We caught up with the founders of MammothDB at the CODE_n contest in Hannover. MammothDB was chosen from amongst various global big data startups to present their technology and services at CeBIT 2014. MammothDB is a specialized analytics database – much like Teradata or Vertica – but offered at a significantly lower cost. On the front-end, MammothDB provides a native MySQL port so that any visualization tool (from Excel to Tableau) can be connected to MammothDB. MammothDB‘s cost offering is transparent – offering both “on-premises” and “cloud” solutions.
Who are you?
We’re MammothDB and we basically do inexpensive analytics. We’re a little bit more old school, we like the classical data warehouse, the enterprise data warehouse; meaning that SQL has been around since the ‘70s, and is a great tool – and a popular one! It’s a bit crazy to throw just ignore the lessons we’ve learned from the past and migrate completely to NoSQL. Certainly there is a place for both, but the trend of going completely to “no schema” or “schema only on read” only makes sense for certain types of (big) data. And frankly, we think the term “big data” itself is misleading – we believe big data can actually be even a hundred and fifty gigabytes; a size which most enterprises struggle with analyzing effectively. If you want an analogy, we’re like the “Easy Jet of airlines in the aviation world. We are practical, we get you there, and we’re inexpensive!
How did you come up with the idea?
We came from the BI Consulting, and Data Warehouse Consulting field, and we kept losing deals because the licenses and hardware costs for enterprise analytics are so expensive. When you’re on a fixed budget, allotting 80% of that to Oracle or Teradata is absolutely not doable, along with expensive servers, is difficult.
The worst part of the high license and hardware costs is thatit’s the consulting that adds value! For example, you don’t just install any ERP solution and it just happens to magically work for your business. You need a lot of consulting and customisations to ensure that the solution properly reflects your business requirements, processes, and needs. It’s the same thing with enterprise BI and data warehousing – the real value comes from making sure the solution represents the needs of the business, not a box of expensive software.
Even more challenging for the BI world is that while important, “business intelligence” is often viewed as a “nice to have”, but they’re not a critical “must have” – it’s not like operational systems – so a high solution cost is likely to be a huge barrier to adoption. And it’s a shame, because as data is getting both larger and more pervasive, the value we can derive is almost limitless. Given this difficult environment and very low rate of adoption due to high cost, we thought to ourselves, “there’s just got to be a cheaper way; a way to flip this value equation upside down so that you’ve got a lot of budget for consulting, where it’s needed, and save your budget on the hardware and software licenses side. A way to disrupt the analytics database industry and commoditize BI databases!”
What is your business model?
Our business model is to go low cost, that’s our absolute strategy. The entire concept we have is that ten percent of businesses have enterprise analytics, due to the high cost (usually a 1+ million capital expenditure barrier), so as a result, people are struggling with Excel and other small solutions in the little corners of offices. This means a lot of waste, reproduction of effort, loss of effort, and lack of access to analytics across the organization. MammothDB wants to change all of that, and provide an enterprise class analytics database at cost businesses can afford. It’s also about flexibility; saying that you can have it on premises or you can have it on your Cloud, and with easy, inexpensive, transparent pricing.
Our prices are on our website and always straight forward, unlike most of our competitors in the industry. Fifteen hundred per year for a node on site, five hundred per month per node on Cloud. Usually each node supports about a 1TB capacity, but frankly you can put more storage on an on-premises node if a customer wanted to. Our minimum node amount is two nodes. From our perspective, what CIO in their right mind is going to say, “Oh, let’s take a brand new technology, with a brand new company and pay the same price as Netezza or Teradata?!?”
What makes you unique?
First off, we’re actually designed for being low cost… Most of our competitors either have a legacy of being high cost, or the newer “big data” companies are taking the approach of saying “hey, this “big data” thing is gaining popularity, and we’re an innovative solution: so despite being a startup, we’re still going to charge you 100,000+ per license, and 7,500+ per node!”
We, on the other hand, are on a mission to allow every business to gain insights from their data, so we designed from the ground up to be low cost. So while we use Hadoop for infrastructure, and basically as an ETL layer, we then take the unique approach of having an open source columnar database on each node. On the front side, we expose a standard MySQL port! This means that any reporting tool you want to use (and most likely, you’ve already either spent money on buying and training costs for a visualization tool, or you want to pick a best of breed visualization tool!). In fact, you can even use Excel for your visualizations if you want – especially given that almost everyone in the organization has access and skills in Excel.
Our point is that an all-in-one solution stack (which includes visualisations) costs more to build, will never be best-of-breed, and ends up wasting your existing resources. With MammothDB, you can use anything really, the options are open, and you can have a very inexpensive approach if you decide to use Excel, Siku, or other inexpensive visualization tools. Of course, if you want to use a more robust tool, such as Tableau or Microstrategy, you have that possibility as well. .
So we provide a truly enterprise scale analytics database which is compatible with data ingest tools (ETL tools), is completely compliant with SQL tools, including with Star and Snowflake data schemas, and with any visualization tool. Additionally, we’ve adopted best-of-breed open source technologies (such as components of Hadoop and MySQL) into our solution – meaning. Finally – we focus on the core functionality of the analytics database! As a result, our entire implementation is scalable, yet inexpensive. That means we can sell fifteen hundred Euros per node, per year on premises – a claim that nobody else can make. .”
Anything else that makes you stand out?
Well, there’s a few things that make us stand out. Low cost is one, but the other is we’re also very responsive. We’re the little guys, so that means we try harder. We’re really out there putting in an effort to be very flexible, and building as many connectors as we can for different types of data, for example. We’re dedicated to building a lot of connectors to popular data sources, such as Facebook and Google Adwords, to make it even easier for customers to bring their data into the data warehouse. We are highly responsive to our clients’ needs, and we also happen to have expertise in logistics, so we even helped them with the algorithm for predictive analytics as well.
We also feel that our options for either on-premises or cloud based provides additional flexibility to customers. In reality, there are very few Cloud based real enterprise BI solutions available, and that’s another offering we have which a lot of our competitors don’t like. Sure, there are dashboarding solutions for limited data-sets, but try brining in operational data from your ERP solutions into these!
Any specific case studies?
One of our customers is a large logistics company, and what we are doing is we are working with them to try to predict the cost of moving air freight between different cities in the future. We’ve developed a predictive model for them based on our own algorithms and integrated this into MammothDB. We run this model against historical data, and of course we also compare the results with actual data, so that we can continue to refine the model. From there, the client can access these reports online, see how well the model is working, as well as get access to specific tariffs.
Another interesting case study is a division of a global media company that we’re working with. They’re interested in bringing the results of campaigns across multiple media channels into one central repository in order to report across channel. We bring all that data into MammothDB, carry out specific transformations of the data, and then allow them to get a broad view of how any specific campaign is performing across an array of social media channels, all in one report. This saves them a lot of manual work, and provides a clear view of the efficacy of campaigns on each platform.
Are organisations willing to readily share their data with you?
Actually we’re seeing that more and more customers are comfortable with using Cloud based solutions. We sign an NDA, of course, and currently we’re hosting the majority of customers in the Cloud. My feeling is that as they grow the solution, they may want to bring the solution in house, and migrate MammothDB on-premises into their own internal data center.
Are you looking for any funding or any special talent(s) to hire?
Looking for funding? Absolutely! We’re looking for European venture capital, as we’re based in Europe, and we also feel that the European market is under-served, as most of the existing providers are heavily focusing on the USA market.
Where do you see yourself in a year from now?
If we get funding we’ll be more aggressive in the European market, and hopefully better known. We would like to also really hit the emerging markets; telecoms in Pakistan, for example, and regions where customers really care about value. Especially with telecoms, it’s the right industry for big data and data analytics.
I think it would be really cool if we could get a developer community built as well, and also to get distribution on AWS and other Cloud platforms such as Rackspace, as well. Also, we’d like to get more integrated with the consulting firms and VARS. Since we lower the cost of hardware and software, we free up budget for the consulting and reseller firms to capture more value. Finally, it would be great to see our solution as part of an overall analytics environment in larger firms.
Our idea is that since we’re low cost, the local departments or divisions of larger companies don’t have to go out to their central procurement organisations because the price is so reasonable. If they want to prototype a solution, or use MammothDB to explore some ideas, they can do so without prohibitive cost, and they can run MammothDB side-by-side with their existing data warehouse software, and also ingest data from their ERP solutions easily. Sort of like a “second car” solution that enables them to try new things without exorbitant costs.
So, one year from now we’re hoping to be in a lot of smaller divisions of departments of larger companies throughout Europe.
We caught up with the founders of MammothDB at the CODE_n contest in Hannover. MammothDB is a specialized analytics database – much like Teradata or Vertica – but offered at a significantly lower cost. On the front-end, MammothDB provides a native MySQL port so that any visualization tool (from Excel to Tableau) can be connected to MammothDB. MammothDB‘s cost offering is transparent – offering both “on-premises” and “cloud” solutions.
