Follow Datanami:
June 17, 2021

Neo4j Sees Graph Data Science Taking Off Following $325 Million Round

(sdecoret/Shutterstock)

Operational use cases such as fraud prediction and recommendation engines have dominated in Neo4j’s existence up to this point. But following today’s announcement of a $325-million funding round that values the graph database company at over $2 billion, Neo4j expects to see more widespread use of graph algorithms for data science and analytics use cases.

2020 turned out to be a pivotal year for graph databases, just as Neo4j founder and CEO Emil Eifrem predicted back in 2015. But the adoption of graph technologies was significantly accelerated last year by the unpredictable presence of COVID-19, which was tailor-made for graph solutions.

“Gartner published a piece in April of last year saying COVID-19 demanded urgent use of graph technologies,” Lance Walter, Neo4j’s senior director of product marketing, told Datanami this week. “On the COVID and corporate front, we did see a lot more appreciation [of graph] We saw three to five years of market maturation happen in what felt like a year, year and a half.”

There are other reasons for the surge in the graph databases, which is a type of NoSQL database that stores information about entities (like people, places, and things) in a highly structured format that enables queries to quickly find other entities that share a connection or a relationship such as an address or a last name. Performing the same function in relational databases typically requires huge SQL jobs known as joins that can bring a database to its knees.

Neo4j, which was founded in 2007, is the oldest and most well-known member of the graph database club. The company has over 800 paying enterprise customers, and its open source database has been downloaded more than 100 million times. This head start has enabled Neo4j to solidify its position as the leader in the graph database market, which owns just a small fraction of the overall $100 billion global database market.

“The naturel connectivity of the world around us is driving that growth,” Walter said. “I think people are starting to appreciate that there is a lot of important connection and relationship information in your data.” In fact, in some cases, the relationship information held in the data is sometimes more important than the data itself, he said.

Graph’s small-fry status started to change a few years ago, as technologists began to recognize the advantages that graph databases can bring. The public cloud vendors launched their own graph databases, including AWS with Amazon Neptune and Microsoft with Azure Cosmos. Other pureplay graph databases, such as TigerGraph, also entered the arena.

Neo4j executives welcome the new players, not because they provide competition, but because they validate Neo4j’s belief that graph approaches have an inherent advantage in cutting through big data clutter and finding the information that really matters.

So when the venture capital firms Eurazeo and GV (formerly Google Ventures) today announced they have invested $325 million in the San Mateo, California company, Neo4j took it more as proof of graph’s emergence on the big stage than a chance to grandstand.

“Obviously we’re excited at the company level, but we feel like the news is saying more about the market than the company,” Walter said. “It’s a very significant transition in the database space. We think the interesting pieces are what it’s saying about the graph category and where data and analytics are going.”

The folks at Gartner suddenly are big believers in graph. In a recent report led by analyst Rita Sallam, Gartner predicted that by 2025, graph technologies will be used in 80% of data and analytics innovations, up from just 10% in 2021. With the $325 million investment, Neo4j, as the incumbent in the graph category, is well positioned to grab a chunk of that growth.

COVID-19 helped accelerate graph adoption (Rost9/Shutterstock)

According to Walter, the company expects to see a shift in use cases from mostly transactional and operational, like fraud detection, product recommendations, and data management, toward more analytical and exploratory.

“We have had a lot of operational use case. A lot of our production use cases are transactional. It’s one of the real strengths of Neo4j,” Walter said. “With that said, we do see more and more people doing analytics use cases, blending transactional and analytics.”

In April 2020, the company launched Neo4j for Graph Data Science, which included dozens of pre-built graph algorithms that enable data scientists to utilize machine learning upon data stored in the Neo4j database. That offering includes a mix community detection algorithms, centrality and importance algorithms, and pathfinding algorithms, all designed to be powered by the Neo4j database and the data that it contains, thereby eliminating the need for time-consuming and expensive ETL jobs.

“It is huge new opportunity for us,” Walter said of the graph algorithms and data science use cases. “It almost feels a company within a company in terms of the audience of data science practitioners, interesting eco-system implementations and the like.”

Neo4j published the free book “Graph Data Science for Dummies” in 2020

Interest in graph algorithms has grown organically at Neo4j over the past two years. With the right level of cultivation from Neo4j and its open source community, including the publication of the free book Graph Data Science for Dummies, it’s now on the verge of flowering into a major revenue driver for Neo4j.

“It’s a little bit like the overnight sensation that’s actually been bubbling for a little while now,” Walter said. “We first made graph algorithms available in our community a few year ago just to see what people would do with them. We saw, in the last few years, everything from Academia publishing a lot more research papers on AI and ML that mention graph in either the title or the abstract. We saw our customers actually doing stuff in production. And we saw the activity in the community. [We said], okay this is a successful community experiment that will  now graduate into a production-ready enterprise project.”

The launch of Neo4j for Graph Data Science is changing how people use Neo4j’s graph database. If you were to take a snapshot of Neo4j today and Neo4j in a few years, the mix of use cases will change dramatically, Walter predicted.

“More of the use cases [today] are the operational use cases than data science,” he said. “But there’s no doubt that the mix is shifting, and the growth is faster on the data science side.”

With that said, there are a lot of other variables impacting Neo4j’s growth at the moment, and a major one is the cloud. Many of the company’s enterprise customers manage their own IT infrastructure, but that is changing thanks to the company’s 2019 launch of Neo4j Aura, its graph database managed service. Today, the company supports all major cloud platforms with Aura.

About 70% of Neo4j’s customers are now running in the cloud, Walters said, and that percentage is growing all the time. “I would say, in new transactions, cloud is the default, unless there’s some particular reason not to,” he said.

The $325 million will allow Neo4j to grow all areas of the business. The company is hiring in all departments and all geographies, Walter said. That includes more engineers and more customer support personnel to help it grow in Europe, North America, Asia Pacific, and the Middle East.

“We’ll be accelerating everywhere now with the fresh capital,” Walter said.

Related Items:

Neo4j Brings Graph Database and Data Science Together

Tracking the Spread of Coronavirus with Graph Databases

Graph Databases Everywhere by 2020, Says Neo4j Chief

 

Datanami