Going mainstream: Neo4j and the future of the graph database

The release of the Panama Papers shined a spotlight onto graph databases and leading supplier Neo4j.
Written by Colin Barker, Contributor

A few years ago, graph database software was little known but that's changing. A big part of the reason was the release earlier this month of the Panama Papers, an event that dramatically changed the perception of the graph database technology, which was used to analyse the vast trove of information.

ZDNet recently spoke to Neo4j's CEO, Emil Eifrem, about the Panama Papers, graph technology, and plans for the company's future.

ZDNet: What's the next step for Neo4j?

Now, for the first time, we are starting to get competitors in our space, and that's a great thing.


A Neo4j graph: Simple in practice, graph databases can illustrate enormously complex concepts and the release of the Panama Papers proved that point.

Graphic: Neo4j

Why? Because it means we must be doing something that is valuable.

Q: Your company is at an interesting stage, isn't it? You might say it's at a crossroads?

I look at it in terms of the chart I showed where in one leg we have the predicted number of devices in the world and by the end of 2020 the prediction is 50 to 75 billion.

On that scale we are not too far along and there is a lot of space to grow in.

One way to see how that change affects us at Neo is that traditionally, at [industry] events, I have mainly been talking to ultra-geeky, early adopters -- and now it's going mainstream. That's what it feels like, it doesn't feel like a crossroads.

Q: An exciting time though?

Very, very exciting. In the beginning it was exhausting and now it's exciting.

Q: Now the Panama Papers are out and the journalists involved used your software, what's your assessment of the impact of that? Presumably you have had a lot of interest from people who now will have heard of graph databases and may be interested in buying you?

Presumably. What we do is incredibly differentiated. Nobody can do what we can do. Now you can part do that with some of these other databases, with relational databases. That's usually a more convenient way of doing it -- for example in a document database.

What we do, you cannot do in a relational database. It's impossible.

We're one of those where we are not 10 percent better, or 20 percent better, we're a 1,000 times better, we are a million times faster.

Now, that's not for everything but it is for some very, very valuable operations. And we have a customer base that is very enterprise oriented, Fortune 500.

Q: Now this next version, Neo4j 3.0, do you see this as a game-changer? When you say that with 3.0 you can have an unlimited number of nodes, what does that mean in practical terms?

Now, as the saying goes, 'in theory, theory and practice are the same but in practice they are different'.

When we say that it is unlimited in the real world, what performance we get is limited by the file systems capability or availability, is limited by questions like, do we have enough hard disk space? So, of course it does have an upper limit, it is just that the database no longer has a physical upper limit.

So you can push the database as far as you can push the hardware and the other physical components of the system. It is just that Neo4j will never be the bottleneck, something else will. And that is very, very significant.

Q: What kind of customers are you attracting?

In general we have a customer base that is very willing to talk about us. You know, people like talking to my executives who have done this many times over.

They will tell you that, uniquely, this company is willing to go on the record that they are using us. Now, we don't really know why, but we think that many of them see it as a recruitment advantage -- they want to be seen as one of the early adopters of this technology, to show that they are on the [leading] edge.

Now we think that is why they want to talk about us, in fraud detection they are all like that. But in government, intelligence, and defence, they don't want to talk about it.

But, then we are doing things like, real-time recommendations, where we have four of the global ten major retailers [as customers]. That is one of the use cases where you get really top-line impact.

In real-time recommendations, that's not funded by IT, that's funded by people like the VP of online sales, the VP of up-sell and cross-sell at these big retailers. So that is really funded by the business and it comes from the retail organisation or the sales organisation.

You know, McKinsey has [research] that shows that more and more of the IT budget is floating over to the CMO and that is where we get lovely use cases and people are more than willing to talk about it.

Q: So when you look back at what you have achieved here, was there a moment when you thought, 'hold on a second, this can be really, really big'?

Linus Torvalds did a talk recently and he said that there was not a single moment when he thought that Linux was going to take off, it was just a series of small things and I think that is the same for us.

I also have this character flaw where I set this really high bar, this goal and when I get close to that, I set that as the new baseline. So I never celebrate success, I am always on to the next thing.

I wouldn't be surprised though, that in 2020, when we look back on this decade, we will think that the Panama Papers was the biggest, mainstream story that was associated with us.

And, of course, in 2020 when people are writing their memoirs they will say, 'oh yes, I could see that was going to happen!'

Read more about Neo4j and graph technology

Editorial standards