engineers understand the financial implications of various actions, and adopt new habits to manage cost.

In return, it made it easier for engineering to afford the tools they needed. “The worst thing for an engineer is to be forced into using substandard tools,” CJ said. But with clear quarterly cost forecasts, juxtaposed with the cloud cost savings enabled by CloudZero, CJ could make a compelling case to the finance team for the right tools.

Used cost per customer to drive pricing strategy

CloudZero facilitated a COGS-based cost per customer, allowing NinjaCat to make informed pricing decisions that would increase revenues without having to add any new functionalities. Not only could that mean charging the highest-use customers more, it could also mean charging smaller customers less.

“Saying to the smaller ones, ‘You’re paying us $X a month, but you really should be paying us $Y,’ because this is how much you’re costing us, is a great way to have transparent conversations with our customers. Smaller customers probably won’t stay small forever. They’ll grow, and as they do, they’ll remember that we looked out for them.”

Like nearly all modern companies, NinjaCat faced pressure to transition to the cloud, and did so successfully. Once there, they realized they had a blind spot with regard to cloud spend — a near-universal blind spot that companies can only fix through cloud cost intelligence.

CloudZero has already fulfilled CJ’s expectations. “The savings paid for CloudZero itself — and then some. It gave us the budget leeway to get the tools we need,” CJ said. “CloudZero helps companies get a handle on their cloud cost, mitigating the overspending they might have incurred in the transition to the cloud.”