Investment Notes: Factor house

We are excited to announce the Blackbird 2022 LP is leading a $5M Seed round into Factor House.

Factor House is building developer tools that allow engineers to manage the processing of massive amounts of real-time data.

‍

Built by engineers for engineers

Factor House was founded by CEO Derek Troy-West and COO Kylie Troy-West in late 2019.

While homeschooling their two young children through Melbourne’s lockdowns, Derek and Kylie built Factor House out of their apartment. Five years on, and with just 5 employees, Factor House surpassed US $1M in revenue and is growing 90% YoY from a customer list that now includes the likes of: Hewlett Packard, Block, Binance,,, Nord LB, and Airwallex, alongside an additional 8 Fortune 500 clients

Derek is a lifetime engineer (going all the way back to the year 2000 when he won the Senior Prize in Computer Science at the University of Auckland). He built his career inside financial institutions, developing a reputation as the guy who built systems that wouldn’t fail and handling screaming phone calls from traders in Hong Kong who were not willing to accept an extra millisecond of lag time on trades.

Derek also runs Melbourne Distributed, a meet-up of engineers in Melbourne working on distributed systems, and is regarded as one of the world’s best Clojure engineers – watch him present at the Clojure 2019 conference. 

Kylie has 20+ years of operational experience at technology and arts management companies and has built trust and reputation in highly technical communities; this has lent itself to an unfair advantage in hiring the world’s best engineers to work with them due to their profile. For example, Eleanor Toulmin (ex BuildKite) started out advising Factor House, and Kylie quickly converted El to sign on as their Chief of Staff.  

Alex Hilton, an Engineering Director at Airwallex, describes Factor House and the team like this: “Factor House is more than a product; it’s a philosophy. It’s built by engineers for engineers with a deep appreciation for the craft. Derek, being an engineer himself, understands the support requirements and the life of an engineer.”

We asked Alex if there was any other product his team used that came close to this level of love, and he said there wasn’t and that he couldn’t recall any other similar tool in his 20-year career – this is core to why we are excited about this investment, it’s a rare engineer that can build a product that people describe in such glowing terms. 

‍

Apache Kafka® & Apache Flink®

To fully appreciate Factor House, we first need to have a basic understanding of Kafka and Flink, the platforms that Factor House is built to manage. 

In 2009, LinkedIn built an internal system (Apache Kafka®) to address the challenges of processing and managing massive volumes of real-time data. In 2011, Kafka was donated to The Apache Software Foundation, a non-profit supporting open-source software projects.

This marked the beginning of Kafka’s journey as an open-source, distributed (can run across many servers) event-streaming platform (like user activity on a website or sensor readings from a machine).

Other companies quickly began leveraging Kafka:

• Tech companies like Netflix & Twitter: manage real-time user activity streams for recommendations and analytics
• Financial institutions like Goldman Sachs & JPMorgan: for real-time processing in trading platforms, fraud detections, and transaction monitoring
• Retailers like Target & Walmart: to track real-time inventory and enable personalised offers or notifications
  
  

Today, 80% of Fortune 100 companies use Kafka. 

At the same time LinkedIn was developing Kafka, a group of researchers at the Technical University of Berlin were creating a better data processing engine for distributed systems (independent computers that work together, like cloud computing platforms), which they would later call Flink (German for quick). In December 2014, Flink was also donated to the Apache Software Foundation and officially released.

Kafka & Flink in practice

We can understand the jobs done by each platform (Kafka and Flink) with an analogy.

Imagine Kafka as a fleet of buses in a busy city, constantly transporting passengers (data) between different locations (systems or applications). The passengers keep getting on the bus in real time, and Kafka ensures they get delivered efficiently and reliably to the right location.

Flink are bus conductors. While a bus (Kafka) is in motion, the conductors (Flink) analyse and process the passengers (data) in real time. The conductors make sure the right actions are taken while the bus is still on its route, not after it has arrived. 

In short: Kafka transports the data, while Flink processes and transforms that data in real time to generate valuable insights or actions.

In an engineering context, each bus is called a cluster. For a customer like Airwallex, each cluster carries specific financial events between different systems in the company, like payments, currency exchanges, or transaction logs. Kafka ensures these events arrive at their destination without any data loss no matter how fast or how many (often thousands), events are being processed simultaneously.  

Flink monitors these events in real time, detects fraud, triggers alerts, or performs currency conversions instantly. For example, if Airwallex needs to check for suspicious transaction patterns, Flink would do the heavy lifting of identifying those patterns while the data is still in motion. 

‍

Kpow for Apache Kafka®!

Factor House has three products, starting with Kpow, their flagship product.

Kpow is an engineering toolkit specifically designed to enhance and simplify the management of Kafka. It provides a comprehensive control centre UI for monitoring and managing Kafka clusters – watch a 90sec demo of Kpow (narrated by Derek). 

Kpow is easy to install and lightweight enough to spread quickly within an engineering organisation. As proof, below is a chart showing MAUs over the last 24 months, reaching 5.1K MAUs in June 2024, growing 100% MoM. The monthly session count over the same period hit highs of 48K sessions in April and May 2024, growing 120% MoM – the average session run time in 2024 is 30 mins. 

‍

‍

In 2024, Kpow totalled 24K users across 86 countries, who have run 481K sessions and counting (2-3x-ing YoY).  

Kpow has 71 commercial customers, which include Airwallex (since 2019, Airwallex has relied on Kpow to aid delivery and operations of their global payments and finance platform) and US Foods, which relies on Kpow to manage its internal e-commerce platform that serves 250K restaurants and foodservice operators.

 The team also has 1.5K freemium customers in their Kpow community using a light version of the product, acting as a great organic lead generator. 

Flex makes it easy for engineers to manage systems and platforms powered by Flink. As companies are generally already on Kafka if they decide to implement Flink, there are promising upsell/cross-sell opportunities with Flex. 

To be released in 2025, Factor House platform unites Kafka and Flink (a world first), letting customers manage their entire streaming ecosystem from a centralised UI. 

Using Airwallex as an example, they would use: 

• Kpow for sending a notification to a customer when a deposit lands in their account 
• Flex to block a transaction that they think might be fraudulent based on analysis of patterns in transaction data
• Flow to unite the data streaming in from the Kpow clusters, and the Flex analysis of transaction data to adjust its risk models across all its systems in real-time

Airwallex would install Kpow in the first instance, and then as usage deepens, buy Flex and Flow as additional modules. 

Critically, the purchase decision is an engineering one – engineers can decide to use Kpow before the CTO decides to buy Flex and the Factor House platform. 

The opportunity here is to build the critical set of tools for any engineer leveraging Kafka and Flink to manage massive datasets in real time. As Kylie describes it: “Ten years from now, Kpow, Flex and Flow will be the industry standard for real-time engineering because, in the end, developers always choose the best solution.”

‍

Over the past decade, Derek and Kylie have discovered a secret about the world: in order to help engineers who build on Kafka be superheroes in their roles, the missing link is Factor House.

Learn more: https://factorhouse.io/