I've written a lot this year about the NoSQL market, looking at how the likes of MongoDB, DataStax, Couchbase and Basho are proving to be reliable and interesting alternatives to the traditional relational database model. As the companies have grown and feathered out their capabilities, they have started to attract the interest of enterprises that want to take advantage of distributed and complex models that are suited to web-based architectures.
For example, the likes of Uber and Netflix are big advocates of the NoSQL market, as they have found that they can operate at speed by making use of unstructured data, at scale, all running on commodity hardware.
As mentioned above, Basho is one of the top four vendors in this space vying for market share. Given that it's early days, it's all very much still to play for. Last time I spoke to Basho it had just announced a new data platform, which it was hoping would introduce a level of operational simplicity to the Riak database. Simplifying Basho's offering was at the top of CEO Adam Wray's agenda when he was appointed to run the company in March last year.
Wray is in London for the month to focus on Basho's European sales and operations – as almost half of the company's business comes from the region – and I sat down with him to discuss what he has been up to since we last spoke and how the company's strategy is progressing.
Unfortunately some of the *really* interesting stuff I can't report on just yet, but I am told that I can reveal more in a couple of months. Needless to say, Basho is worth keeping an eye on. Nonetheless, Wray still gave me plenty of details that I can share.
First up, we discussed the company's recent opening of its HQ in Bellevue, Washington, which is just across from Seattle. It's always interesting to hear why a CEO chooses a location to set up and run their business, just because it sometimes gives you a hint to their culture and approach to development. It's also interesting when a company chooses NOT to put their business in San Francisco, because we are so often told that anyone outside of the bubble is missing out.
However, Wray disagrees and reckons that there are advantages to locating a couple of hours flying time north of the Bay area. And mostly this comes down to the companies you are surround yourself with and developer talent. He says:
There was just a lot more developer talent available, from the spin off of AWS and Microsoft. We have developers in the Bay, but more of our developers are non-Bay centric. More of our developers are in Denver, Dallas, got a couple in Edinburgh. And the value is that we can cherry pick these incredibly highly talented people and when they come on board they're not looking to change their job every twelve months, which is very common in Silicon Valley.The joke in Seattle is that it has become a suburb of San Francisco. Whereas San Francisco is very focused on social, gaming and things of that nature, what you see in Seattle is a lot more enterprise and cloud. It's still by far a much better place to build a business in tech, because you're right in the middle of the scene, but you're not necessarily in the midst of all the craziness that happens in the Valley.
Part of Basho's agenda with the announcement of its Data Platform, was that it wanted to remove some of the complexities around operating Riak and introduce more integrations that would help IT focus on the services it needs to deliver, instead of thinking about the data tier or what it needs to mesh together.
Wray explains that this has proved to be a differentiator for Basho, as he believes whilst his competitors are focusing on developers and point solutions, his approach allows a broader discussion to take place with those people higher up the chain of command.
We had a thesis that the core routing foundation of Riak could be used to develop a platform. If you look back at March of last year when I came on board, we went down this path of purposeful collaboration with our partners, with our clients to build on that foundation that has got five years of distributed scale, high availability for production environments. What you are seeing is the output of that in the last 60 days, where we are going public with what we are doing.
We are rolling out search, rolling out in-memory and we are rolling out analytics with Spark. From my perspective we are transforming to become very agile, whereas historically we did very monolithic releases. We want to be constantly innovating and updating. For example, we had to turn the entire engineering team into an agile group, where they are now doing code drops daily. Every other week we have a demo day internally, etc.
We also put the strategy together in the marketing and sales engine to be able to support what is now a more CTO and chief architect type sell as a persona. People that are looking holistically at the perplexities that their organisation will face, trying to simplify a multi-model world, simplify the ability to replicate and synchronise their data, to be able to build solutions on top of that, go after the business need. I hope that that story will over time resonate very clearly all the way down from developers all the way up to the CTO.
Wray says that he is going into deal discussions that are $2 million to $7 million in commitments, which tohim highlights that the discussions are about impacting top-line revenues. He admits that because some of his competitors (e.g. MongoDB) have gone after the developer community and have a good foothold there, this means that he won't necessarily be “first at the table” when it comes to sales discussions. However, he argues that his experience tells him that Basho often wins these discussions when it gets there because of its scalability and operational simplicity. He says:
Half or more of my business comes from companies where they have found that relational databases were never the right thing to start with and now they need to find something that is more open and flexible for many datasets. The other 30% of the last 50% are from MongoDB or Couchbase customers where they have suddenly got production level needs that scale across multiple data centres and the operational overhead and accuracy of data is too critical to leave in those systems. The last 20% is new stuff.
The business case
I always ask the NoSQL players what sorts of discussions they're having with buyers, what the business cases and objectives are that are driving them to purchase. And that's because this is frequently changing – we have noticed that things have moved from experimentation to wide-scale deployments in recent years.
Wray says that one of the things he often talks about with customers is business agility. For example, he is currently in discussions with a $3 billion software service company (the name of which he can't mention) that has a strategy to go to a web services architecture. He explains:
As part of that they need a level of agility they can't get with their current relational core. They have 17,000 Microsoft SQL Server licences, they've sharded the crap out of that thing. Their fear is their inability to collect data, but more importantly to make that data actionable in real time, which is not feasible with this architecture. So the traditional approach was to get a data lake, which is code word for Hadoop, and they threw everything in there and realised it was no more accessible than it was sitting somewhere else.
People are saying let's drop all that in a key value store, query it and make it actionable.
And the other, more important, discussion that comes up with customers is about boosting revenues or going after new revenue streams. Wray says:
What you might not expect is that it's not usually operational cost savings, they're not leaving Oracle to save 70 cents on the dollar. The biggest thing that people forget is those companies have got an operational team that understands Oracle. So there is already a big overhead cost that is pre-invested. It's not going away. The cost to leave is as painful as the savings.
So if you look at Bet365 (which has swapped out a relational database for Riak) as an example, they were literally being capped on their top line revenue opportunity by not being able to change bets midstream. They couldn't do that in a SQL database. With us they could increase the take per individual engaged in their platform – that's a top line revenue opportunity. I literally couldn't have charged enough for them to care.
And what of the future of the NoSQL market? Does Wray think that the likes of Basho are going to standalone and continue to carve a market out independently, or does he think we are on the verge of an acquisition game? To be honest, it's a little bit surprising to me that some of the other bigger players haven't yet made a move to acquire one of the top four, but this is probably down the market needing to stabilise a bit – which Wray recognises. He says:
There hasn't been this buying spree yet, that sometimes happens in segments. Companies that are protecting that much revenue (e.g. Oracle, Micorosft) are going to defensively make a move at some point in time. I do not think that it's binary though that their desire to acquire to drive position ensures that not one or two or multiple of us could make it to critical mass and become a standalone entity for the long-haul.
The top three or four of us are all within $20 million in revenue of each other. We are all within spitting distance of each other. We all have two to three hundred real clients with us. It will be intriguing. The reason that we are incredibly bullish in our position, we are doing large enterprisedeals.
I've said it before and I'll say it again – I love talking to these guys. It's an incredibly interesting area. And that's not because NoSQL databases are the most interesting thing ever – they're not. What's interesting is how this is going to play out and what it means for the other players in the market e.g. SAP, Oracle, Microsoft.
Let's face it, the NoSQL players aren't going to stick to databases – we've seen that already with Basho and its data platform. There's a bigger play here for all of them. And as that unfolds, what will that mean for the bigger incumbents? As this market and the products expand and continue to stabilise and gain traction, they pose more of a problem. And more of an opportunity.
Disclosure: Oracle and SAP are diginomica premier partners at time of writing.